• Copy the reference
  • Tutorial video
Source(s) of the individual document(s):

Lawyers, other representatives, expert(s), tribunal’s secretary

Report of the Panel

I. INTRODUCTION

1.1.
This proceeding has been initiated by two complaining parties, the United States and New Zealand.
1.2.
In a communication dated 8 October 1997 (WT/DS103/1), the United States requested consultations with Canada in accordance with Article 4 of the Understanding on Rules and Procedures Governing the Settlement of Disputes ("DSU"), pursuant to Article XXII:1 of the General Agreement on Tariffs and Trade 1994 ("GATT 1994"), Article 19 of the Agreement on Agriculture, Article 30 of the Agreement on Subsidies and Countervailing Duties ("the SCM Agreement") and Article 6 of the Agreement on Import Licensing Procedures with respect to export subsidies of Canada on dairy products and the administration by Canada of its tariff-rate quota for fluid milk and cream. The United States and Canada held consultations in Geneva on 19 November 1997 but these consultations did not result in a resolution of the dispute.
1.3.
On 29 December 1997 New Zealand requested consultations with Canada pursuant to Article 4 of the DSU, under Article 19 of the Agreement on Agriculture and Article XXII:1 of the GATT 1994 with regard to Canada's Special Milk Classes Scheme. New Zealand and Canada held consultations on 28 January 1998 but these consultations did not result in a resolution of the dispute.
1.4.
On 2 February 1998, the United States (WT/DS103/4) and on 12 March 1998, New Zealand (WT/DS113/4), each requested the establishment of a panel with standard terms of reference.
1.5.
In its request, the United States claims that:

(a) "The Government of Canada is providing subsidies, and in particular export subsidies, on dairy products through its national and provincial pricing arrangements for milk and other dairy products without regard to the export subsidy reduction and other WTO commitments undertaken by Canada. Specifically, the Government of Canada established and maintains a system of special milk classes through which it maintains high domestic prices, promotes import substitution, and provides export subsidies for dairy products going into world markets. These practices distort markets for dairy products and adversely affect US sales of dairy products."

(b) "In addition, although Canada committed under the Marrakesh Agreement Establishing the World Trade Organization to permit access to an in-quota quantity of 64,500 tonnes (product weight basis) under a tariff-rate quota for imports of fluid milk and cream, Canada has refused to permit commercial import shipments within the quota. Instead, Canada is administering this tariff-rate quota in a manner that denies market access."

(c) "These measures appear to be inconsistent with the obligations of Canada under the General Agreement on Tariffs and Trade 1994 (GATT 1994), the Agreement on Agriculture, the Agreement on Subsidies and Countervailing Measures, and the Agreement on Import Licensing Procedures. The measures in question are the Canadian Dairy Commission Act, agreements of the Canadian Dairy Commission, the Interprovincial Comprehensive Agreement on Special Class Pooling (as well as the P-4, P-6, and P-9 interprovincial pooling agreements), the National Milk Marketing Plan (and amendments thereto), operations of the Canadian Milk Supply Management Committee, the Dairy Products Marketing Regulations, and Canada’s administration of its tariff-rate quota on fluid milk and cream (as reflected in its implementation of its WTO Schedule of Concessions)."

(d) "These measures are inconsistent with the obligations of Canada under Articles II, X, XI, and XIII of the GATT 1994; Articles 3, 4, 8, 9, and 10 of the Agreement on Agriculture; Article 3 of the Agreement on Subsidies and Countervailing Measures; and Articles 1, 2 and 3 of the Agreement on Import Licensing Procedures."

1.6.
In its request, New Zealand claims that:

(a) "The Government of Canada is providing export subsidies on dairy products in contravention of its export subsidy reduction and other WTO commitments as encapsulated by the Agreement on Agriculture and the General Agreement on Tariffs and Trade 1994 (GATT 1994). The dairy export subsidy scheme in question is commonly referred to as the "special milk classes" scheme. The background to, and details of, the "special milk classes" scheme is contained, though not necessarily exclusively, in the following documents:

(i) the Canadian Dairy Commission Act;

(ii) the Comprehensive Agreement on Special Class Pooling (the P9 Agreement);

(iii) the National Milk Marketing Plan (NMMP);

(iv) the Agreement on All Milk Pooling (the P6 Agreement); and

(v) the Western Milk Pooling Agreement (the P4 Agreement)."

(b) "The "special milk classes" scheme referred to above is inconsistent with Canada's obligations under the following provisions:

(i) Articles 3, 8, 9 and 10 of the Agreement on Agriculture; and

(ii) Article X:1 of the GATT 1994."

1.7.
The Dispute Settlement Body (DSB) agreed to each of these requests for a panel at its meeting of 25 March 1998 (WT/DSB/M/44). The DSB further agreed that the two panels be consolidated as a single panel pursuant to Article 9.1 of the DSU with the following standard terms of reference:

"To examine, in the light of the relevant provisions of the covered agreements cited by the United States in document WT/DS103/4 and by New Zealand in document WT/DS113/4, the matters referred to the DSB respectively by the United States and New Zealand in these documents, and to make such findings as will assist the DSB in making the recommendations or in giving the rulings provided for in those agreements".

1.8.
On 12 August 1998, the parties to the dispute agreed on the following composition of the Panel:

Chairman: Professor Tommy Koh

Members: Mr. Guillermo Aguilar Alvarez

Professor Ernst-Ulrich Petersmann

1.9.
Australia and Japan, and the United States in respect of the New Zealand claims, reserved their rights to participate in the Panel proceedings as third parties.

II. FACTUAL ASPECTS

A. THE CANADIAN DAIRY SECTOR

1. General

2. Components of the Canadian Dairy Policy

3. The Canadian Dairy Commission (the "CDC")

4. Provincial Milk Marketing Boards

5. The NMMP

6. The CMSMC and the MSQ.

B. THE CANADIAN SPECIAL MILK CLASSES SCHEME

1. Background

2. The Special Classes

3. In-quota milk and over-quota milk

4. The price of milk to the processor

(a) Other Classes (Classes 1 – 4)

(b) Class 5 (Special Classes)

5. Returns to producers from exports

(a) General

(b) In-quota exports

(c) Over-quota exports

6. Pooling

C. CANADA’S TARIFF-RATE QUOTA ON FLUID MILK AND CREAM

III. CLAIMS OF THE PARTIES

A. EXPORTATION OF DAIRY PRODUCTS

1. Product coverage and period of time

3.1.
The dairy products and marketing years covered by the claims of New Zealand and the United States are set out in Table 1 below.

Table Error! Unknown switch argument. - Products1 and marketing years2 subject to the complainants' claims
Butter Cheese Other Milk Products
New Zealand 1995/96 1996/97 1995/96 1996/97 1996/97
United States 1996/97 1996/97 1996/97
1 Butter consists of products classified in 0405.10 and 0405.90. Cheese consists of products provided for in 0406.10, 0406.20, 0406.30, 0406.40, and 0406.90. Other Milk Products includes milk and cream in 0401.10, 0401.20, 0401.30; powdered whole milk and cream in 0402.21 and 0402.29; condensed evaporated milk in 0402.91 and 0402.99; buttermilk and yoghurt in 0403.10 and 0403.90; milk protein concentrate, 0404.90; and ice cream, 2105.00. Although the United States understood that Canadian exports of Skim Milk Powder were not in excess of Canada's WTO commitments, the United States considered that all exports under the SMP category that were exported through the Special Milk Classes Scheme should have been notified as subsidies to the WTO. 2 New Zealand did not refer to the marketing year 1997/98 because official figures for that period were not available. Nevertheless, if those figures were to indicate that Canada's actual exports also for that period exceeded its reduction commitments in respect of the products mentioned in the table, New Zealand would consider that Canada had also breached its WTO obligations in respect of those products for the 1997/98 marketing year. The United States noted that although Canada had not yet reported to the WTO its export quantities for the 1997/98 period, based on preliminary information for that period, the volume of exports appeared to remain at levels exceeding the pertinent reduction commitments. After our first substantive meeting, the figures for marketing year 1997/1998 became available and are incorporated above in Table 2 in para.2.41.

2. Nature of Measure

3.2.
New Zealand and the United States claimed that there was extensive government involvement in all critical aspects of Canada's Special Milk Classes Scheme, from its initiation through to its administration and operation. Canada's Special Milk Classes Scheme was a product of governmental authority and was operated under the auspices of the federal and provincial governments. This government involvement in the scheme was sufficient to constitute government action within the meaning of the jurisprudence developed by GATT and WTO panels.
3.3.
Canada claimed that the Complainants' assumptions of government control, direction or mandate were without basis in fact and were, therefore unsustainable. Government involvement was limited to providing an appropriate regulatory framework and essentially responsive to the initiatives of the Canadian dairy industry.

3. Agreement on Agriculture

(a) Article 1(e)

3.4.
Both New Zealand and the United States claimed that the Special Milk Classes Scheme was an export subsidy in the sense of Article 1(e) of the Agreement on Agriculture.
3.5.
Canada claimed that as the sales of milk at differing prices under Special Classes 5(d) and (e) did not constitute a "subsidy" pursuant to the definition of the SCM Agreement, it followed that these sales could not constitute a subsidy for the purposes of the Agreement on Agriculture. Therefore, by definition, such sales could not constitute an "export subsidy" within the meaning of the definition in Article 1(e) of the Agreement on Agriculture.

(b) Article 9.1(a) and (c)

3.6.
New Zealand and the United States claimed that the Special Milk Classes Scheme constituted export subsidy practices listed in Article 9.1(a) and (c). As such, these practices were subject to reduction commitments under the Agreement on Agriculture. Canada refuted both these claims.

(c) Article 3.3 and Article 8

3.7.
New Zealand and the United States claimed that Canada's provision of export subsidies under Article 9.1(a) and (c) of the Agreement on Agriculture in excess of its scheduled export subsidy commitments was a violation of Article 3.3 of that Agreement. Furthermore, Canada was in violation of its obligation under Article 8 of the Agreement on Agriculture not to provide export subsidies otherwise than in conformity with the Agreement on Agriculture.
3.8.
Canada claimed that since the sales of milk at differing prices for domestic and export markets did not constitute an "export subsidy" as that term was defined in Article 1(e) of Agreement on Agriculture, the practice at issue did not fall within the scope of Article 8; that article could therefore not apply.

(d) Article 10

3.9.
Alternatively, New Zealand and the United States claimed that Special Classes 5(d) and (e) of the Special Milk Classes Scheme constituted an export subsidy not listed in Article 9.1 that was being applied in a manner which circumvented or threatened to lead to circumvention of Canada’s export subsidy commitments contrary to Article 10.1 and 10.3 of the Agreement on Agriculture.
3.10.
Canada claimed that Article 10 did not apply in the present case as it could not be established that there existed "export subsidies", including those export subsidies listed in Article 9.1. Nor could it be established that there was actual or threatened circumvention of Canadian export subsidy commitments.

4. Agreement on Subsidies and Countervailing Measures ("SCM Agreement")

(a) Article 1 and Paragraph (d) of the Illustrative List of Export Subsidies in Annex I

3.11.
New Zealand and the United States claimed that even on the basis of Canada's own approach to the interpretation of the term "subsidy" Canada had not shown that the Special Milk Classes Scheme fell outside the definition of subsidy under the SCM Agreement. The Scheme constituted a subsidy within the meaning of Article 1 of the SCM Agreement. In addition, that the Special Milk Classes Scheme constituted the provision of an export subsidy within the meaning of Paragraph (d) of the Illustrative List of Export Subsidies in Annex I of the SCM Agreement.
3.12.
Canada claimed that the sale of milk at differing prices did not constitute a "subsidy" within the meaning of Article 1 of the SCM Agreement. Further, Canada claimed that the practices at issue were not "export subsidies" in the sense of Paragraph (d) of the Illustrative List of Export Subsidies in Annex I of the SCM Agreement.

(b) Article 3

3.13.
The United States claimed that as Canada’s Special Milk Classes Scheme was inconsistent with Canada’s obligations under the Agreement on Agriculture it was consequently in violation of Article 3 of the SCM Agreement.

B. IMPORTATION OF MILK

1. Article II of GATT 1994 and the Agreement on Import Licensing Procedures

3.14.
The United States claimed that Canada’s administration of its tariff-rate quota on fluid milk37 which restricted access to the in-quota quantity of its tariff-rate quota for fluid milk to entries that were valued at less than C$20 and that were for the personal consumption of Canadian residents, was inconsistent with its obligations under Article II:1(b) of GATT 1994 and Article 3 of the Agreement on Import Licensing Procedures.
3.15.
Canada claimed that its current treatment of fluid milk imports was fully consistent with the terms and conditions of the tariff concession for fluid milk (HS 0401.10.10) in its Schedule. Canada further refuted any alleged violation of the Import Licensing Agreement.

C. RECOMMENDATIONS REQUESTED BY THE PARTIES

3.16.
New Zealand requested that the Panel, in accordance with Article 19 of the DSU, recommend that Canada bring its measures into conformity with the Agreement on Agriculture.
3.17.
The United States requested that the Panel find the Canadian Special Milk Classes Scheme and the denial of access to imports under the tariff-rate quota on fluid milk and cream to be inconsistent with Canada’s WTO obligations. Accordingly, the Panel should recommend that Canada bring those measures into conformity with its obligations under the GATT 1994, the Agreement on Agriculture, the Agreement on Subsidies and Countervailing Measures, and the Agreement on Import Licensing Procedures. More specifically, the Panel should recommend (i) that Canada either withdraw its export subsidies or reduce the level of its subsidized exports of dairy products to a level commensurate with its reduction commitments and (ii) that such action be taken without delay. In this regard, the United States saw no reason why Canada could not bring its export subsidies into compliance within 30 days of the adoption by the Dispute Settlement Body of recommendations and rulings. With respect to market access, the United States respectfully submitted that the Panel should recommend that Canada not apply its tariff-rate quota in a manner that denies entry at the in-quota rate to any fluid milk imports made within the quantitative limit of the tariff-rate quota.
3.18.
Canada requested the Panel to find that (i) Canada’s Special Milk Classes did not provide an export subsidy and thus did not violate Canada’s obligations under Articles 8, 9 or 10 of theAgreement on Agriculture nor under Article 3 of the SCM Agreement; and (ii) that Canada’s administration of its tariff-rate quota on fluid milk and cream was consistent with Canada’s obligations under Article II:1(b) of the GATT 1994 and Article 3 of the Agreement on Import Licensing Procedures. Canada requested the Panel to dismiss all claims brought against Canada in this case by the United States and New Zealand.

IV. ARGUMENTS OF THE PARTIES

A. EXPORTATION OF DAIRY PRODUCTS

1. Nature of Canada's Special Milk Classes Scheme

(a) Outline

4.1.
New Zealand argued that the government was implicated in all critical aspects of Canada's Special Milk Classes Scheme, from its initiation through to its administration and operation. Likewise, the United States claimed that Canada's Special Milk Classes Scheme was a product of governmental authority and was operated under the auspices of the federal and provincial governments. The Complainants claimed that the government involvement in the scheme was sufficient to constitute government action within the meaning of the jurisprudence developed by GATT and WTO panels.
4.2.
Canada claimed that the Complainants’ arguments rested entirely on erroneous descriptions of the Canadian dairy system. This was particularly true of the way that they had attempted to miscast the means by which milk was marketed for export use in Canada. Exports of milk from Canada were controlled and directed by Canadian dairy producers, not governments. As such, the assumptions of government control, direction or mandate that premised each of the Complainants' arguments was without basis in fact. Therefore, each of these arguments was unsustainable. Governments did play a role, but it was limited and essentially responsive to the initiatives of the industry.
4.3.
Canada emphasized that the objective behind the institution of the Special Milk Classes Scheme had been to provide export opportunities that were consistent with Canada’s WTO commitments, while providing reasonable continuity to dairy producers. Canada noted that the current dispute was not about domestic dairy supply management in Canada: it was about the marketing of milk in Special Classes 5(d) and (e). Even if the Canadian domestic supply management regime were found to be governmental, in Canada's view, that would not necessarily imply that the marketing of milk in Special Classes 5(d) and (e) would be governmental as such.

(b) The Authority and Role of the Canadian Dairy Commission (CDC)

4.4.
New Zealand argued that the price that processors paid for milk, both for domestic and export use, was determined by the exercise of governmental authority through the operation of the CDC and provincial marketing boards and agencies. The CDC and the provincial marketing agencies were in effect a part of the executive branch of government. The fact that they were composed largely of producers could not be allowed to disguise that fact. This was evident, in New Zealand's view, from several Canadian sources. The Act establishing the CDC granted it authority to "establish the price, or minimum or maximum price, to be paid... to producers of milk or cream... " (paragraph 2.12 and following).38 The British Columbia Milk Marketing Board, whose members were appointed by the Lieutenant Governor in Council, was "authorised to regulate the marketing of milk in inter-provincial and export trade... ".39 The CDC’s internet website40, describing the relationship between the federal and provincial authorities in the marketing of milk stated: "certain marketing activities related to industrial milk are carried out jointly between the federal government and participating provinces". The Annual Report of the CDC for 1996-97 described the Commission as a "crown corporation" and referred to the "framework" provided by the Commission for "the federal/provincial participation that is crucial to the success of the dairy sector".41 It referred to the authority of the CDC "to purchase, store, process or sell dairy products" and "to make payments to milk and cream producers".42
4.5.
New Zealand noted, in respect of a few key features of the operating procedures of the CMSMC43, that it was noteworthy that the CDC, as Chair of the CMSMC, had a de facto veto power over almost all aspects of that Committee’s decision-making.44 In other words, if the CDC disagreed with the rest of the CMSMC, it could then resolve the disagreement by its own unilateral decision. In this and in other respects, the CMSMC was intrinsically linked to the CDC, a Crown agency, and to the federal government.
4.6.
The United States argued that the Canadian Special Milk Classes Scheme depended for its existence on legislation enacted by the Government of Canada. The Canadian Government’s role did not stop with the planning and enactment of the authorizing federal legislation. The CMSMC established and revised the annual national production quota. The CDC established the target prices and Special Milk Class prices. The CDC issued the permits that were required to initiate surplus removal, which was then exported. The CDC, working with the provincial marketing boards (the powers of which were derived from the provincial and federal governments), calculated the sales returns received by the provinces, and adjusted those returns to reflect participation in the Special Milk Classes. The CDC had even financed the Special Class distributions by obtaining a line of credit.45
4.7.
The United States argued that the Canadian federal and provincial governments had demonstrated the compulsory nature of the dairy regime’s production quotas, administered price levels, and revenue pooling; the CDC and the Province of British Columbia, had for over a decade, sought through legal action to prevent producers in that province from shipping milk without the benefit of a quota allotment under the federal/provincial MSQ.46 Those milk producers had contested the authority of the provincial government to regulate the production and marketing of milk. The reaction of the federal and provincial government was instructive regarding the urgency with which they met this perceived challenge to supply management and the levies which subsidized dairy exports.47 Following an initial set back in litigation in British Columbia, both the federal government and provincial governments throughout Canada amended the governing legislative authority to address the court’s finding that the provincial marketing board in British Columbia possessed only the power to regulate intra-provincial trade and, therefore, could not either regulate production or impose the levies necessary to finance the system relating to industrial milk. In response to the court’s decision, the applicable federal legislation was amended to delegate authority to the provinces to regulate inter-provincial and export trade and the provinces then amended their own authority to reflect this delegation of administrative powers from the Government of Canada.48
4.8.
The United States noted that the Attorney-General for Canada joined the litigation to preserve the authority to collect levies to finance dairy product exports and to dispute the contentions of the unlicensed dairy farmers that the exercise of power by the provincial marketing board and the Canadian Dairy Commission was ultra vires.49 In submissions to the court, the Attorney General stated that "it was intended by Parliament that the Governor in Council delegate to others the administration of such a scheme [administration of a quota based regulatory system] and, in the Federal Regulations there was a valid delegation of administrative powers to the Commission [the CDC], the Committee [the CMSMC], and to provincial Boards."50 Later in the same document, the Attorney General declared that the:

"Federal Regulations were passed to provide federal legislative support to the continued regulation of the dairy industry in Canada. At its core, regulation of this industry is accomplished by the joint participation of both federal and provincial authorities. This is reflected in the National Plan and in the marketing scheme created by the Federal Regulations".51

4.9.
The United States noted that the Special Milk Classes Scheme was established through the collaborative efforts of the federal and provincial governments. Specifically, the Special Milk Classes Scheme was created by the Comprehensive Agreement on Special Class Pooling (P9 Agreement, see also paragraph 2.24).52 That Agreement was an agreement between the federal government in Canada and the provincial governments. The powers necessary to create the Special Classes and to administer the Special Milk Classes Scheme had been conferred on the Canadian Dairy Commission, a Crown corporation, by amendment to federal legislation, the Canadian Dairy Commission Act.53
4.10.
The United States emphasized that the powers of the Canadian Dairy Commission, as set forth in Section 9 of the CDC Act (paragraph 2.13)54, were numerous and broad. The Act conferred the authority to the CDC: (i) to purchase any dairy product and sell, or otherwise dispose of, any dairy product purchased by it; (ii) to establish and operate a pool or pools in respect of the marketing of milk and cream; (iii) to establish the price, or minimum or maximum price, paid or to be paid to the Commission, or to producers of milk or cream; (iv) to collect the price paid or to be paid to the Commission, or to any producer in respect of the marketing of any quantity of milk or cream; and (v) to do all acts and things necessary or incidental to the exercise of any of its powers or the carrying out of its functions under the Act. The power to establish a pool, to establish prices, and to collect the price to be paid had all been added in 1995 at the time of the creation of the Special Class system. Furthermore, conspicuous by its absence in the CDC’s enumerated powers was any qualification that those powers were subject to the decision of milk producers.
4.11.
The United States further noted that the Comprehensive Agreement on Special Class Pooling contained several interrelated elements. First, the Agreement provided for the adoption of both the Memorandum of Understanding on Special Class Pooling ("MOU on Special Classes") and also an Addendum to the Memorandum of Understanding. The MOU on Special Classes provided that the provincial governments would enter into a revenue pooling arrangement for milk in the Special Classes. The MOU directed the CDC to determine the percentage of total production by special class utilization in each province. The MOU also contained an agreement to establish and harmonize prices for Special Classes in the CMSMC. Significantly, section 11 of the MOU stated that a province could join the MOU only by the action of the individual provincial governments, not through the action of the marketing boards, or the decision of milk producers, either individually or collectively. Annex B of the Comprehensive Agreement addressed the question of surplus removal and confirmed the role of the CDC in the operation of surplus removal. The Annex provided that there would be both a CDC-initiated surplus removal plan, as well as provision for processor-initiated surplus removal. Paragraph C(1)(iii) of Annex B provided that the CDC would remove the surplus milk by authorizing dairy processors to acquire milk under Special Class 5(e) and manufacture dairy products for purchase by the CDC for export. Sub-paragraph (vii) stipulated that the processor would receive an assured margin and that the level of the margin would be negotiated by the CDC with the processor. No mention was made at all in this Annex for a decision-making role for milk producers in any of these decisions. Nor were producers given such a role in connection with a decision of processor-initiated surplus removal. While the MOU established an advisory group, consisting of an equal number of processors and milk producers, their role was to advise the CDC on when the CDC-initiated surplus removal should be initiated.
4.12.
Canada did not deny that the CDC played an important role in the Canadian dairy system but it was not the central directing agency that the Complainants alleged. The CDC acted as a centre of technical expertise that was available to the dairy industry as a whole and to the producer-dominated decision-makers at the CMSMC. It supplied recommendations and data for the consideration of the CMSMC to assist it in deciding on the annual MSQ. It also calculated a "Support Price"55 used by the producer boards to assist them in negotiating and establishing domestic price levels. Furthermore, the CDC's role as chair of the CMSMC was one of facilitation. The role of the CDC for practical, legal, historical and political reasons, was necessarily that of facilitator and consensus-builder.
4.13.
Canadaargued that although the CDC did indeed act as the chair of the CMSMC, the CDC did not act in the manner suggested by the Complainants: that of dominant director, telling the industry how they were going to carry out federal government policy. Its role was that of facilitator and technical advisor. Ultimately, the CDC implemented the policies and programmes agreed upon in the CMSMC; it neither dictated nor directed them. Canada argued that the CDC, in its role as the administrator of aspects of the dairy system, operated under the direction and control of the CMSMC and was routinely subject to rigorous scrutiny at each meeting of the CMSMC. The CMSMC, unlike the CDC, was not a governmental body. For example, CDC-initiated surplus removal did not begin with a decision of the CDC. It began with a decision of the Advisory Group on the Surplus Removal Program (known as the SRC, paragraph 2.54 and footnote thereto refers), a body composed of producers and processors.56 Only when the industry representatives on the SRC had decided that the domestic market was being satisfactorily supplied, did it instruct the CDC to open such a programme. Once the CDC-initiated surplus removal programme had been opened, the CDC entertained proposals from private exporters with whom it negotiated on behalf of the producers as agent. Canada rejected the US suggestion that Annex B of the Comprehensive Agreement on Special Class Pooling confirmed the role of the CDC in surplus removal (paragraph 4.11). What the Comprehensive Agreement on Special Class Pooling did was empower the continuation of the CDC role at the direction of the CMSMC. The difference was important because the role of the CDC had changed profoundly in 1995. The old Offer to Purchase Program, under which butter was bought on open offer by the CDC and then sometimes exported, had been eliminated. Instead, the CDC was instructed to be guided by the decisions of the SRC in deciding whether surplus removal activity was required. In other words, the empowerment for any activity was now squarely with the industry, not the CDC.
4.14.
Canada emphasized, in respect of the CDC's role as Chair of the CMSMC and its "de facto veto power", as alleged by New Zealand (paragraph 4.5), that the Comprehensive Agreement on Special Class Pooling had now all but superseded the provisions in the NMMP. Pursuant to Article 1 of Schedule I of the Comprehensive Agreement on Special Class Pooling, all decisions relating to matters covered by that Agreement required unanimity, including all matters with respect to export trade. Given the breadth of coverage of the Comprehensive Agreement on Special Class Pooling, little of significance was left without a unanimity requirement. Hence, very little scope had been left for the CDC to take a decision where no consensus was reached in the CMSMC. In addition, Canada noted that the Comprehensive Agreement on Special Class Pooling had a formal dispute settlement process to ensure that no "de facto" CDC veto existed. The Complainants would have been better informed if they had taken note of the clear direction given in the Comprehensive Agreement on Special Class Pooling:

"The Canadian Milk Supply Management Committee (CMSMC) will be the supervisory body which will oversee the implementation of this agreement."57

4.15.
Canada further argued that similarly, the process of negotiation with exporters with respect to the price for Special Class 5(d) and (e) sales was subject to CMSMC control and direction. Even more importantly, the CDC was required under the terms of the Comprehensive Agreement on Special Class Pooling to act in these negotiations as agent. Section 2 of Schedule II of the Agreement stated that the "CDC shall act as agent in carrying out the administrative functions in the operations of the programme".
4.16.
Canada emphasized the importance of the character of the Government's participation as it occurred in practice. Canada argued that to suggest that the mere presence of legislative authority under which an industry operated made that business "governmental" in character was untenable. All businesses operated to some degree within legal and regulatory frameworks established by government to ensure that the public interest was protected. If it were true that a legislative framework made business "governmental" then any regulated industry, including banking or utilities, would be deemed to be "governmental" in nature. It would even be possible to argue that private corporations established under corporation law were "governmental". This would be absurd. Obviously, to establish that business was "governmental" in nature, significantly more governmental intervention had to be shown.
4.17.
Canada argued that once established, producer boards were provided with the authority they required to operate their affairs. This was done through enabling legislation.58 It established the framework for producer board operations and enabled the board to exercise certain functions when and as required. These discretionary functions related to the issuance and administration of quota, the pooling of returns, prices, producer record-keeping and reporting, inspection and the ability to enter into co-operative arrangements with other producer boards and the CDC. The critical point in this regard was that the authority provided to the producer boards was enabling, not mandatory. The producer boards were not directed or obligated by the enabling legislation to carry out certain tasks or functions, as might be the case with mandatory legislation or regulation. The result of establishing producer boards equipped with authority under enabling legislation had to be to allow producers to join together to run their own affairs, subject to a government oversight function to ensure that this authority was used in the public interest. It was absurd to suggest that, where governments had taken steps to enable citizens or industries to govern their own affairs, and withdrawn from or avoided imposing government direction and control, that the resulting self-governing regime was an arm or extension of government.
4.18.
Canada noted that the authority provided to the producer boards was now essentially exercised primarily with respect to the domestic market. Producers had to be licensed to participate in the industry either with respect to milk for domestic or export sales. One of the criteria for such licensing was the holding of a minimal amount of marketing quota, quota that could be purchased in the open market. Producer boards only used their pricing authority with respect to the domestic and general use classes (Classes 1 through 5(c)) and, even in those cases, prices were usually the result of negotiations between the producer boards and the processors. More particularly, the pricing authority was not used with respect to the export classes: Special Classes 5(d) and (e). Prices for sale in these classes were the result of transaction by transaction negotiations between the processor/exporter and the CDC, acting as agent for the producers. In addition, the pooling function was not exercised with respect to over-quota sales under Special Class 5(e).
4.19.
Canada acknowledged that the CDC was involved in the negotiation with exporters of prices to be paid for the milk for export purposes. However, the CDC was acting under the direction of the CMSMC, whose policies were driven by the producer-run marketing boards. Further, in acting as intermediary with the exporter, the CDC was acting as an agent on behalf of producers and in furtherance of their interests. The CDC was expected to obtain the best possible price for the producers based on prevailing returns in the world market. CDC performance in this regard was carefully monitored by the CMSMC. The producers proceeded on the expectation that the CDC would, in its negotiations, obtain the highest possible return for them. If, on review at the regular meeting of the CMSMC, there was some question that the CDC had not maximised producer returns on the export market, the CMSMC could direct the CDC to abandon the market in question or work to obtain better prices.
4.20.
Furthermore, Canada argued that these negotiations were true commercial negotiations. The CDC, acting for the producers, negotiated with competing processors, pressing for the best price for the producers. The exporter sought the lowest price. The competition among exporters coupled with the forces of the international marketplace drove the negotiations. The CDC was in no position to offer, contrary to the interests of the producers, prices lower than those dictated by the world market. Nor could the CDC force an exporter to pay for a particular transaction more than world prices permitted for it to be profitable. The issuance of a "permit" by the CDC was merely a recommendation to the respective board that milk be supplied for a particular proposal. The producers, through their boards, were under no obligation to accept that recommendation. Nor was the permit recipient required to actually carry out the proposed export. The CDC did not control milk supply. Thus, it was evident that the ultimate control over decisions to produce milk for export and to pursue any particular sale rested with the producers. It was equally evident that control of production and sales relating to exports did not rest with government.
4.21.
New Zealand noted that Canada did not seek to deny that the CDC and the provincial milk marketing boards and agencies derived their authority from statute. Instead, Canada argued that reference to the statutory basis of these bodies gave a misleading picture because many of the powers that existed in legislation were not used in practice (paragraph 4.16). But this ignored the fact that the statutory powers of a body gave it a status and character that existed regardless of whether all aspects of those powers were exercised in fact. It did not need to exercise every power that it had on every occasion, or at all, in order to maintain the authority that the government had given it. Its governmental authority derived just as much from its residual as from its exercised authority. Thus, describing legislation as "enabling" did not prove anything. The fact that the federal and provincial legislation in question "enabled" the operation of the Special Milk Classes Scheme was not disputed. The question was whether that scheme entailed sufficient government involvement to meet the relevant definition of an export subsidy. In this respect, enabling legislation was an important first step. It "enabled" the construction of the market for milk into two separate markets. It "enabled" the compelling of producers to ship their milk into one or the other of those markets, and it "enabled" the provision of lower-priced milk to exporters, the subsidy that was the subject of this case.
4.22.
New Zealand argued that the above could be demonstrated by reference to the authority of the CDC to resolve differences in the Canadian Milk Supply Management Committee (the CMSMC) by its own unilateral decision.59 The fact that it might not do so in practice, as Canada alleged, was irrelevant. The fact that it had the authority to make such decisions would have a significant impact on the reaching of consensus within the CMSMC. The process of reaching consensus when one party had the ultimate power of decision was fundamentally different from the process of reaching consensus when no one could individually make that decision. Moreover, in this case, the power of decision rested, not just with a private entity, but with an agency vested with all the authority of government.
4.23.
Furthermore, Canada had amended the legislation establishing the CDC in order to enable it to implement the Special Milk Classes Scheme. Thus, there was no lack of opportunity for legislative amendment in Canada. If the Canadian Government’s argument that the regulation and pricing of milk was not a matter of governmental authority, New Zealand questioned why then it had been necessary to amend both provincial and federal legislation to provide such powers; for example, why had it been necessary to amend the CDC Act to provide that the CDC may "establish the price, or minimum or maximum price, paid or to be paid to the Commission, or to producers of milk or cream... "60
4.24.
New Zealand noted that the process by which milk was accessed for export implicated federal and provincial authorities in an integral way. Exporters had to obtain a permit under Special Class 5(d) or 5(e) from the CDC. That permit was presented to the relevant provincial milk marketing agency in order to obtain milk at the Special Class 5(d) or (e) price. Canada claimed that this permit was "merely a recommendation" (paragraph 4.20). But that was simply a statement about the relationship of the CDC to the provincial marketing agencies, reflecting constitutional authority regarding the setting of prices for milk in a federal system. The exporter had no choice in the matter. The only way to access milk for export in Classes 5(d) and (e) was with a permit. The system was mandatory and the source of that authority was not producer agreement; it was derived from the statutory authority of the CDC and the provincial marketing agencies to compel such behaviour.
4.25.
New Zealand argued that Canada’s assertion that the CDC operated as a collective bargaining agent for producers, and was controlled by those producers acting through their producer-run marketing boards, equally did not withstand analysis. Canada claimed that the CDC’s performance in negotiating prices for exports had to satisfy the boards and the producers or they would require a re‑evaluation of CDC practices through the CMSMC (paragraph 4.15). The impression sought to be conveyed was that the CDC was somehow subservient to producers and not exercising government authority. New Zealand claimed that the opposite was the case. The CMSMC could not control the CDC which ultimately had a veto within the CMSMC, whether or not in practice that veto ever needed to be exercised. Naturally, the CDC would want to maximise returns for producers; that was part of its statutory mandate. And, if producers were unhappy with the CDC, political pressure would be brought to bear on it. Yet this was a normal description of how government agencies functioned. It did not demonstrate that somehow the CDC lost its governmental character through participating in the administration of the Special Milk Classes Scheme. New Zealand noted that Canada downplayed the role of the CDC, arguing that it did not exercise the statutory functions that it had in fact been given. Yet, in New Zealand's view, the actual functions exercised by the CDC were more than sufficient to show the necessary government involvement in the administration and operation of the Special Milk Classes Scheme.
4.26.
New Zealand noted that Canada also sought to argue that New Zealand had wrongly focused on the CDC when it was the CMSMC which was the key decision-maker in respect of special milk classes. The CMSMC, Canada argued, was not a governmental body (paragraph 4.13). Yet the CMSMC was created under the National Milk Marketing Plan (the NMMP)61 and the NMMP was described by Canada as a "contractual agreement", hence the CMSMC "is a contractual body not a creation of government".62 New Zealand agreed that the NMMP was indeed an agreement. Its opening words were "This Plan is a federal-provincial agreement... ". It was entered into by the CDC, a federal Crown Corporation, and the provinces. In some instances the agreement was signed by the Ministers of Agriculture for the province as well as the representative of the provincial milk marketing board; in other instances it was signed by the milk marketing board for the province. The NMMP was an intergovernmental agreement entered into between the representative of the federal government, the CDC, and the authorised representatives of the provinces who varied province by province. As a result, the CMSMC was a creature of intergovernmental agreement. Thus, to characterise it as a "contractual body not a creature of government" was misleading. The preamble to the Plan recognised that "the participation of the Federal and Provincial authorities is required to assure the adoption and implementation of such Plan".63
4.27.
New Zealand noted that although Canada did not deny that the CDC was a Crown agency and was the representative of the Government of Canada, it argued that its role on the CMSMC, which it chaired, was not governmental. Canada said that the CDC’s role was that of "facilitator and technical adviser," and the CDC "implements the policies and programmes agreed upon in the CMSMC; it neither dictates nor directs them". Such a view ignored the ultimate decision-making powers of the CDC whose influence through research and technical expertise, as well as its authority derived from the fact that it was the representative of the government of Canada, could not be hidden by focusing on a collegial decision-making process within CMSMC meetings. The CDC did not need to "dictate" or "direct" the policies of the CMSMC. The fact that at the end of the day it could do so, gave the CDC the central role in the CMSMC. New Zealand noted that in the Bari case, the British Columbia Supreme Court had found that the provincial boards and the CMSMC had valid administrative powers – derived from the CDC Act – "to carry out... the efficient marketing of milk and milk products".64
4.28.
New Zealand noted that Canada denied that the CDC had a veto power within the CMSMC claiming that the Comprehensive Agreement on Special Class Pooling had "all but superseded" the provisions of the NMMP which grant such authority to the CDC (paragraph 4.14). It was notable that Canada carefully qualified this statement about the redundancy of the NMMP. It was particularly important to do so because what Canada sought to rely on to support its position simply did not prove its point. The "veto power", Canada said, had been replaced by a unanimity rule. But that made New Zealand’s point; it did not contradict it. Under a unanimity rule, the CDC would have a "de facto veto power". The Chairman of the CDC, Mr Guy Jacob, when describing the role of the CDC within the CMSMC to the Canadian House of Commons Standing Committee on Agriculture and Agrifood, in March of 199865, said:

"On occasion, the [CDC] Commissioners may take decisions on issues when [the CMSMC] Committee members are not unanimous."

4.29.
New Zealand pointed out that in its Annual Report for 1996/97, the CDC stated that it "is largely responsible for the administration of the National Milk Marketing Plan, [and] the federal/provincial agreement governing industrial milk production and management in Canada... ".66 The reality was that the Special Milk Classes Scheme operated through the combined actions of the CDC, a federal government agency, and provincial milk marketing boards. New Zealand noted that these institutions, acting individually, or collectively within the CMSMC, exercised governmental functions that were essential for the operation of the Special Milk Classes Scheme. They established and administered the quota regime on which the Special Milk Classes Scheme was based. They set prices and determined whether milk was to be sold in domestic or export markets. They prohibited entry by new producers except in accordance with the quota regime. They exercised enforcement authority over both quota holders and those outside the system. In respect of the federal government agency - the CDC - it had specific authority within the framework of the CMSMC when unanimity was not reached. It also issued permits to exporters, which constituted the only way in which access to lower-priced milk could be obtained.
4.30.
New Zealand noted that regardless of the composition of provincial milk marketing boards, these boards exercised governmental functions - functions that had been expressly mandated by government in the boards’ constituent statutes or regulations, or functions that had been delegated to them by the federal government’s agency, the CDC. The CDC both functioned independently as a governmental actor under the Special Milk Classes Scheme and was the source of delegated governmental authority exercised by the provincial milk marketing agencies.
4.31.
The United States stressed that the Special Milk Classes Scheme could not exist without the CDC to oversee its operations. As noted by New Zealand (paragraph 4.23), if this had not been the case, it would have been unnecessary to amend the Canadian Dairy Commission Act to grant specific additional powers to the CDC to supervise establishment of the Special Classes, to empower it to pool revenue from the Special Classes, and to establish the price to be paid.
4.32.
In respect of the producers' involvement in the price-setting of milk, the United States argued that while Canada contended that the MOU on Special Class pooling appointed the CDC to negotiate prices with the processors as agent for the milk producers, Canada had been unable to point to specific language that established this principal-agent relationship. Canada failed to identify any provision that demonstrated that the CDC negotiated prices as agent for the producers. In fact, Section 4 of the CDC Act specifically stated that the CDC was the agent of the Crown for all purposes of the Act. Indeed, the facts suggested to the contrary that the CDC was primarily negotiating a price for inputs for the exporters that would allow them to be competitive in world markets for processed dairy products. Furthermore, the language of the MOU specifically contradicted Canada’s assertion that the producers negotiated the processors’ assured margin on the latter’s export sales. Paragraph C(1)(vii) of Annex B to the MOU directed that this was the role of the CDC, not the milk producers.67
4.33.
The United States argued that an examination of the documents that provided the foundation for the Special Milk Classes Scheme revealed that the Canadian argument that producers retained the ultimate decision-making authority was baseless. The United States had already detailed the scope of the powers that Canada’s Parliament granted to the CDC to operate the Special Milk Classes Scheme, as well as the authority provided to the CDC to delegate some of its newly-created powers to both the provincial governments and the provincial milk marketing boards – powers which had been provided mainly through amendment of the CDC Act and the Dairy Product Marketing Regulations (paragraph 4.10 and following).68 The Memorandum of Understanding establishing the Special Classes was an agreement between the provinces and the CDC, a Canadian Crown corporation. Paragraph 11(a) of Schedule I of the MOU stated that "[i]f a Province wishes to become a party to this agreement, its Provincial Government representative shall send a note of its intent to the CDC" (emphasis added). The MOU did not state that provinces would join the Special Classes Agreement when the milk producers or the provincial milk marketing boards so decided. The MOU very plainly stated that Provinces joined the Agreement by the action of their Provincial Government representative. While milk marketing boards were also signatories to the Agreement, the dispositive fact for each province was whether its Government representative indicated an intent to join. The United States did not dispute that most Provincial governments conferred with their industry representatives to determine whether the industry supported the Special Classes Agreement. However, government/industry consultations and a government’s receptivity to citizen’s input did not alter the essential legal fact that the Comprehensive Agreement on Special Class Pooling was first, and foremost, between the Provincial governments and the CDC.
4.34.
The terms of the National Milk Marketing Plan, which had long been the centre piece of Canada’s dairy supply management, were similar in effect to the Special Class MOU. The opening paragraph of the NMMP stated "[t]his Plan is a federal-provincial agreement in respect of the establishment of a National Milk Marketing Plan... ". The Preamble, in fact, stated that "the participation of the Federal and Provincial authorities was required to assure the adoption and implementation of the Plan." In contrast, the terms of the Plan made it clear that the participation of milk marketing boards was not essential to the working of either the Plan or the CMSMC. Paragraph H(3) of the Plan provided:

"In the event that there are no Signatories of a province which are producer boards, representatives of producer organizations shall be seated as full participants in the deliberations of the Committee, except that they shall not have the right to vote." (Emphasis added.)

4.35.
Moreover, the United States noted that the Governor in Council had delegated some of his powers to the CDC. For example, the Dairy Products Marketing Regulations provided that the CDC shall cause a federal licence to be issued only to a person to whom a share of the portion of the federal quota had been allocated. Moreover, the regulations stipulated that no person should engage in the marketing in inter-provincial or export trade of a dairy product unless the dairy product was, or was made from, milk or cream that was produced by a person who held a federal licence. The CDC, thus, could deny any milk producer that did not possess a federal quota the right to market milk. The regulations were also important in that they provided specifically for the delegation of federal powers to the dairy marketing boards which operated in the various provinces. Thus, when those dairy boards acted with respect to the Special Classes and inter-provincial or export trade, they functioned under powers delegated by the federal government. Hence, the source of the authority for the regulation of the milk marketing in inter-provincial and export trade was federal law and the Canadian Parliament. The entities that had been tasked with performing the functions necessary to implement the legislated regime were also federal and provincial governments, or marketing boards acting with authority delegated from the federal government, not from the Canadian dairy industry. The price the processors paid for milk, both domestically and for export, was determined by the exercise of governmental authority through the operation of the CDC and provincial marketing boards and agencies. The United States noted that Section 3 of Bill C-86, which was the legislation which amended the CDC Act in 199569 gave the CDC the authority, with the approval of the Governor in Council, to enter into agreements with a province or a marketing board to set prices, establish pools, and collect prices to be paid.
4.36.
The United States noted that Canada had stated that although the CDC issued permits to processors to enable them to purchase milk at reduced prices, such "permits" did not actually require that milk be made available to the processors at such discounted prices. Instead, Canada claimed that the permits were only recommendations, and that the milk boards were not compelled to provide milk to the processors according to the terms of the permit. However, there was no question that the processors could not obtain the milk at the lower price without the permits.70 In this sense, Canada minimized two equally, if not more, important practical considerations: (i) the milk producers were essentially price takers; and (ii) processors could not access the lower priced Class 5(d) and (e) milk without a permit. Thus, a CDC issued permit was a condition for receipt by the processor of the Special Class 5(d) and (e) milk, and boards possessed few options but to accept the price that was offered.
4.37.
Canada argued that to the extent that authority in respect of the Special Milk Classes Scheme was conferred on the CDC, this was to assist in the implementation of the decisions taken through the producer-boards and the CMSMC as reflected in the Comprehensive Agreement on Special Class Pooling. Canada stressed that the amendments to the CDC Act were intended to supplement the existing authority of provincial producer boards with the necessary federal enabling authority so that the producer boards could fulfil their tasks effectively. It was not intended by Parliament that the enabling amendments to the Act would result in the CDC directing the producer boards to conduct their functions in a particular way. On the contrary, Section 9.1 of the amended Act contemplated agreements providing for the performance by those very boards of functions otherwise vested in the CDC in respect of inter-provincial and export trade. In the specific case of pricing, Sections 9(1)(g) and 9.1 of the amended Act were together intended to permit the producer boards’ pricing functions in respect of Classes 1 – 4 and 5(a), (b) and (c) to be supplemented with the requisite federal authority, thereby filling the legal gap identified in the Bari II case. Contrary to what was suggested by the Complainants, the CDC did not, as a result of the inclusion of Section 9(1)(g) in the Act, exercise price-setting powers. Pricing for Classes 1 – 4 and Classes 5(a), (b) and (c) was negotiated and decided by the provincial boards either directly or through the CMSMC. Pricing for Classes 5(d) and (e) was negotiated on a contract-by-contract basis, with the producer boards exercising ultimate control over the supply of milk at the negotiated price recommended by the CDC. Class 5(d) and Class 5(e) prices were not fixed through the exercise of regulatory powers, whether under Section 9(1)(g) or otherwise, but were established on a commercial basis based on world market conditions.
4.38.
In respect of the US assertion that Canada had not identified specific language that established a principle-agent relationship in respect of the CDC's role as agent for the milk producers in price negotiations with processors (paragraph 4.32), Canada drew the Panel's attention to Section 2, Schedule II, Addendum of the Comprehensive Agreement on Special Class Pooling, which stated that: "... that the Canadian Diary Commission (CDC) shall act as agent in carrying out administrative functions in the operation of the program" and, also, in Section 1, Schedule I to the same Agreement which stated that the CMSMC "... will be the supervisory body which will oversee the implementation of the [Comprehensive Agreement on Special Class Pooling] agreement."
4.39.
Canada emphasized that the role of the CDC as a technical advisor and consensus builder – rather than as a decision-maker at the CMSMC – was evident in the setting and allocation of the MSQ. In practice, the CMSMC asked the CMSMC Secretariat to develop estimates of market requirements for Canadian industrial milk (paragraph 2.27 and following refer). The CDC presented the estimates developed by the Secretariat to the CMSMC for consideration in advance of each dairy year. The technical estimates prepared by the Secretariat were then actively debated by the CMSMC. Typically, the market requirements initially calculated by the CMSMC Secretariat were revised on the basis of directions received from the CMSMC. Canada stressed that the CMSMC was in no way bound to accept the figures suggested by the Secretariat. Nor did the CMSMC merely rubber stamp these suggested market requirements. Only after the CMSMC was satisfied with the market requirements estimates was a decision made on the MSQ allocations among the various provinces. Canada reiterated that it was the CMSMC, not the CDC, that decided on the MSQ allocation. Such allocation required unanimity.
4.40.
Canada argued that the Complainants had suggested that because many types of discretionary authority had been conferred on the CDC through the CDC Act, this somehow indicated that the CDC carried out all of the things it was empowered to do. This was insupportable both in principle and in fact. First, simply because enabling legislation permitted an entity to carry out certain acts, this did not mean that they would be exercised. The entity could be constrained from acting by other legal obligations, or may simply find that certain actions were not needed in a particular circumstance. Alternatively, it could not be in a political or technical position to undertake the action it was authorized to do. The point was: enabling legislation did not compel any action. In practice, the list of functions actually carried out by the CDC was much more circumscribed than the list of functions in the CDC Act.71 Further, Canada argued that the Complainants' argumentation ignored the fundamental principle in the WTO and the GATT that obligations were only attached to what a Member actually did, not what they could do. For example, an entity could have all the necessary power to provide subsidies in excess of a Member’s obligations, but this was of no consequence unless such subsidies were actually provided.

(c) Government Involvement

4.41.
New Zealand argued that the Special Milk Classes Scheme was a response to the belief that export subsidies, which existed under the producer levy scheme, would no longer be compatible with Canada’s international trading obligations. The fact that Canada had chosen to abandon the producer levy-based subsidies in favour of an alternative scheme made it clear that the reduction and eventual elimination of the incentives provided by those subsidies was not an option Canada was prepared to follow. The Special Milk Classes Scheme was not, as Canada maintained, a response to the fact that as a result of the new WTO Agreements Canada no longer had to limit production of milk. New Zealand argued that the Special Milk Classes Scheme was a substitute for the old producer levy-based subsidy. Moreover, it was a federal government agency, the CDC, that had spearheaded the process that led to the development of the Special Milk Classes Scheme.
4.42.
New Zealand noted that the question of the degree of governmental involvement necessary for measures to be regarded as government measures had arisen on several occasions under both the GATT 1947 and the WTO. The 1960 GATT Panel studying the obligation on states to notify subsidies financed by a non‑governmental levy under GATT Article XVI, spoke of schemes "which are dependent for their enforcement on some form of government action."72 The Panel on Japan - Photographic Film emphasised the fact that where non‑binding action of government "creates incentives or disincentives largely dependent on governmental action for private parties to act in a particular manner, it may be considered a governmental measure."73
4.43.
New Zealand contended that in the present case, an obvious parallel could be drawn with the decision of the Panel in EEC - Restrictions on Imports of Dessert Apples.74 In deciding whether the EEC regime relating to the marketing of apples constituted a governmental measure within the meaning of GATT Article XI:2(c)(i), the Panel noted that:

" … the EEC internal régime for apples was a hybrid one, which combined elements of public and private responsibility. Legally there were two possible systems, direct buying-in of apples by Member State authorities and withdrawal by producer groups. Under the system of withdrawals by producer groups, which was the EEC’s preferred option, the operational involvement by public authorities was indirect. However, the régime as a whole was established by Community regulations which set out its structure. Its operation depended on Community decisions fixing prices, and on public financing; apples withdrawn were disposed of in ways prescribed by regulation. The Panel therefore found that both the buying-in and withdrawal systems established for apples under EEC Regulation 1035/72 (as amended) could be considered to be governmental measures for the purposes of Article XI:2(c)(i)."75

4.44.
Hence, New Zealand noted that in such a situation, of combined public and private responsibility, the Panel had considered the EEC regime to be governmental in nature - even although the involvement by public authorities was indirect.
4.45.
New Zealand contended that the Special Milk Classes Scheme had been initiated with direct government involvement. The CDC, a federal Crown corporation, had identified the need for changes to the programmes it offered as early as 1992.76 It was a participant in the industry Consultation Committee which concluded that export subsidy reduction commitments would "render the use of levies ineffective".77 A federal-provincial task force was established to review the matter. The CDC chaired the "Dairy Industry Strategic Planning Committee" that recommended a classified pricing system for milk based on end-use and a national pooling system. A negotiating sub-committee of the CMSMC brought those recommendations to federal and provincial Ministers of Agriculture in December 1994. This was the genesis of the government-initiated Special Milk Classes Scheme.
4.46.
New Zealand contended that the Special Milk Classes Scheme was implemented through government action. It was embodied in a federal-provincial agreement, the Comprehensive Agreement on Special Class Pooling. The CDC ACT was amended to allow the CDC to administer the Special Milk Class permit system and the pooling arrangements. Under the Comprehensive Agreement on Special Class Pooling the CDC was to "act as agent [of the federal Government] in carrying out administrative functions in the operation of the programme" (Schedule II).
4.47.
New Zealand stressed that the scheme required continued government involvement for its operation and enforcement. In order to be effective and to provide the appropriate incentives, the scheme had to be mandatory and new entrants prohibited. This was done through the exercise of statutory authority. Government involvement was therefore essential to the scheme’s existence. The mandatory character of pooling, the administrative functions of the CDC and of the provincial milk marketing boards or agencies in the operation of "special milk class" access, pricing and pooling, were all activities of government. Indeed, in Canada - Import Restrictions on Ice Cream and Yoghurt78, Canada itself had claimed that its milk supply management system constituted governmental measures within the meaning of GATT Article XI:2(c)(i).
4.48.
New Zealand claimed that the centrality of governmental involvement in the Special Milk Classes Scheme was readily apparent when the question was asked whether the scheme could continue to operate if the governmental presence were removed. There would be no legislative basis for the operation of the scheme. There would be no NMMP, there would be no Comprehensive Agreement on Special Class Pooling. In the absence of government authority, there would be no mechanism to set prices or to compel compliance except through the agreement of the members of the "producers’ club". Nothing could prevent those outside the "club" from marketing milk domestically. There would be no government agency, no CDC, to chair the CMSMC and resolve differences where unanimity could not be reached, and there would be no delegated governmental powers residing in provincial marketing boards. Furthermore, to the extent that some agency was needed to administer a permit system under which access to "special class" milk was provided, it would have to be a private agency established by the producer members themselves. New Zealand argued that the Special Milk Classes Scheme would simply not function if everything was left to private producer agreement. Indeed, any attempt at price setting by private producer agreement would raise questions about compliance with competition laws. It did not do so under Canada’s supply management system because of the very involvement of government in the scheme. Hence, government involvement was critical to the functioning of all aspects of the Special Milk Classes Scheme. Canada’s argument that the scheme operated through the private activity of producers with only a government oversight function simply was not credible.
4.49.
New Zealand recalled that Canada had not argued that there had been any real change to the role of government after the introduction of special milk classes. At no point had Canada sought to suggest that special milk classes heralded a shift in the extent of federal and provincial governments’ involvement in the dairy marketing system. Instead, the picture Canada painted of government activity within the dairy supply management system was one of continuity. To Canada, it had remained a producer-driven and dominated dairy marketing system (paragraph 4.62). New Zealand argued that, if this were the case, there would not have been sufficient government involvement within Canadian terms under the old producer levy scheme to meet the requirement of government involvement in order to constitute an export subsidy. It would simply have been a system whereby producers within their own organisations - milk marketing agencies and the CMSMC - decided to levy themselves in order to support exports, and government involvement would simply have been to act as the designated agent of the producers to assist in implementing the system. Although this was the consequence of Canada’s position, even Canada had admitted that its old producer levy system would have constituted a subsidy under Article 9.1(c) of the Agreement on Agriculture.79 In New Zealand's view, such an admission undermined Canada’s portrayal of government involvement in the Special Milk Classes Scheme.
4.50.
New Zealand noted that Canada claimed that the government role was one of oversight only and provincial and federal governments simply provided a framework for the activities of producer-run marketing boards; within the CMSMC, the government's role was to ensure that the system was operated in accordance with the general public interest (paragraph 4.16). Canada sought to equate the role of government in the operation of the Special Milk Classes Scheme, and in dairy supply management more broadly, to that of its role in society generally - to act in the public interest. However, the role of government in the Special Milk Classes Scheme was much more intrusive than the exercise of its general function of oversight in the public interest. Without the active participation of government in the administration and operation of the scheme, including its residual enforcement authority, the system could not work. Instead of describing the nature of the government involvement, Canada had gone to the other extreme and sought to have the government disappear altogether from the Special Milk Classes Scheme.
4.51.
New Zealand argued that in the case of non-mandatory measures, the determining factor in deciding whether conduct could be ascribed as resulting from governmental action had been whether there were sufficient incentives or disincentives for the measures to take effect.80 Most recently, a WTO Panel had noted that "the fact that an action is taken by private parties does not rule out the possibility that it may be deemed to be governmental if there is sufficient government involvement with it".81 The Panel had gone on to note that it was difficult to establish definitive rules, and that a case-by-case analysis was required.
4.52.
New Zealand maintained that under any of these tests the operation of the Special Milk Classes Scheme was a government activity. The scheme derived from the agreement of agencies of the federal and provincial governments created by statute. These agencies administered a supply management scheme that could function only through the interplay of the exercise of federal and provincial authority. The Bari litigation demonstrated that provincial authority alone was not sufficient to give effect to a quota regime affecting inter-provincial and export trade. There had to be the joint action of federal and provincial governments to enable the system to function.
4.53.
The Special Milk Classes Scheme, as an aspect of Canada’s supply management system, was compulsory: essentially, the only way that milk could be sold on the export market was through special milk classes. Producers did not have the option of selling on the export market independently of a government-mandated scheme. The subsidy that exporters received under the Special Milk Classes Scheme was provided through the cooperative activity of the CDC and the provincial milk marketing agencies. Exporters had to obtain a permit from the CDC and obtain milk through the provincial milk marketing agency. The scheme was thus government-mandated, maintained through the actions of government agencies, and enforced by government authority.
4.54.
The United States also agreed that milk would not be available to processors of dairy products for export at the indicated prices absent the structure of the dairy regime established by the Canadian federal and provincial governments and which was administered and enforced by those governments. The pervasiveness of the government’s role in the Special Milk Classes Scheme was evident from consideration of the legislatively granted authority that those entities possessed and exercised (paragraph 4.4 and following). Despite Canada’s claims to the contrary, government action and authority was not transformed into private action simply because private parties, in this case dairy farmer organizations, could approve in specific instances of the actions taken by their government. If this were the case, any action by a government to benefit a portion of its citizenry would be converted into private action. For example, most anti-dumping measures would be "private actions" by definition and the WTO Agreement on Anti-Dumping would be nullified.
4.55.
The United States argued that it was sufficient to look to the marketing boards’ own statements about the source of their powers to lay to rest Canada’s contention that the boards receive their power from the dairy farmers. For example, the British Columbia Milk Marketing Board’s Consolidated Order of 1 August 1997, described both its purpose and the basis for the Board’s authority. The stated purpose of the Order prominently referenced both the provincial and federal authority that permitted the Board to act.82 Also, a regulation issued by the Ontario Milk Marketing Board in June 1995 had a similar effect with respect to the powers delegated to it by the federal government as the authority for its control over the marketing of milk produced in the Province of Ontario.83
4.56.
The United States argued that both the manner of creation of the Special Classes and the actions by the provincial and federal governments following the Bari II litigation also refuted Canada’s allegations that the Special Milk Classes Scheme represented an agreement between private parties that was simply pursued within an overall legislative and regulatory framework that was government created.84 Indeed, if the Special Milk Classes Scheme were simply an agreement between the various producer dominated provincial marketing boards, the United States questioned: (i) why any government involvement was required; (ii) why did the Comprehensive Agreement state that it applied only to those provinces whose provincial governments had approved it85; (iii) why was it necessary for the Canadian Parliament to amend the CDC Act to provide specific powers to the CDC to operate the Special Classes; and (iv) why such powers could not simply be conferred by the dairy marketing boards in their capacity as representatives of the dairy farmers. Canada had not, in the US view, provided responses to any of these questions. Its only answer was that the provincial marketing boards (which performed most, if not all, of their relevant responsibilities by virtue of powers delegated to them by both the federal and provincial governments) could not remain in office if they did not satisfy the desires of their dairy farmer constituency (paragraph 4.19). The United States argued that if this were the test for determining whether action was governmental or not, any action by a popularly-elected government would be deemed not to be governmental action. Moreover, every time a government agency or legislature took action which benefited a class of individuals, that action would no longer be considered to be governmental in character.
4.57.
The United States noted that a body of legal authority had developed in the GATT and WTO that was relevant to the issues before this Panel. Several GATT and WTO panels had considered, primarily in the context of Article XI of the GATT, whether actions by a government that did not impose specific requirements on private parties were, nonetheless, government measures. While this analysis had necessarily to be conducted on a case-by-case basis, the consistent conclusion that each Panel had reached was that action need not be mandated by a government to constitute enforcement of a government measure. This issue was first addressed in Japan - Restrictions on Imports of Certain Agricultural Products (hereafter "Japan – Certain Agricultural Products").86 There the Panel had wrestled with the question of whether the Japanese system relating to restrictions on domestic production provided for "enforcement of government measures". The Panel found that the restrictions emanated from the government and that "administrative guidance" from the Government of Japan played an important role in the enforcement of those measures.87 This principle was taken a step further by the Panel in Japan - Semiconductors.88 In that dispute, the Panel found that "an administrative structure had been created by the Government of Japan which operated to place maximum possible pressure on the private sector to cease exporting at prices below company-specific costs."89 The Panel concluded that despite the absence of any legally binding obligation, the complex of measures that Japan had adopted operated in a manner equivalent to mandatory requirements.90
4.58.
Also, the United States claimed that the analysis of the panel which examined Japan - Photographic Film91 demonstrated exactly how schemes such as the Special Milk Classes Scheme fitted within the body of WTO law. The Photographic Film panel addressed the related issues of whether certain governmental actions were "measures" for the purposes of the non-violation nullification or impairment remedy under GATT Article XXIII:1(b) and were "laws, regulations or requirements" for the purposes of GATT Article III. Agricultural trade measures of a hybrid nature had been the subject of GATT panel findings. As the Photographic Film panel observed, "a 1989 panel on EEC - Restrictions on Imports of Dessert Apples noted that ‘the EEC internal regime for apples was a hybrid one, which combined elements of public and private responsibility. Legally there were two possible systems, direct buying-in of apples by Member State authorities and withdrawals by producer groups’. That panel found that both the buying-in and withdrawal systems established for apples under the EEC regulation could be considered to be governmental measures for the purposes of Article XI:2(c)(i)."92 The rule formulated by the Photographic Film panel was that

"... the fact that an action is taken by private parties does not rule out the possibility that it may be deemed to be governmental if there is sufficient government involvement with it. It is difficult to establish bright-line rules in this regard, however. Thus, that possibility will need to be examined on a case-by-case basis."93

4.59.
The United States noted that the Photographic Film panel also examined the Fair Trade Promotion Council’s 1984 Self-Regulating Standards. Whereas Japan argued that this Council was a mere private entity, the panel had noted the extensive links between the Council and the Japanese government – "dependence of the Fair Trade Promotion Council on liaison with the JFTC for the establishment of these standards" – and found that these standards were attributable to the Japanese government.94 The Photographic Film panel examined a Retailers Fair Competition Code and its enforcement body, the Retailers Fair Trade Council. Japan argued that this Code was only self-regulation among business entities, and the Council was a voluntary organ to implement this self-regulation. The panel rejected Japan’s position:

"... Viewed in the context of the JFTC having approved the Fair Competition Code and the Retailers Council, and of Article 10(5) appearing to give a governmental exemption from certain provisions of the Antimonopoly Law to actions by the Retailers Council and code members under the code, it is difficult to conclude that investigation, enforcement and governmental liaison actions of the Retailers Council under the code are purely private actions of a private trade association.... we note that a finding to the contrary would create a risk that WTO obligations could be evaded through a Member's delegation of quasi-governmental authority to private bodies. In respect of obligations concerning state trading, the provisions of GATT explicitly recognize this possibility. In this regard, an interpretative note to Articles XI, XII, XIII, XIV and XVIII states: "Throughout Articles XI, XII, XIII, XIV and XVIII, the terms "import restrictions" or "export restrictions" include restrictions made effective through state-trading operations". The existence of this note demonstrates that the drafters of the General Agreement recognized a need to address explicitly one aspect of the government-delegation-of-authority problem. In our view, it supports our finding that measure for purposes of Article XXIII:1(b) should be interpreted so as to prevent actions by entities with governmental-like powers from nullifying or impairing expected benefits."95

4.60.
The United States argued that the same dangers that the Photographic Film Panel found, also existed in the context of Canada’s Special Milk Classes Scheme. Canada was giving disproportionate weight to the involvement of private dairy farmers in the operation of the marketing boards, and would minimize the greater importance of the boards’ dependence on powers delegated by the federal and provincial governments in Canada. The United States argued that if the current Panel found that the Special Milk Classes Scheme was outside the WTO Agreements simply because it incorporated some private elements into an essentially government scheme, other countries, led by the United States, would be impelled to similarly rearrange their affairs. The economics of dairy trade provided overwhelming pressure to imitate Canada’s regime if it was determined to be WTO consistent, a result which the United States believed would be entirely unjustified.
4.61.
The United States emphasized that the Special Class prices for exported milk were determined by the CDC or the provincial marketing boards. This was essentially achieved by the CDC negotiating an assured margin for processors, which was then subtracted with other costs to provide a net return to the milk producers. In cases of exports by the CDC, this was the end of the matter. Where the CDC reached agreement with a processor on a price to be paid to the producers, Canada insisted that the marketing boards then could determine whether to accept the price obtained. Nonetheless, Canada admitted that the boards rarely failed to accept the price. But more importantly, when the boards accepted that price, they were exercising the governmental powers that had been delegated to them. Their actions, therefore, were no less governmental than those of the CDC. It was not an exaggeration that the boards were essentially extensions of the executive branch of the Government of Canada for most purposes relating to regulation of milk marketing and the Special Classes in particular.96
4.62.
Canada argued that governments did play a role, in that they had taken the necessary steps to provide enabling authority to the producers and their organisations to ensure that the system could fulfill its supply management objectives while retaining an oversight function to ensure that such enabling authority was not misused and that the public interest was protected. Subject to this oversight function, governments in Canada had devolved discretionary authority to the dairy industry so it could run its own affairs. This function was diametrically opposed to the fanciful image suggested by the Complainants of coercive government control and direction.
4.63.
Canada refuted the US argument that producer boards were essentially extensions of the executive branch of its government (paragraph 4.61). The executive branch of the Government of Canada consisted of officials and departments directed by ministers of the Crown and the principal executive body, the Cabinet, headed by the Prime Minister. Producer-run and producer-controlled provincial marketing boards could in no way be equated with the executive branch of any government, merely because they carried out certain activities pursuant to enabling legislation and were subject to government oversight. Canada rebutted as equally ill-founded the characterization by New Zealand of the producer boards as "government agencies". The hallmark of a government "agency" (referred to in Canadian, and New Zealand, legal parlance as a "Crown agency") was a "body which was subject at every turn in executing its powers to the control of the Crown".97 The producer boards in the Canadian dairy sector had a vastly greater degree of independence, private accountability and discretion than "government agencies". Hence, far from being government agencies, the producer boards had the character of private agents representing dairy producers. The boards were collective agents for producers as a group. Producers could revoke the agency at any time. However, as was characteristic of a collective agency role performed by trade unions, revocation decisions were made on a collective not an individual basis. Canada argued that the activities carried out by the producer-run boards (the private bodies in question) were necessary for the proper operations of their affairs. The authority provided to them by governments, fell short of the test of being "governmental" in character.98
4.64.
Canada further refuted the US claim that the recent Panel Report in Japan – Photographic Film provided support for the proposition that the Canadian dairy export measures at issue were to be treated as governmental in character (paragraph 4.58). Canada argued that that panel report did not provide a rule with respect to determining whether or not a measure is governmental:

"These past GATT cases demonstrate the fact that an action taken by private parties does not rule out the possibility that it may be deemed to be governmental if there is sufficient government involvement with it. It is difficult to establish bright line rules in this regard, however. Thus, that possibility will need to be examined on a case by case basis."99 (emphasis added)

4.65.
Canada argued that an examination of GATT cases showed, if anything, that the opposite conclusion than that argued by the United States could be drawn. The Japan – Semi-Conductors case,100 involved a situation where the Government of Japan had initiated a policy of its own volition and then sought to impose it on the industry. In the words of the panel in Japan – Photographic Film, this policy was "operated to exert maximum possible pressure on the private sector".101 Thus, this was a clear case of top-down direction from government, through the private sector. By contrast, it was clear that in the present case there was no policy being imposed by governments. In particular, there was no government imposed policy whatsoever on export methods or levels. This was left to the producers and producer boards to determine. Involvement by government had been limited to oversight. The Complainants had referred to the EEC – Dessert Apples case (paragraphs 4.43 and 4.58) suggesting that the facts in that case were relevant to the current matter. As in the case of Japan – Semi-Conductors, Canada pointed out that the fact was that the measures in question in EEC – Dessert Apples represented top-down government direction in a mixed government/private sector environment. In particular, the Panel in EEC – Dessert Apples noted:

"The regime as a whole was established by Community regulations which set out its structure. Its operation depended on Community decisions fixing prices, and on public financing; apples withdrawn were disposed of in ways prescribed by regulation."102

4.66.
Canada argued that the distinction between the EEC – Dessert Apples situation, with direct action by government to implement the policy, and the present case where government provided enabling discretionary authority without policy direction was obvious. In sum, the theme through all these cases was one of governments participating in a top-down, policy-directing and initiating role. The Complainants had failed to show any evidence of governments in Canada setting and dictating policies with respect to the operation of the Canadian dairy system and, in particular, the Special Class export practices at issue. Accordingly, this was suggestive that, using the case-by-case approach, the current case fell outside of the type of situation that could be considered to be governmental.
4.67.
Canada further refuted the significance attributed by both New Zealand and the United States to the Bari case. The Bari cases involved a group of non-licensed British Columbia producers and a processor who were custom processing milk for those producers for marketing in inter-provincial trade. The litigation arose well before the Special Milk Classes Scheme was introduced and, more importantly, had nothing whatsoever to do with export trade in milk or milk products. In response to constitutional gaps in the domestic milk marketing regime identified at an earlier stage in the litigation, regulations were established in 1994 pursuant to the CDC Act. The Bari III case referred to by New Zealand addressed the validity and applicability of the Regulations in a purely domestic context. The Bari litigation had little relevance to the export of products made from milk sold under Special Classes 5(d) and (e).
4.68.
Canada also refuted the Complainant's assertion that milk would not be available to producers of dairy products for export at the indicated prices absent the structure of the dairy regime established by the Canadian federal and provincial governments (paragraphs 4.48 and 4.54). Canada argued that the reality was that export milk was sold to processors at prices negotiated in arms-length transactions directly responsive to world market conditions. Canada argued that even if the CDC were to cease to perform the negotiating activities on behalf of producers that it currently performed, the realities of the world market which drove those export prices would still be the same. Indeed, Canada argued that if the CDC were to exit the export pricing negotiations altogether and producers were to negotiate export sales through their producer boards, processors might easily succeed in negotiating a slightly lower price with nine producer boards having less experience than the CDC in world markets. In this context, to assert the existence of a "benefit" to Canadian processors flowing from the negotiating role performed by the CDC was not only counter-intuitive, it was nonsensical.
4.69.
Canada cautioned the Panel in respect of New Zealand's arguments relating to the 1960 Working Party Report regarding the notification of export subsidies under Article XVI of GATT 1947 (paragraph 4.41). That Working Party Report occurred in a context in which there was no consensus among GATT Contracting Parties as to what constituted a subsidy. The entire focus of the report was to decide what measures were to be notified. Accordingly, the report advanced the Contracting Parties trade policy objectives by requiring notification of certain matters so that the Contracting Parties could assess the resulting trade impacts. Given that there now existed binding obligations with respect to subsidies, the Panel had to be careful in using that Report for the purpose of reading context into the definition of "subsidy".
4.70.
The United States noted that Canada argued that the Bari case was irrelevant because it allegedly did not address the issue of exports. This was a remarkable statement by Canada in light of the fact that the principal issue was the authority of the provincial boards over export and inter-provincial trade. In addition, the Canadian court specifically addressed the question of the authority to impose a levy on production. That levy, of course, was used to fund exports and to allow exports of dairy products to be competitive on world markets. Those issues appeared to the United States to be relevant to exports despite Canada’s protestations to the contrary.
4.71.
In respect of the 1960 Panel Report, the United States noted that Canada's argument that the 1960 Panel Report regarding notification of subsidies should not be given weight because it was issued at a time when the current consensus on those subsidies under the current Agreements had not been concluded totally ignored the fact that the view of producer-financed subsidies had not changed in over 30 years since the Report was issued. In fact, relevant language to this dispute first appeared in that Report and now had been incorporated into the SCM Agreement. Those considerations argued for greater not less weight for the 1960 Report.

(d) Producers' Involvement

4.72.
New Zealand noted that the Canadian arguments focused almost exclusively on the role of producers within the system in order to distract the Panel from the fact that it was exporters, and not producers, who were being subsidized. New Zealand recognized that producers operating individually and collectively were involved in the export regime for Canadian milk products. The fact that a body was composed of producer members did not alter the character it had been granted through its statutory mandate (paragraph 4.21). The power "to regulate the marketing of milk in inter-provincial and export trade" - a power possessed, for example, by the British Columbia Milk Marketing Board103 - did not cease to be the exercise of a governmental regulatory function in respect of export trade because the Board happens to be composed of producers. New Zealand argued that contrary to the impression conveyed by the Canadian arguments, provincial milk marketing boards were not producer clubs. They had regulatory functions, and their authority to regulate was derived from statute, not from the agreement of producers. Milk producers had to comply with the decisions of the provincial boards. They had no right to produce milk and market it except in accordance with the systems established through the combined actions of the CDC and the provincial marketing boards and agencies. Access to the system could only be gained through the purchase of existing quota.
4.73.
New Zealand noted that Canada had also sought to portray the production of milk for export as the result of the individual decisions of producers relying on their own business sense (as argued by Canada in paragraph 4,109). Yet the power of individual producers was largely limited to deciding upon their level of milk production. The revenue they received for that milk did not depend on market forces. The revenue a producer received depended on whether the milk produced was classified as in-quota or over-quota, and whether it was exported or sold on the domestic market. And those decisions were not made by the individual producer. They were made through the interaction of the CDC and the provincial milk marketing boards and agencies, under the auspices of the CMSMC (further argued in paragraph 4.93 and following).
4.74.
New Zealand argued that Canada's description of the operation of the Special Milk Classes Scheme as one that was "developed on a bottom-upbasis by Canadian producers" where Governments simply implemented what they were directed to do by producers was an implausible explanation of how governments operated. In any event, it did not prevent the conclusion that there was an export subsidy.
4.75.
The United States noted that Canada argued that the Special Milk Classes Scheme gave milk marketing in Canada a market-orientation that it had lacked under the producer levies that were eliminated during 1995. In particular, Canada contended that whereas the old system was dependent largely on supply management, the new Special Milk Classes Scheme allowed Canada’s milk producers to price to the various markets available for their products. At the core of Canada’s argument was its assertion that the over-quota levy that existed under the pre-1995 system served as a disincentive to production above quota limits.104 The United Statesargued that since the Special Class prices were set at approximations of the world market prices, there appeared to be comparatively little difference between the economic treatment of over-quota production under the levy system as contrasted with the present Special Class 5(e), which was applicable to over-quota production. Canada’s assertion that the change to the Special Milk Classes Scheme heralded a new day of producer independence was, in the US view, unfounded.
4.76.
The United States stressed that producers could not on their own achieve the national coordination of prices and production that was essential to the operation of the Special Milk Classes Scheme. Canada had, in their view, exaggerated the legal basis for the producers’ role in the system and particularly their role in decision-making. The United States claimed that Canada ignored two important factors. The part played by producers was permitted by: (i) the delegation of government powers to producer marketing boards; and (ii) the provincial governments’ designation of producers as representatives to the CMSMC. In both instances the producers’ participation was at the discretion of the governments involved.
4.77.
The United States argued that the omission of several critical facts distorted Canada’s portrayal of the scope of the legal authority possessed by the milk marketing boards in their role as designated representatives for the Provinces on the CMSMC. While Canada contended that most of the voting representatives participating in the Canadian Milk Supply Management Committee were from producer marketing boards, Canada neglected to give any weight to the most critical fact. The voting representative from each province was selected by the provincial government, the provincial government commission overseeing the industry, and the provincial marketing board.105 Thus, any producer who was a representative of a province at the CMSMC did so by virtue of the decision of three separate entities, two of which were entirely governmental. Therefore, any producer representative designated to sit at the CMSMC table did so at the discretion of the provincial governments. Even were this not the case, the fact that producers sat on the CMSMC, a policy making committee, again did not alter the fact that the CMSMC itself was a creation of a federal-provincial government agreement. Canada, which contended that the CMSMC was run by milk producers and was the decision-maker for dairy policy, also failed to explain why Canada’s Attorney-General in the Bari litigation described the CMSMC as "a federal body, a federal functionary...

"In discharging its duties under the Federal Regulations, the Committee acts as a federal body, a federal functionary concerned solely with matters related to the federal quota.

This exercise of co-operative federalism, suggested by many judicial decisions as the only practical and effective way to regulate in Canada the marketing of agricultural products under divided jurisdiction, is at the heart of the Governor in Council’s recognition of the Committee. The Committee’s composition includes, inter alia, representatives appointed by provincial signatories who are bodies established under provincial law to exercise statutory powers in relation to the marketing of dairy products in intraprovincial trade." (Emphasis added.)106

4.78.
The United States argued that while Canada’s portrayal of the provincial milk marketing boards admitted that federal and provincial delegation of powers were essential to their functioning, Canada failed to acknowledge the binding decision-making powers of the provincial government authorities respecting the boards’ operations. The CDC’s website description of the operation of dairy management in each of the provinces was enlightening for this purpose.107 In connection with the province of Nova Scotia, the CDC’s internet site stated: "The Nova Scotia Dairy Commission is a government agency which controls the province’s marketing." In Ontario, according to the same report, "[t]he Farm Products Marketing Commission, a branch of the Ontario Ministry of Agriculture and Food, acts as a supervisory board for the industry." With respect to Quebec, the website stated that: "In all cases of unresolved dispute, the Regie des marches agricoles et alimentaires - a government organization - has the authority to intervene, and will act as a tribunal and hand down a final and binding decision." In New Brunswick, the same report explained that while the milk marketing board possessed the main responsibility for milk marketing, the Farm Products Marketing Commission, a government agency, which administered the Farm Products Marketing Act, oversaw the activities of the milk marketing board. As Canada already admitted that the milk marketing boards in the provinces of Alberta and Saskatchewan were government operated, there was no need to examine them further. Thus, in at least six of the provinces, accounting for an overwhelming majority of milk production, the provincial governments retained ultimate authority respecting the operation of the milk marketing boards. Although the milk producers could choose who sat on the board, it was the provincial and federal governments that determined the respective board’s authority, and it was the provincial government that had the final say on the operation of the boards.
4.79.
Hence, the United States argued, regardless of the fact that producers had a role in the Special Milk Classes Scheme, which was not denied by United States, the fundamental authority, practices, and operations necessary for the national application of the Special Milk Classes Scheme were governmental. That the governments at both the federal and provincial levels had entrusted certain powers to the milk marketing boards, and thus indirectly to producers, did not alter the fact that the powers exercised by those boards were governmental in origin and that the boards as institutions were ultimately answerable to the overseeing government authorities.
4.80.
The United States argued that the Special Milk Classes Scheme, as outlined in its own and New Zealand's arguments above, was created under the authority of the federal and provincial governments, and was administered, maintained, and enforced by them. Now that it existed, milk producers did not have a choice in its application. Milk producers could not market the milk that they produced without a government assigned quota and a CDC issued license. Milk producers also did not have any input into whether their milk was sold into one of the Special Classes or another. They could not opt not to participate in the Special Class pool.
4.81.
In respect of price, the United States stressed that the dairy farmer had no input in the pricing of milk for over-quota production. Canada stated that all over-quota production received the 3 month rolling-average Special Class 5(e) price. The Special Class 5(e) price was approximately 50 per cent of the price obtained for the same milk components in the domestic market in Canada. The milk producer had absolutely no ability to sell over-quota milk at the domestic price levels. The Special Milk Classes Scheme closed this alternative to the milk producer. Although Canada asserted that the CDC, a Crown corporation, only negotiated the Special Class 5(e) price for the farmers and their boards, Canada had conceded that the price was rarely, if ever, rejected by either. Yet, for example, the British Columbia Consolidated Milk Marketing Order provided in Part VIII, paragraph 31, that "[t]he Board will determine the minimum Producer price for over quota production based upon the calculated world price published by the Commission."108 It was noteworthy that this regulation did not say that the Board could base the Producer price on the Commission published price, but that it will base the price on the CDC published price. Similarly, the Manitoba Milk Producers newsletter published monthly the CDC Special Class price for over-quota production and indicated that this was the price that would appear on producers’ monthly pool statements.109 Again, this left little question but that as a practical matter, the CDC calculated price was the only price available for over-quota milk.110 In this sense, the United States argued that it would be more accurate to state that the CDC was negotiating a milk price for the dairy exporter, rather than for the dairy farmers. This was because a dairy processor with an opportunity for an export sale approached the CDC with the price that it could obtain in the international market for the dairy product export, such as butter, or cheese. The CDC then negotiated with the processor/exporter an "assured margin" which incorporated an

amount for profit as well as the processor’s costs.111 This amount was then subtracted from the processor’s selling price for butter, cheese, or some other dairy export in world markets. The international sales price for the dairy export, less the processor’s "assured margin," was the amount paid to the dairy farmer for his or her milk.

4.82.
Thus, the entire calculation procedure was aimed at assuring a price to the processor/exporter that allowed the export to take place at world market prices. While presumably the interests of the dairy producers were considered, the over-riding objective was to ensure that surplus dairy products were exported. The major benefit derived by the milk producers from such exports was that they assisted in administering the supply management of dairy products and producers got some return for this surplus production. Without such exports, the surplus would be destroyed with obvious political ramifications.
4.83.
Hence, the United States submitted that Canadian exporters of dairy products were provided with milk priced at the lower Special Class 5(d) and (e) levels only as a result of the Special Milk Classes Scheme, that had been put into place by the joint action of the federal Government of Canada and the provincial governments. Without the government action that required the removal of surplus in-quota and over-quota production at prices negotiated or set by the CDC, those producers of dairy product exports would not receive the lower priced milk necessary for them to compete in international markets. It was not normal commercial practice for a government entity to customize the price of an input to match particular sales opportunities.
4.84.
Canada maintained that the creation of the Special Milk Classes Scheme was a producer-initiated process and that contrary to the assertions of the Complainants, the Special Milk Classes Scheme was not imposed by governments. It was developed on a bottom-up basis by Canadian producers through their producer organizations and the producer-dominated CMSMC. Through their decisions in these bodies, the producers had reached agreement on the principles that would form the foundations of such a system. In those resolutions, they called on governments to take the necessary steps to provide the enabling authority that would allow such a system to come into being.
4.85.
Canada noted that as early as 1993, proposals were being circulated for a revision of the Canadian supply management system to include new features such as possible pooling systems and an optional export programme. The purpose of these proposals had been to respond to perceived new market conditions and to introduce added flexibility and competitiveness so as to take advantage of the new export opportunities that would emerge from the conclusion of the Uruguay Round. In Quebec, for example, an early initiative in this direction was taken at the 14-15 April 1993 meeting of the General Assembly of Fédération des producteurs de lait du Québec to begin the process of renegotiating the terms of the NMMP. Further decisions in the General Assembly in 1994 and 1995 advanced this process, particularly with respect to regional pooling arrangements. These decisions fundamentally modified the basis of the Quebec dairy marketing plan and the conditions that applied to the marketing negotiations conducted between producers and processors in Quebec. Parallel developments were taking place within the industry in other provinces, developments that ultimately culminated in the decisions taken at the CMSMC to propose the Special Milk Classes Scheme. Thus, Canada argued that governments in Canada had not imposed arrangements on the dairy producers and processors. Governments had responded to the initiatives of the industry by providing the discretionary authority required to implement industry proposals, provided those proposals were in the public interest. On the other hand, had the processors and producers rejected the concept of a national Special Milk Classes Scheme, it was quite certain that such a system would never have been implemented.
4.86.
Canada argued that the object of the supply-management system in Canada had been to provide the Canadian dairy industry with the means by which they could effectively govern their own affairs, so as to yield a fair return to producers while balancing the interests of processors and consumers. The purpose of the supply management system was to take the necessary steps to match, as closely as possible, the quantity of milk to be marketed for domestic use with Canadian domestic demand. This, coupled with border measures to control imports into the domestic market, allowed for the maintenance of price levels for milk in the domestic market higher than would otherwise be obtained in the absence of such measures. In order to provide Canadian dairy producers with the means to operate such a system, governments in Canada had put in place the legislative and regulatory framework to allow such a self-governing regime to function. Provincial governments had passed enabling legislation to provide for the establishment of producer-run marketing boards.
4.87.
It followed that the Canadian dairy marketing system rested on the foundation of the producers themselves. Canada noted that the United States conceded that the producer-run boards were "private entities", "private parties" and "dairy producer organizations".112 They were organised throughout Canada, from the local level, including co-operatives, and then up to the provincial113 and national114 levels. Through these producer organisations, the individual producers maintained close links with each other and had an effective means for the development of policies and other marketing initiatives. At the provincial level, the key producer-controlled institutions were the milk marketing boards (the "producer boards"). In each province, these boards only came into being through an affirmative vote of the producers. In most provinces,115 and in Ontario and Quebec in particular, the membership of the boards consisted exclusively of producers. Through on-going consultations at the district and provincial levels, the producers held these elected representatives responsible for their actions. In Quebec, many of the important decisions of the board were the subject of individual voting at general meetings of the producers. Thus, without question, the boards were controlled by, and act on behalf of, the producers.116
4.88.
Canada argued that the heart of the Canadian dairy system was to be found in the CMSMC (paragraph 2.27). The central actors in the CMSMC were the various provincial producer boards. Each province sent a delegation. These delegations were led by "designated representatives" who were senior executives with the respective producer boards.117 Provincial government officials participate in an oversight role, without having voting rights. It was the producer board representatives who had the right to vote on and thus directly participate in the CMSMC decisions. As a result, for seven of the nine provinces, representing 92 per cent of Canada’s dairy producers, the designated representative was a dairy producer elected to the producer board by fellow dairy producers.118 Canada noted that even a cursory review of the signature pages of the Comprehensive Agreement on Special Class Pooling showed the signatures of producer boards in addition to the signature of government officials. Nevertheless, Canada did not argue that only the producers or the industry were involved. The Comprehensive Agreement on Special Class Pooling, like the NMMP and the CMSMC, represented a co-operative arrangement in which all interested parties, the producers, through the producer boards, and government representatives in their public interest capacity119 participated.
4.89.
Canada argued even a minimal knowledge of Canadian political life would confirm that no federal government institution could ever begin to dictate to provincial representatives where provincial jurisdiction and interests were at stake. Indeed, in this case, the chief provincial speakers were representatives of the industry, not government, this made the picture of unilateral federal government action painted by the Complainants even less credible. The true character of the CMSMC, based on co-operation and consensus, was representative of the character of the Canadian dairy system as a whole. Although the processors did not have voting status in the formal arrangements for the CMSMC, they had a prominent role at the CMSMC table. This was because, as a matter of practice, every effort was made to reach a consensus that could be supported by all provincial producer representatives as well as the other stakeholders in the industry.
4.90.
Canada noted that the Complainants, rather than addressing the CMSMC, preferred to stress the CDC, a federal crown corporation, suggesting that it was the true controller of the Canadian dairy system. Their approach was understandable given the true character of the CMSMC, i.e., control in the hands of the producers in consultation with other concerned parties, and the difficulties that this presented for the Complainants’ case. Canada argued that whereas the NMMP had provided that the CDC could take decisions for the CMSMC where no consensus was reached in the CMSMC, the provisions of the Comprehensive Agreement on Special Class Pooling took that override authority away with respect to all matters covered by that Agreement (paragraph 2.33). Since the matters covered by the Comprehensive Agreement on Special Class Pooling were very broad, this had left very little subject to the original override provision in the NMMP. This reality fundamentally undermined the attempt by the Complainants to allege that it was the CDC that controlled the Canadian dairy system, not the CMSMC.
4.91.
Alternatively, Canada noted that the Complainants had argued that even to the extent that the CMSMC ran the system, it was a body composed of government officials. Canada recalled that for meetings of the CMSMC, for seven of the nine provinces, it was the representative of the producers who took the lead in discussions between provincial delegations. This producer representative, the "designated representative", was elected to the respective producer board by fellow dairy producers. As a result, it was producers who spoke most prominently with respect to the various aspects of the Canadian dairy system as they were brought before the CMSMC and this provided an important flavour to the nature of those proceedings. Nonetheless, representatives of provincial governments did attend and did speak at the meetings of the CMSMC when matters arose touching on their responsibility to oversee the public interest. Thus, through its consensus-based decision making structure, the CMSMC served to bring together the representatives of industry who took the lead on operational matters and representatives of government who ensured that the public interest was respected.
4.92.
Canada noted that with respect to the processor margins (paragraph 4.81), transactions occurred at the highest milk price which the CDC believed it could achieve, subject to the potential exporter's willingness to participate. Canada argued that there were no standard mark-ups or margins, although the CDC used the return for milk producers that would be generated by making butter and SMP for export as the baseline in negotiating the milk price. The processor margin for CDC export sales was the subject of extensive negotiations between producers and processors following the completion of the Comprehensive Agreement on Special Class Pooling on how the surplus removal system would be implemented.

(e) In- and Over-quota Milk and the Producers' Choice

4.93.
New Zealand noted that Canada placed considerable emphasis on the distinction between in-quota and over-quota milk that was destined for export. New Zealand maintained that the objective of the Canadian distinction between in-quota and over-quota milk was to distance over-quota milk even further from government agencies such as the CDC and the provincial milk marketing agencies. However, the distinction between in-quota and over-quota milk was an irrelevant distraction. It was an artificial distinction created for Canadian regulatory purposes that had no reflection in reality. Milk was milk in the tank or in Special Class 5(d) or (e). What Canada had done through the Special Milk Classes Scheme was to determine that the revenue that producers received for their milk would be based on one price for a certain quantity of milk and on a different price for another quantity of milk. Canada had decided that exporters were to pay a different price for milk for products destined for export than that for products destined for domestic consumption; it had thus provided an export subsidy regardless of whether the milk sold to processors for export was classified as in-quota or over-quota milk.
4.94.
New Zealand noted that the distinction between in-quota and over-quota milk was not as clear-cut as Canada described it. Whether or not an increase in producers’ supply would constitute over-quota production was a determination made by others, not by the producers. Each province managed a complex system according to its own rules. As the Chairman of the CDC had said in April 1998, "there are almost as many ways to manage provincial quota, over-quota production and payment mechanisms as there are provinces".120 Most provinces had a complex monthly credit and debit system according to which producers who did not fill their quota one month could carry over "quota credits" for future months, or indeed, in the case of Manitoba, for future years.
4.95.
New Zealand noted that Canada claimed that over-quota production arose when producers in a province produced milk in excess of their quotas and as a result the province as a whole exceeded its share of the national MSQ in a month. Thus, whether a producer produced what ultimately would be determined to be over-quota milk might depend on the level of production in the province as a whole. Furthermore, it was ultimately the CMSMC which decided what would be sold as in-quota and what would be sold as over-quota milk. Hence, a producer could not know whether milk produced on any particular day, week or month would be treated as in-quota or over-quota, or whether it would be used for domestic or for export purposes. A producer could simply decide to produce more milk. The consequences of producing more milk were in the hands of the milk marketing boards and the CMSMC. Thus, the concept of producers systematically deciding to produce over-quota milk for export purposes after considering current world market prices was far-fetched.
4.96.
Nevertheless, New Zealand emphasized that focusing attention on whether a producer produced in-quota or over-quota milk ignored what was really at issue in this case. That was, whether providing exporters with access to lower-priced milk for products destined for export - as occurred with Special Class 5(e) in- and over-quota, as well as with Special Class 5(d) - constituted a subsidy within the meaning of the Agreement on Agriculture.
4.97.
New Zealand argued that Canada’s claim that individual producers decided, on market considerations, whether to produce over-quota milk misrepresented the facts. Over-quota milk production was not always a result of a deliberate decision by a producer to produce for export.121 It made sense for producers to slightly overshoot their quotas to allow for fluctuations in milk production caused by weather or biological factors. This was not a deliberate decision to produce for the export market; it was a rational decision to maximise revenue by ensuring they do not underfill their more lucrative in-quota entitlement.
4.98.
The United States argued that milk was milk; it was not labelled in-quota or over-quota when it was sold – the distinction was not of concern to processors or exporters. What was of concern to exporters was the ability to access their major input at a low cost, which was precisely what classes 5(d) and (e) set out to achieve. Like New Zealand, the United States emphasized that there were a variety of factors, mostly beyond the control of Canadian dairy farmers, that determined whether there was over-quota production in the first place. Many factors, including weather, quality of feed, and the biological condition of the dairy herd, would affect the amount of milk produced in any given time period. Thus, despite a farmer’s best efforts to confine production to the amount of his/her quota, doing so was more an art than a science. Production could not be regulated with precision. Moreover, a farmer had an incentive to try to produce the full amount of his/her quota. This was because producing the full amount of quota provided the best opportunity to recover as much of the applicable fixed costs of production as possible since within-quota production was entitled to receive the higher domestic prices, or, at least, a blended price that included primarily higher domestic prices.
4.99.
The United States argued that several authorities had testified regarding the difficulty of precisely producing to the level of the quota. Mr. Rick Phillips, Director of Government Affairs, Dairy Farmers of Canada, made the following statement regarding the uncertainty of over-quota production:

"In 1997, Dairy Farmers of Ontario conducted a survey to determine the extent of producer interest in providing milk for the Optional Export Program. Now, as you probably differentiate this milk supply, which represents a conscious and voluntary exposure to world markets from over-quota milk - and I wouldn’t want that to be said in public, as well - producers produced or filled their quota, and that’s basically a producer behaviour, and the level of over-quota milk as Mr. Core [President of Dairy Farmers of Ontario] has stated a bit earlier, is sort of dependent on the biological conditions that they find on the farm. In fact, if you happen to be in a state where the feed is good and the cows are calving properly and there is not a whole bunch of diseases, these can add together to create a fairly significant level of over-quota milk. When in fact, under normal chance circumstances, if you didn’t have a lot of good things happening at the same time, the level of over-quota milk would be much less."122

4.100.
The United States noted that Mr. Phillips’ views were confirmed by the testimony of Mr. Guy Jacob, President, Canadian Dairy Commission, before Canada’s Parliamentary Standing Committee on Agriculture and Agri-Food.

"Last year the quota was cut by 3%. A farmer had the choice of reducing his production by 3%. He could sell one cow out of that barn and reduce his net revenue. He had to reduce his production because the quota was cut last year by 3%. That’s the choice he has. He may decide to keep his production at the level of the previous year and then produce 3% over quota. Then it may happen that this year the feed was a little better, the climate was a little better, and it just happens that he’s producing 6% over his quota. That milk is being removed by the CDC at the international price."123

4.101.
The United States argued that when the level of production by dairy farmers was so greatly influenced by factors largely beyond the producers’ control, Canada’s assertion that an increase in over-quota production in a single year evidenced the willingness of its farmers to sell at world price levels was simply not credible. Rather, farmers’ willingness to sell milk for use in the OEP124 would be a far better gauge of interest in selling at the Special Class 5(e) price for export. Yet testimony from Mr. Phillips125, indicated that dairy farmers in Ontario had shown very little interest in participating in the OEP. There had been virtually no use of the OEP for the first two years of its authorized usage.126 Any increase in usage during the 1997/98 year was most likely attributable to the unusual level of over-quota production in that year and the fact that farmers had an opportunity to obtain a higher price for their milk under the OEP than from the CDC dictated price under Special Class 5(e). For example, Manitoba was selling milk for OEP contracts in 1997/98 at $32 per hectolitre compared to the Special Class 5(e) price of between $23 and $25.127
4.102.
The United States further noted that Mr. Phillips had predicted in hearings before the Canadian International Trade Tribunal that it was very unlikely that dairy producers would voluntarily participate in a new special class:

"So, again, I would note that a 5-B, which is a typical Class 5 price, most of the variable costs are covered. But when we go down to the world price, the $23.38 in this instance, look across there, you find that the cutoff point is around 18 percent of producers whose variable costs would be covered. That means that a vast majority of producers would definitely not want to produce milk at the world price."128

4.103.
The United States also observed that Mr. Phillips had testified that only one-half of one per cent of producers in Ontario had shown any interest in participating in the OEP that, like the Special Classes, involved the sale of milk for export at approximations of world market prices.129
4.104.
The United States stressed that Canada’s argument that an increase in over-quota production since the implementation of the Special Milk Classes was evidence that milk producers were deliberately choosing to export milk at world market prices was flawed. First, over-quota production actually declined in the first full year, 1996/97, after the implementation of the Special Milk Classes Scheme. The annual report of the Dairy Farmers of Canada showed that over-quota production actually declined between 1995/96 and 1996/97, both in actual volume and as a percentage of total production.130 Furthermore, the purported increase in over-quota production between 1996/97 and 1997/98 appeared to be attributable in large part to a decision to reduce the Market Share Quota (MSQ) for 1997/98. The United States noted that information contained in Canada’s Exhibit 16 showed that the MSQ had been reduced by one million hectolitres between 1996/97 and 1997/98. When such a major reduction in MSQ occurred, it all but compelled an increase in over-quota production, as it was difficult, if not impossible, for milk producers to reduce production in such a precipitous manner. A one million hectolitre reduction in the MSQ equalled almost one-half of the total over-quota production in 1996/97, which according to the DFC was 2.21 million hectolitres.
4.105.
The United States argued that there was the definitional question of what actually constituted over-quota production. Various "flexibility" provisions authorized by several provinces allowed for a departure from the pre-existing practice of determining whether a particular milk producer was over-quota based on an analysis of a dairy farmer’s daily or monthly production levels. Canada had confirmed the United States understanding that both Alberta and Manitoba permitted such adjustments.131 Alberta and Saskatchewan performed a year-end price adjustment which applied under-delivered quota against over-quota deliveries. Manitoba currently allowed flexibility for up to 25 days of quota production. This was characterized as a credit that the dairy farmer could use on a rolling basis to apply against over-quota production. The United States understood that other provinces had similar provisions. In addition, Manitoba had introduced a so-called "cover-off" that provided yet another hedge against over-production.132 While this "cover-off" mechanism was originally instituted for only the months of August through November, Manitoba later extended it to additional months. To the extent that other provinces had similar arrangements, their existence undermined Canada’s contention that there existed a consistent definition of over-quota production that dairy farmers consider in their daily production plans.
4.106.
The United States noted that whether milk was in-quota or over-quota and what Class price it received was in fact so confusing, that many milk producers apparently did not know whether their production was over-quota and, if it was, what price they would receive for the over-quota production. The Manitoba Milk Marketing Board’s newsletter Milkline has responded to milk producers confusion with a number of articles attempting to explain the mechanics of these various schemes.133 In the face of such uncertainty, it was difficult to comprehend Canada’s assertion that dairy farmers were producing over-quota milk in response to price signals from the world market. Moreover, the United States stressed that the price earnings information provided to producers was largely, if not entirely, retrospective.134 While Canada stated that producers knew that Special Class 5(d) and (e) prices were set on the basis of negotiated transactions, Canada omitted to mention that the returns that producers received from Special Class sales were pooled under the Special Class Agreement if they consisted of in-quota production. In the case of over-quota production, the producer also received a weighted average return based instead on all Class 5(e) transactions during the year. Thus, any individual negotiated transaction price was of no direct consequence to an individual producer; his ultimate return from in-quota exports was determined based on the Special Class pooled price and over-quota sales were based on a weighted average Class 5(e) price.
4.107.
Canada recalled that milk to be used in exported products under the Special Classes arose from two sources of milk: in-quota production and over-quota production.

(a) In-quota milk: Producers, acting collectively through their milk marketing boards and the CMSMC, controlled the quota level and the amount of milk that was likely to be exported from in-quota production. If prices obtained for milk used to make products for export were not high enough, producers could decide, through their boards, to reduce the quantity of MSQ.135

(b) Over-quota milk: Individual producers decided whether to supply milk above their individual production quota in the full knowledge136 that over-quota shipments would receive world market returns. In fact, considerable numbers of producers voluntarily, as a business matter, chose to engage in over-quota production.

4.108.
Canada noted that the Complainants suggested that there was no real distinction between over-quota and in-quota milk: "milk was milk", it was all fungible. Canada argued that this was irrelevant: the molecules of milk were not tracked into export or domestic markets. What was relevant was that the producer was perfectly aware when his milk was picked up at the farm gate that his shipment was within his marketing quota or was "over-quota". If it was "over-quota", then the producer knew that his return for this shipment would be in accordance with actual Class 5(e) prices, world market-based prices. This was true regardless of where the molecules in that truckload of milk actually ended up.
4.109.
Canada argued that for both in- and over-quota milk, the essence of the Canadian system was that it exposed milk producers to market signals from the export market and allowed them to make business decisions based on those signals. In contrast to the allegations of the Complainants, the government did not direct milk to be used for manufacturing products for export. On the contrary, while milk sold on the domestic market was subject to marketing quotas and price regulation or approval, the quantity or price of milk sold for use in products destined for export markets was determined on a strictly commercial basis. There was no government involvement whatsoever in decisions to participate in the export market. That was a choice left entirely to the producers. Over-quota production for exports did not constitute any sort of pre-condition for a producer’s annual allocation of quota for milk sales into domestic markets. In short, the decision to produce for the export market or not was one that was made by producers alone on the basis of true price signals with the objective of profit maximization. The essential feature was that the milk producer was exposed to world market-driven prices for dairy products and responded entirely on a commercial, market-driven basis. Canada emphasized that the individual farmers knew their individual quota level and knew that any production above their individual quota would be paid to them at world market prices.137
4.110.
Canada argued that in the case of in-quota milk, the decision to provide for a certain amount of milk within the annual Market Sharing Quota (the "MSQ") was taken by the producers collectively. These decisions were taken at the producer board-dominated Canadian Milk Supply Management Committee (the "CMSMC"), in consultation with the processors. Producer representatives on the CMSMC were accountable through a system of producer democracy that began at the district level, with elected district or regional milk committees. Generally elected members of these committees were directors at the provincial level, and provincial boards were the main voice in deciding on production targets in the CMSMC. Thus, the producers were free to collectively determine whether and to what extent they wished to provide in-quota milk for export purposes. There was no evidence of government control, direction or coercion in this process.
4.111.
In the case of over-quota milk, any qualified dairy producer in Canada was free to produce as much milk as he or she chose. Specifically, the producer was free to produce any amount of milk over his or her domestic marketing quota, i.e., over-quota production with the understanding that their return for over-quota milk would be based on actual world market-based prices, i.e., the prices realised from Special Class 5(e) sales, taking into account their individual cost structures. The individual over-quota producer received a Special Class 5(e) return whether or not the province’s producers taken collectively were in an over-quota position.138 Accordingly, decisions to participate in over-quota production and to supply milk for export use were market-driven choices made by individual producers. As such, the absence of any government control or direction was clear and unequivocal.
4.112.
In respect of the assertions of the United States with respect to testimony of officials of the Dairy Farmers of Canada (DFC) before the Canadian International Trade Tribunal (CITT), Canada argued that the testimony had been taken completely out of context and did not support the US proposition (paragraph 4,102). The context of the DFC testimony was an inquiry by the CITT, initiated at the request of the Canadian government, into issues raised by increased imports of blends of dairy products, particularly butteroil/sugar blends, into Canada. Among the various options considered by the CITT was the possibility that producers may wish to create a Special Class price to service the domestic butterfat market at world prices. The DFC testimony to which the United States referred was addressing this option, not the question of individual producers deciding to produce milk at world market prices for the export markets. This distinction had been made abundantly clear in the DFC's Final Agreement before the CITT:

"7.1.6 There is evidence that some dairy producers produce quantities in excess of MSQ. This is done by producers who voluntarily seek to increase production for participation in world markets. Certain low cost producers may also voluntarily decide to actively participate in world markets through the Optional Export programe. These producer decisions, however, must not be confused with a proposal to service the domestic butterfat market using within quota production at world prices. The recent decision to reduce MSQ is clear evidence that dairy producers are not willing to produce irrespective of domestic market requirements."139

4.113.
In the case of both in-quota and over-quota milk production, the claim of the Complainants that the Canadian dairy system was government-controlled and directed had to fail. Particularly with respect to the marketing of over-quota milk for export purposes, this represented a decision, by the governments, not to intervene, to avoid the use of the discretionary authority provided to the boards and rely instead on market-driven results. To suggest that such a restraint from intervention constituted government action resulting in export subsidies was not logical.
4.114.
In respect of pooling, Canada argued that contrary to suggestions from the Complainants, pooling was not an obligation that had been forced on the producers by coercive governments. Pooling was a consensus-based arrangement that the producers, through their boards, had agreed to, pursuant to the terms in the Comprehensive Agreement on Special Class Pooling (the P9 Agreement) and the P6 and P4 Agreements (paragraph 2.24). The producer boards were full signatories of those agreements, which were co-operative agreements involving all interested stakeholders. These agreements were not agreements between governments to impose on the dairy industry and the dairy producers, in particular, certain arrangements and requirements, as suggested by the Complainants. Each producer board had joined in pooling freely, and they were equally free to leave the pooling arrangements. The provincial producer boards could agree at any time to cease any sharing of revenues and markets. Indeed, to give a practical example, the Manitoba producer board had temporarily opted out of the P6 pool, pending their evaluation of their participation. Under the enabling legislative framework, such decisions could not be overridden by provincial or federal governments.
4.115.
Canada noted that it was also possible for a provincial producer board to partially withdraw from the pooling of revenues if it so chose. For example, as outlined in US Exhibit 39, under an experimental programme introduced in the province of Manitoba, two per cent of each producer’s daily quota had become optional and was no longer pooled. Producers could choose to ship this quantity of their quota, at a known, non-pooled return, based on realised returns in Class 5. The volumes associated with this experimental programme approximated the share of in-quota Class 5(d) and (e) production in Manitoba. By filling this portion of the provincial MSQ through voluntary shipments by producers at non-pooled prices, the producer board reduced the exposure of other producers who did not ship the optional quantity to Class 5 returns. In other words, contrary to US assertions, this was an example of a provincial producer board giving its members an option to increase or decrease their participation in Class 5 sales. Significantly, this was a unilateral decision by the Manitoba producers acting through their board. Contrary to the image of government coercion suggested by the Complainants, no permission was required from the Government of Canada or the CDC. No government sanctions followed this decision by producers to reduce their participation in the pooling of returns. Canada emphasized that the Special Milk Classes Scheme was producer-driven and necessarily based on co-operation and consensus.
4.116.
Canada noted that the United States had stated that the amount of in-quota milk sold for export use exceeded that from over-quota sources. While it was true that in-quota export sales did exceed over-quota sales in 1995/96, by 1996/97 the two were in balance. Most recently, in the 1997/98 dairy year, with the growth of over-quota and the full use of the "sleeve" in domestic markets, over-quota export sales had begun to greatly exceed in-quota export sales. It was also suggested that over-quota production was actually in decline. The United States noted that there was a decline from 1995/96 to 1997/98. There was indeed a small decline between those two years but it was misleading to suggest that this was the general trend.140 In fact, there had been substantial growth the previous year and an even greater increase in 1997-98. It had also been suggested that growth in over-quota production was the result of the reduction of MSQ. Yet this was equally invalid. The growth in over-quota had greatly exceeded any reduction in MSQ.141
4.117.
Canada noted that the Complainants assumed that when products were exported at lower prices than those same products would command in the home market, there had to be a subsidization of the lower-priced exported products through profits obtained on the domestic market. Canada rejected these arguments. The Complainants had offered no evidence or explanation of why, in the absence of government direction or without the linking of domestic sales quota to export performance, producers would give away their profits from domestic sales in order to make unprofitable export sales. Indeed, they had failed to demonstrate any incentive to finance export sales from domestic profits. This argument was based on an assumption that milk producers behaved irrationally, or that governments somehow, in some unexplained way, forced them to reduce their net profits to engage in export sales. Canada argued that as a result of the supply management system and the presence of border protection, there were two very different markets: (i) a limited domestic market allocated to producers via quotas; and (ii) an open international market available to any producer willing to supply under conditions prevailing in that market. There was nothing in the Canadian milk marketing system that forced producers to supply the international market if they did not want to do so. The Complainants had failed to provide any evidence supporting the existence of any such mandatory performance requirement imposed on Canadian producers that would confer a benefit to exporting processors. Hence, Canada argued that the decision to produce or not to produce was one made on the basis of the same criteria that every commodity producer, indeed, any business person would make – the enhancement of the producer's net profits. The only reason for Canadian producers to sell products for the export market was because they could do so profitably. The protection of the domestic market afforded by tariffs did not alter this basic fact.
4.118.
In respect of the price difference between OEP sales and Class 5(e) sales, Canada emphasized that the price difference between OEP sales and Class 5(e) sales stemmed from the commercial terms of the contract under which producers produced OEP milk. In OEP transactions, both the volume of milk supplied to the processor by the individual producer and the price paid for that OEP milk by the processor were contractually negotiated well in advance of the milk production with a view to fulfilling a pre-planned export contract made by the processor with a foreign buyer. By way of contrast, Class 5(e) milk sales were not pre-planned. Class 5(e) milk was normally used in dairy products sold on international spot markets. The pre-planned nature of an OEP transaction provided the processor with a secure supply of milk, an assured level of plant utilization and a guaranteed export sale price for a transaction identified in advance for its superior returns. For this assurance of supply, processors were willing to pay a premium for OEP milk. Canada argued that these market characteristics would exist even if producers negotiated Class 5(e) sales prices on their own or through their producer boards rather than using the CDC as a collective sales agent. As for the matter of processor margins, OEP processor margins were commercially confidential to the individual processor and were not made known to the producers, the producer boards or the CDC.
4.119.
Canada argued that over-quota and OEP production came down to a matter of individual producer choice, as was illustrated by the dairy producer leaders who were CMSMC representatives and members of the Board of Directors of Dairy Farmers of Canada. Some of those producer representatives simply aimed to fill their domestic quota and not to participate in export market opportunities through the production of over-quota or OEP milk. For example, the DFC Director for Saskatchewan, Leo Bertoin, had been 0.04 per cent over, 0.03 per cent under and 0,004 per cent over his producer quota for 1995/1996, 1996/1997 and 1997/1998, respectively. By way of contrast, the President of DFC, Barron Blois, a producer from Nova Scotia who was also a provincial spokesperson at the CMSMC, had actively participated in over-quota export opportunities made possible by changing his feeding programme to reduce cash costs. Mr. Blois had consistently been in an over-quota position since the Special Milk Classes Scheme was introduced and was currently producing 20 per cent above his producer quota. The DFC Director for New Brunswick, Jacques Laforge, who was also the provincial spokesperson for New Brunswick at the CMSMC, had only produced to the level of his domestic quota but had also actively concentrated on available OEP markets.142
4.120.
The United States noted that Canada’s argument that producers collectively decided to produce for export as part of their determination of MSQ for the year was contradicted by Canada’s experience in 1997. In 1997/98, the MSQ was set at the beginning of the year at a level that presumably built in a specific amount for planned exports, in the so-called "sleeve". However, Canada had stated that almost all of the sleeve had been used last year for the domestic market. Consequently, any decision to produce for export within in-quota production, based on the MSQ at the beginning of the year, simply had not been realized when the domestic marketplace required more milk. Instead, those producers received primarily domestic prices for almost all of their in-quota production. As had been noted, whether milk was classified as in-quota or over-quota was subject to a variety of factors, and especially the fluctuation of MSQ from year to year.
4.121.
The United States emphasized the significant increase in MSQ for the 1998/99 year (see table under paragraph 2.31). This increase was the clearest evidence that the CDC/CMSMC had seriously misjudged domestic requirements in the previous year. The reduction in MSQ in 1997/98 also was in large measure responsible for the unusual change in the relative proportion of over-quota versus in-quota exports in that year.143 Even if domestic consumption in Canada remained at the high levels enjoyed in 1997/98, the 1998/99 increase in MSQ would necessarily result in a significantly higher percentage of in-quota exports in the current marketing year. The United States submitted that the 1997/98 marketing year was an aberration with respect to the low level of in-quota milk that was used for planned exports or determined to be surplus. That situation was primarily a result of a reduction in MSQ that turned out to be totally unjustified by market circumstances. In this connection, the United States noted that this judgment error had been corrected for the 1998/99 marketing year as the MSQ for the current year had been readjusted to restore the full amount of this reduction and actually included an increase over the 1996/97 MSQ level. This should result in a decline in over-quota production. The restoration of MSQ to earlier levels no doubt was a result of milk producers’ complaints that they were being compelled to sell milk at Class 5(e) prices by the reduction in MSQ in 1997/98 which bore no relationship to market conditions. The United States argued that this significant shift of MSQ highlighted the artificiality of the over-quota/in-quota distinction, and the inability of milk producers to quickly adjust to dramatic swings in the MSQ such as occurred between 1996/97 and 1997/98. In fact, the reduction in MSQ in 1997/98 of approximately 3 per cent, when factored with Canada’s statement that only over-quota production by individual producers above the 105 per cent level could be considered to be deliberate (see Footnote 121), accounted for at least two thirds of all over-quota production during the 1997/98 marketing period. Thus, Canada’s assertion that over-quota production reflected milk producers' deliberate decisions was belied by the very information that Canada had submitted.
4.122.
Like New Zealand, the United States stressed that the designation of milk as "surplus" was essentially an arbitrary one. The Canadian system was predicated on the assumption that the domestic population would only consume so much butter, so much cheese, so much ice cream and yoghurt, etc., at desired price levels in a given year and that this translated into class prices for milk. However, many of these dairy products were storable. Therefore, the decision to designate milk as "surplus" was really a decision not to build domestic stocks of dairy products for later consumption. Processors resisted stock building because as stocks grew they exercised downward pressure on the prices at which processors could sell their dairy products, and processors' profit margins would be eroded as a result. Producers disliked stocks because dairy product stocks represented a quantity of milk that would not be required at a future date, i.e., milk production would have to be cut while Canadians consumed the products in stock. Canada’s solution was to export the surpluses using subsidies, and thereby maintain both high domestic price and production levels.
4.123.
The United States refuted Canada's comparison of how prices were established under the OEP versus under Class 5(e) milk (paragraph 4,118). Canada's own explanation supported the conclusion that the Special Class prices were particularly low, given the international dairy marketing conditions. Canada was over-simplifying the commercial context of these sales. While Canada contended that OEP prices could be higher, in part, because they generally involved advance purchases of milk and, therefore, established a secure source of milk for the processors – Canada had earlier stated that many of the export transactions under Special Class 5(e) involved repeat sales involving the same exporters and international customers. If that was true, then the distinction that Canada now attempted to draw between OEP and Class 5(e) sales lacked a factual foundation.

2. "Export Subsidy" - the Interpretive Context

(a) The SCM Agreement and the Agreement on Agriculture

4.124.
Canada argued that the definition of "export subsidies" for the purposes of the Agreement on Agriculture was found in Article 1(e) of the Agreement. This definition contained two components:

(a) the first component was "subsidies contingent on export performance";

(b) the second component included as export subsidies for the purposes of the definition the export subsidies specifically listed in Article 9 of the Agreement.

4.125.
Canada further noted that Article 10.1 of the Agreement on Agriculture made an explicit reference to "(e)xport subsidies not listed in paragraph 1 of Article 9." Article 10.3 spoke of "no export subsidy, whether listed in Article 9 or not." Thus, Canada submitted that Article 10 in general, and Article 10.1 in particular, recognised the dual nature of the definition found at Article 1(e) of the Agreement. Article 10.1 applied only144 with respect to "export subsidies" as defined in Article 1(e) other than those export subsidies that were included in the definition by virtue of having been explicitly listed in Article 9. What was left of the definition of "export subsidies" was the first component of the definition (i.e., "subsidies contingent on export performance" without including export subsidies specifically listed in Article 9 of the Agreement). Thus, the "other export subsidies" referred to in Article 10.1 meant "subsidies contingent on export performance", excluding the export subsidies specifically listed in Article 9 of the Agreement.
4.126.
The term "subsidies contingent on export performance" itself had two components: (i) "subsidies" and (ii) "contingent on export performance". The term "subsidy" was not defined in the Agreement on Agriculture. Canada's position was that the prime contextual interpretative source for the meaning of the term "subsidy" was the definition of "subsidy" found in the SCM Agreement.145 While the potential application of other sources for the interpretation of the term "subsidy" in the Agreement on Agriculture could not be excluded, the Complainants had not been able to identify any source having anything like the interpretative force of the definition of "subsidy" found in the SCM Agreement. Accordingly, "other export subsidies", as used in Article 10.1 of the Agreement on Agriculture, included "subsidies" (as interpreted in the context of the SCM Agreement) that were contingent on export performance, other than the export subsidies specifically listed in Article 9 of the Agreement on Agriculture.
4.127.
Canada argued that the list of export subsidy practices found in Article 9.1 served two purposes in the Agreement on Agriculture. On one hand, it served as an exhaustive list of export subsidy practices that were subject to the reduction requirements set out in the Agreement. It also provided an illustrative list of export subsidy practices to be included within the definition of "export subsidy" in Article 1(e) of the Agreement. Both as an exhaustive list under Article 9 and as an illustrative list for the definition in Article 1, the text precisely reflected the common agreement by all Members regarding which practices should be considered "export subsidies" for the purposes of the Agreement on Agriculture. By the same token, the text also reflected where there was an absence of common agreement with respect to whether a particular practice should be considered to be an export subsidy (paragraphs 4,446 and following). In respect of criteria for export subsidies relating to differences between export and domestic prices, no reference could be found amongst the items listed in Article 9.1, with the exception of Article 9.1(b). The fact that a reference to domestic and export prices was included only with respect to paragraph (b) of Article 9.1 suggested that the negotiators could agree to such a reference only with respect to Article 9.1(b).146
4.128.
Canada emphasized the importance of the origin and character of the definition of a "subsidy" in the SCM Agreement. The SCM Agreement achieved what had not been possible during the life of the GATT 1947: a definition for the term "subsidy". This definition reflected a compromise reached by the negotiators in the Uruguay Round. As such, it reflected the practical realities of negotiations and represented a statement of what, by the end of the negotiations, the negotiators had been able to agree would be a "subsidy" for WTO purposes. It followed, therefore, that any measure or practice that did not fall within the terms of the definition could not be considered to be a subsidy for WTO purposes, regardless of any other conceptions or proposals as to what ought to constitute a "subsidy". Canada argued that it was not suggesting the agreed definitions of "subsidy" and "export subsidy" were perfect conceptions. There was no doubt that many WTO Members would prefer amendments in pursuit of their own policy objectives preferences. Perhaps proposals would be brought to the negotiating table for the next round of WTO negotiations. What the Complainants could not be allowed to do was to create an effective alteration of the negotiated texts through litigation.
4.129.
New Zealand claimed that sources relevant to the interpretation of the word "subsidy" were: the Agreement on Agriculture, which provided the immediate context for the interpretation of the term "subsidy", and subsequently, the broader context of the WTO. The latter included, in particular, the SCM Agreement together with its Illustrative List of Export Subsidies, GATT 1994 and WTO/GATT practice.
4.130.
New Zealand argued that in the present case, the question of the meaning of the term "subsidy" arose on two occasions: in the interpretation of Article 9.1(a) which referred to "direct subsidies"; and in the interpretation of Article 10 which referred to "export subsidies... applied in a manner which results in, or which threatens to lead to, circumvention... ". In each case, the question had to be asked what constituted a subsidy. New Zealand contended that the answer to this question had to be sought initially in the context of the Agreement on Agriculture and then more broadly in the context of the WTO Agreements as a whole. In this regard, the particular relationship of the Agreement on Agriculture and the SCM Agreement, recognised by the Appellate Body in Brazil - Measures Affecting Desiccated Coconut147, meant that the SCM Agreement was clearly part of the context to be referred to within the meaning of Article 31 of the Vienna Convention on the Law of Treaties (the "Vienna Convention") when interpreting the Agreement on Agriculture.148
4.131.
However, referring to the SCM Agreement as part of the context for the interpretation of the concept of "subsidy" in the Agreement on Agriculture was not the same as simply fastening onto one definition from the SCM Agreement and treating it as if it were an overriding definition for the purposes of all of the WTO Agreements. In fact, reference to the SCM Agreement could include reference to the definition in Article 1, or to the Illustrative List of Export Subsidies in Annex I, or to parts of those definitions and lists. Reference could also be made within this broader context to what constituted a subsidy under GATT 1947 and under GATT practice.
4.132.
In New Zealand's view Canada’s approach in this case was based on a fundamental interpretative error. Rather than addressing the key issue of interpretation in this case - the meaning of the term "export subsidy" as used in Article 9.1 and Article 10 of the Agreement on Agriculture - Canada instead focussed on the meaning of the term "subsidy", and proceeded to argue that the exercise of interpretation of this term as used in the Agreement on Agriculture could be largely confined to a consideration of the definition of subsidy found in Article 1 of the SCM Agreement. Canada’s failure to locate the interpretation of export subsidies within the Agreement on Agriculture resulted in its ignoring the rules of interpretation applicable to the WTO Agreements which had been endorsed by the Appellate Body, and in limiting the scope of the disciplines that were carefully negotiated in the Agreement on Agriculture. Canada was hence inviting the Panel to read into the WTO Agreements an extravagant interpretative relationship between the Agreement on Agriculture and the SCM Agreement, and to make broad pronouncements on the scope of the SCM Agreement that were not necessary for this dispute.
4.133.
New Zealand argued that by its very wording Article 1 of the SCM Agreement was limited to that Agreement. The opening words of Article 1 were "For the purposes of this Agreement" (emphasis added). Article 1 went on to say, "a subsidy shall be deemed to exist if..." (emphasis added). Hence, New Zealand argued that the drafters did not intend this to be a definition for all purposes; it was simply a listing of what was deemed to be a subsidy for the purposes of the SCM Agreement. Canada had, in New Zealand's view, interpreted the terms of Article 1 of the SCM Agreement in isolation and therefore ignored the specific context of the SCM Agreement itself. If Canada had interpreted Article 1 in the context of the SCM Agreement as a whole, it would have been forced to conclude that, in the context of a discussion on export subsidies, the meaning of Article 1 had to be read also in the light of the Illustrative List of Export Subsidies. Yet, in New Zealand's view, for Canada, the Illustrative List appeared to stand alone as a separate list of export subsidies having no necessary relationship to Article 1 of the SCM Agreement.
4.134.
New Zealand further argued that there was nothing in the Agreement on Agriculture that incorporated the SCM Agreement definition. Indeed, the implication of Article 21 of the Agreement on Agriculture, which subordinated the provisions of GATT 1994 and the other Multilateral Trade Agreements in Annex 1A to the WTO Agreement to any contrary provision in the Agreement on Agriculture, was that if the Agreement on Agriculture was to be dependent on another Agreement, that dependency would have to be express. The fact that at the end of the implementation period agricultural export subsidies would be subject to the disciplines of the SCM Agreement, as contemplated in Article 13(c) of the Agreement on Agriculture, did not necessitate that there had to be a coincidence of definition between the two Agreements on what constituted a subsidy. New Zealand argued that the relevance of Article 21 of the Agreement on Agriculture was that it made plain that in the event of a conflict between the Agreement on Agriculture and another WTO Agreement, the Agreement on Agriculture was to prevail. In other words, if an export subsidy were to meet the terms of one of the sub-paragraphs of Article 9.1 (relating to the category of export subsidies subject to reduction commitments) but yet did not meet the definition of Article 1 of the SCM Agreement, it would nonetheless still constitute an export subsidy for the purposes of the Agreement on Agriculture.
4.135.
New Zealand argued that to simply transpose the definition provided in Article 1 of the SCM Agreement made no sense within the context of the Agreement on Agriculture. Article 1(e) defined the term "export subsidies" as referring to "subsidies contingent upon export performance, including the export subsidies listed in Article 9 of this Agreement." At the outset, this was potentially a broader definition than Article 1 of the SCM Agreement because Article 9 subsidies were included within it whether or not they met the definition of subsidy in Article 1 of the SCM Agreement or any other definition. They were export subsidies by virtue of the definition in Article 1(e) of the Agreement on Agriculture. It made no sense to re-test them by reference to some other definition of subsidy.
4.136.
New Zealand argued that in effect, the relationship between the export subsidies listed in Article 9.1 of the Agreement on Agriculture and the definition of export subsidies set out in Article 1 of that Agreement was similar to the relationship that existed in the case of the Illustrative List of Export Subsidies in Annex I of the SCM Agreement and Article 3.1 of that Agreement. These were export subsidies by definition and they did not need any further testing against a definition of subsidy or export subsidy set out in the respective provisions of the two Agreements.
4.137.
New Zealand further argued that while the export subsidies listed in Article 9 of the Agreement on Agriculture were exhaustive of the export subsidies for which reduction commitments had to be entered by Members, they were also, by virtue of their inclusion in the definition of "export subsidies" in Article 1 of the Agreement on Agriculture, indicative or illustrative of the broader category of export subsidies referred to elsewhere in the Agreement. This, too, argued against an interpretation of the term "subsidy" in the Agreement that limited it to the specific requirements of Article 1 of the SCM Agreement. It also had implications for the determination of what constituted an "export subsidy" under Article 10 to which the non-circumvention provisions of that Article applied.
4.138.
The United States argued that the term "export subsidy" was defined in Article 1(e) of the Agreement on Agriculture, although the term "subsidy" was not. Article 9 of the Agreement on Agriculture, however, set forth a non-exhaustive list of export subsidies. That list served to inform the meaning of the term "subsidy" for the purpose of the Agreement on Agriculture.
4.139.
The United States stressed that the meaning of the term "subsidy" in the Agreement on Agriculture had to be determined in its context, including the other WTO Agreements and the GATT 1994; it was not governed either exclusively or primarily by Article 1 of the SCM Agreement. The United States maintained that Canada’s argument that the SCM Agreement’s definition of "subsidy" was the exclusive basis for discerning the meaning of that term for purposes of all the WTO Agreements, including the Agreement on Agriculture, disregarded the plain meaning of the opening words of Article 1 of SCM Agreement as well as the views of the Appellate Body.149 The definition of "subsidy" in the SCM Agreement was relevant for purposes of interpreting the same term used in the Agreement on Agriculture, but it was not to be given more weight, however, than the provisions of the Agreement on Agriculture.

(b) The Relevance of the Vienna Convention and Negotiating History

4.140.
New Zealand argued that there was no need to define the term "subsidy" in the abstract. The term needed definition when it arose in the context of a particular provision of the Agreement. What constituted an export subsidy under Article 10 would necessarily differ from what constituted an export subsidy under Article 9, because Article 10 was concerned with export subsidies that were not listed in Article 9. Yet both articles used the term "subsidy". The correct approach to interpretation in the case at issue was to apply the customary principles of interpretation of public international law as required by Article 3 of the DSU. This meant applying the principles of interpretation set out in the Vienna Convention on the Law of Treaties - that is, giving words their ordinary meaning in their context and in the light of the object and purpose of the treaty as a whole.
4.141.
New Zealand argued that there was nothing in the negotiating history of the Agreement on Agriculture or the SCM Agreement that suggested that the negotiators intended that the definition of "subsidy" in the SCM Agreement would be simply accepted as the definition of subsidy in the Agreement on Agriculture, or for that matter in other WTO Agreements. Indeed, the fact that the negotiations were conducted separately, and that two separate Agreements were concluded, suggested that, contrary to the Canadian position, there was no understanding reached that the definition of Article 1 of the SCM Agreement would be the exhaustive or governing definition of subsidy for the purposes of the Agreement on Agriculture. Furthermore, the negotiators did not purport to work out all the details of the relationship between the two Agreements. It was well understood at the end of the Uruguay Round that the relationship between the Agreement on Agriculture and the SCM Agreement was to be worked out in practice.150
4.142.
New Zealand argued that the Uruguay Round did not result in a universal definition of the term "subsidy" for the purposes of all of the WTO Agreements. There was a definition for the purposes of the SCM Agreement in Article 1 of that Agreement, although that definition had also to be read in the light of the list of export subsidies in the Illustrative List. The Agreement on Agriculture did not contain any such definition. Thus, in interpreting the Agreement on Agriculture other definitions in the WTO Agreements, in particular the SCM Agreement, were relevant but not determinative.
4.143.
The United States noted that although the Agreement on Agriculture did not list all export subsidies or define the term "subsidy," the Appellate Body had indicated that it was neither necessary nor appropriate to confine the interpretative analysis of the terms of an Agreement to the text of the particular treaty provision.151 In addition, treaty interpreters could look to the customary rules of interpretation of public international law reflected in the Vienna Convention for assistance in construing the terms of a treaty provision. As an additional aid to interpretation the Appellate Body had emphasized Article XVI:1 of the WTO Agreement152 which provided that:

"Except as otherwise provided under this Agreement or the Multilateral Trade Agreements, the WTO shall be guided by the decisions, procedures and customary practices followed by the CONTRACTING PARTIES to GATT 1947 and the bodies established in the framework of GATT 1947."153

4.144.
The United States noted that the reports developed under GATT Article XVI, the Tokyo Round Agreement on Interpretation and Application of Articles VI, XVI, and XXIII of the GATT ("the Subsidies Code"), and the SCM Agreement were therefore relevant to interpretation of the term "export subsidy" as used in the Agreement on Agriculture.
4.145.
The United States noted that a recurring theme appeared in the GATT and WTO discussions of subsidies. There had long been a reluctance to provide a specific and exhaustive definition of the term "subsidy" for fear of inadvertent exclusion of a particular practice or inability to foresee the development of some new type of subsidy.154 Thus, a 1960 panel on subsidies noted the longstanding lack of an explicit definition for the term.155 Another panel "considered that it was neither necessary or feasible to seek an agreed interpretation of what constituted a subsidy."156 Despite this initial reluctance to confine the scope of the term subsidy by compiling lists of recognized subsidy practices, panels had identified various guidelines inherent in the GATT rules on subsidies as to what could be considered to fall within the term subsidy and what lay outside.157 Article 1 of the SCM Agreement now provided a definition of subsidy.
4.146.
The United States noted that one of the first illustrative lists of subsidies was contained in a 1960 Working Party Report, which was the precursor for the illustrative lists contained in both the Tokyo Round Subsidies Code and the WTO SCM Agreement.158 That Report set forth a detailed list of measures that had been considered to be export subsidies by a number of Contracting Parties, including both the United States and Canada.159 Among the practices determined to be an export subsidy were the deliveries by government or governmental agencies of production inputs "for export business on different terms than for domestic business... "160 This was, in the view of the United States, the essence of the Special Milk Classes Scheme, which provided for provincial government dairy boards to deliver discounted milk to manufacturers of dairy products for export. During the same year, another panel examined the scope of the export subsidy reporting requirement under Article XVI of GATT 1947.161 The panel in its review considered whether producer-financed subsidies were notifiable under Article XVI and determined that they were subject to the reporting requirement where the "levy/subsidy schemes affecting imports or exports... are dependent for their enforcement on some form of government action."162 The panel’s conclusion was that producer-financed export subsidies that were part of a government-directed system were essentially indistinguishable from direct government export subsidies.163

"The Panel examined the question whether subsidies financed by a non-governmental levy were notifiable under Article XVI. The GATT does not concern itself with such action by private persons acting independently of their government except insofar as it allows importing countries to take action under other provisions of the Agreement. In general there was no obligation to notify schemes in which a group of producers voluntarily taxed themselves in order to subsidize exports of a product. The Panel felt that in view of the many forms which action of this kind could take, it would not be possible to draw a clear line between types of action which were and those which were not notifiable. On the other hand, there was no doubt that there was an obligation to notify all schemes of levy/subsidy affecting imports or exports in which the government took a part either by making payments into the common fund or by entrusting to a private body the functions of taxation and subsidization with the result that the practice would in no real sense differ from those normally followed by governments. In view of these considerations the Panel feels that the question of notifying levy/subsidy arrangements depends upon the source of the funds and the extent of government action,164 if any, in their collection. Therefore, rather than attempt to formulate a precisely worded recommendation designed to cover all contingencies, the Panel feels that the CONTRACTING PARTIES should ask governments to notify all levy/subsidy schemes affecting imports or exports which are dependent for their enforcement on some form of government action." (Emphasis added)165

4.147.
The United States argued that the analysis of the May 1960 panel was also reflected in work prepared by the GATT Secretariat in support of the Uruguay Round negotiations on the Subsidies and Countervailing Measures Agreement. In a paper prepared by the Secretariat, broadly examining a variety of issues pertinent to the subsidies negotiations,166 the Secretariat recounted the discussion of the 1960 Panel Report and its conclusions regarding the lack of a meaningful distinction between the producer-financed export subsidies that were enforceable by a government and direct government subsidization of exports.167 The United States noted that the SCM Agreement recognized in Article 1.1(a)(1)(iv) that entities other than governments could perform functions normally vested in the government, and that that could also constitute a subsidy (this argument is further developed in paragraph 4,333 and following).
4.148.
The United States pointed out that in the Uruguay Round, producer-funded export assistance with significant government related action continued to be viewed as an export subsidy. Thus, the discussions during the Uruguay Round in the Secretariat papers prepared for the Chairman of the Committee on Agriculture168, as well as the proposed framework agreement developed by Chairman de Zeeuw169 and the notes prepared by Chairman Dunkel170 all characterized producer-financed export assistance as export subsidies. The only issue that surrounded producer-financed export payments and other producer-financed export assistance was not whether they were export subsidies, but whether they should be exempted from any disciplines imposed on export subsidies on agricultural products. The Agreement on Agriculture resolved this question by including such subsidies in Article 9.1 of the Agreement.
4.149.
The United States argued that producer-financed export subsidies had been treated as export subsidies because they were likely to have the same distorting effect on trade and competition as subsidies paid from a government treasury. Producer-financed subsidies conferred a benefit on the exported product and, when substantial government involvement was present, as in Canada’s Special Milk Classes Scheme, the resulting subsidy could not be distinguished from purely governmental action. This fact was recognized in Article 9.1(c) of the Agreement on Agriculture that provided that producer-financed export payments were an export subsidy subject to the Agreement’s reduction commitments (further developed in paragraph 4,196 and following).
4.150.
In this respect, the United States pointed out that Article 9.1 of the Agreement on Agriculture referenced six broad classes of export subsidy practices which were subject to reduction commitments. There had been no apparent effort in Article 9.1 to specify particular subsidy programmes. Instead the intent had been, as initially set forth in Chairman de Zeeuw’s "Framework" and later in Chairman Dunkel’s "Checklist" and "Notes", to develop a list as all-encompassing as possible to address "direct budgetary assistance to exports, other payments on products exported and other forms of export assistance."171 Thus, these export subsidy categories were defined broadly, similar to the illustrative list in the SCM Agreement. In fact, there was substantial commonality between several paragraphs of Article 9.1 of the Agreement on Agriculture and the Illustrative List which results from the drafters’ initial consideration of the Illustrative List, as well as the SCM Agreement’s definition of export subsidy, as possible models for the export subsidy disciplines in the Agreement on Agriculture. Indeed, the six categories were sufficiently broad that there was potential (and actual) overlap between the coverage of the subsections. For example, subsidies used to reduce the cost of marketing exports of agricultural products that were addressed in subsection (d) of Article 9.1 could also be captured under the broader category of direct export subsidies set forth in Article 9.1(a) of Article 9.1.
4.151.
The United States noted that both the SCM Agreement and the Agreement on Agriculture considered the provision of products or services at a price lower for export than the comparable price charged for the like product to buyers in the domestic market to be a central feature of many export subsidies.172 This principle was also reflected in the language of Article XVI:4 of the GATT 1947, which provided that:

"Further, as from 1 January 1958 or the earliest practicable date thereafter, contracting parties shall cease to grant either directly or indirectly any form of subsidy on the export of any product other than a primary product which subsidy results in the sale of such product for export at a price lower than the comparable price charged for the like product to buyers in the domestic market."

4.152.
The United States submitted that in the unlikely event of a conflict between the definition of subsidy contained in the SCM Agreement and the non-exhaustive list of export subsidies in the Agreement on Agriculture, the latter would prevail for purposes of interpretation of a provision in the latter Agreement. In other words, if one of the six categories of export subsidies set forth in Article 9.1 of the Agreement on Agriculture were determined not to be a "subsidy" within the meaning of Article 1 of the SCM Agreement, the practice would still comprise a "subsidy" for the purposes of the Agreement on Agriculture. This result was dictated by Paragraph 1 of Article 21 of the Agreement on Agriculture which provides that:

"The provisions of GATT 1994 and of the other Multilateral Trade Agreements in Annex 1A to the WTO Agreement shall apply subject to the provisions of this Agreement."

4.153.
Canada argued that Article 3.2 of the DSU recognized that the covered agreements were to be interpreted in accordance with "customary rules of interpretation of public international law." The Vienna Convention set out some of the applicable rules of international law. The rights and obligations of the Parties under the SCM Agreement and the Agreement on Agriculture must therefore be interpreted in accordance with the Vienna Convention and, in particular, Articles 31 and 32.173 Canada argued that based on these governing provisions, the interpretation of the text of WTO agreements had to proceed on the following basis:

(a) the starting point was always the actual text and its ordinary meaning;

(b) the ordinary meaning of the terms of the agreement was to be read in their context;

(c) "context" had to comprise the text of the agreement, including other contemporaneous agreements reached by all the parties, as outlined in Article 31.2 of the Vienna Convention; and

(d) recourse could be taken to supplementary means of interpretation, such as negotiating history, in the event of ambiguity or to confirm a meaning.

4.154.
Canada noted that the Appellate Body reiterated in its Report in India - Pharmaceuticals,

"The legitimate expectations of the parties to a treaty are reflected in the language of the treaty itself. The duty of a treaty interpreter is to examine the words of the treaty to determine the intentions of the parties. This should be done in accordance with the principles of treaty interpretation set out in Article 31 of the Vienna Convention. But these principles of interpretation neither require nor condone the importation into a treaty of words that are not there or the importation into a treaty of concepts that were not intended."174 (emphasis added)

4.155.
Canada argued that under Article XVI:1 of the WTO Agreement, the body of adopted pre-WTO GATT panel reports and other "decisions, procedures and practices" had been given a certain status as points of non-binding reference.175 It was evident from the views of the Appellate Body in Japan-Taxes on Alcoholic Beverages176 that pre-WTO GATT materials carried some weight but it was strictly limited. At best, these materials could be considered as supplementary means of interpretation in the context of the rules of the Vienna Convention. Supplementary means of interpretation were, by definition, supplementary. They were not to be given primacy over the principles set out in Article 31. Nor were they to be treated as having on the same significance as the text or the "context" of the terms in question.
4.156.
Canada argued that the weight to be ascribed to pre-WTO materials was also governed by relevance. Where the WTO text in question incorporated a pre-WTO text, then the relevance was stronger. However, in those circumstances where new agreements were reached in the negotiations in the Uruguay Round, materials relating to previous agreements that had been overtaken by the new regime could be of less interpretative value.
4.157.
Therefore, Canada argued that the role of the Vienna Convention was fundamental. It was the starting point for any interpretative inquiry and that inquiry had be conducted in accordance with the principles that were laid out in its provisions. Canada maintained that the United States and New Zealand appeared to be reluctant to begin with the "ordinary meaning" of the term "export subsidies" in its context in the Agreement on Agriculture or to consider the close relationship between the provisions of the SCM Agreement and the Agreement on Agriculture respecting "export subsidies". However, this was not an option open to a treaty interpreter. The starting point under Article 31 of the Vienna Convention must be ordinary meaning of the actual words agreed upon by the parties in their context, which in this case included the SCM Agreement.
4.158.
Canada maintained that the Complainants, while acknowledging the application of the Vienna Convention, had departed from its principles; they did not appear to give the Vienna Convention adequate recognition as the primary guide to interpretation of WTO Agreements. In particular, the United States appeared to confuse the relationship between Article XVI:1 of the WTO Agreement, which brought prior GATT practice into the WTO, with the principles of the Vienna Convention. This led to an analysis that mixed past GATT practice with the treatment of "export subsidies" under the SCM Agreement.
4.159.
Canada reiterated that the Agreement on Agriculture defined an "export subsidy" as a "subsidy" contingent upon export performance. However, the Agreement on Agriculture did not provide a definition of the term "subsidy". Canada recalled that Article 31 of the Vienna Convention directed that the meaning of a term had to be sought in its ordinary meaning and in its context. It then specifically included companion agreements within "context". While there was no express link between the definition of "subsidy" in the SCM Agreement and the definition of the "export subsidy" in the Agreement in Agriculture, such a link should be inferred since the SCM Agreement was part of the "context" of the Agreement on Agriculture. The definition of "subsidy" in the SCM Agreement was the only place a definition of "subsidy" appeared in the WTO agreements as a whole. As was well established, all of the WTO agreements were to be considered as part of a single integrated system and there was a particular relationship between the Agreement on Agriculture and the SCM Agreement with respect to export subsidies on agricultural products.177 Given the very similar definitions of "export subsidies" in both the SCM Agreement and the Agreement on Agriculture and the fact that the definition of "subsidy" in Article 1 of the SCM Agreement formed part of the definition of "export subsidy" as found in the SCM Agreement, the "subsidy" definition in the SCM Agreement was also applicable to the Agreement on Agriculture.178 Hence, Canada submitted that the definition of "subsidy" provided in Article 1 of the SCM Agreement was the applicable definition of "subsidy" for the purposes of the Agreement on Agriculture.. This was also consistent with the comments of the Appellate Body on the relationship of the SCM Agreement and the Agreement on Agriculture:

"[W]ith respect to subsidies on agricultural products, the Agreement on Agriculture and the SCM Agreement reflect the latest statement of the WTO members as to their rights and obligations concerning agricultural subsidies."179

4.160.
Canada further noted that the term "subsidy" was found in both the definition of "export subsidy" in Article 1 and in several items listed in Article 9.1, such as Article 9.1(a). The interpretation of this term had then to be the same in all these provisions. Accordingly, if the practices at issue did not constitute "subsidies" under the SCM Agreement definition, there could not be a subsidy for the purposes of either Article 1 or the relevant items in Article 9.1 of the Agreement on Agriculture.180
4.161.
Moreover, Canada argued that pursuant to the customary principles of treaty interpretation, a treaty had to be interpreted and applied to reflect the underlying common intention of the parties. Parties had therefore to resist seeking through dispute resolution benefits that were not obtained from negotiation. A trade agreement, in particular, expressed a delicate and carefully achieved balance of economic rights and obligations between the parties within a specific historical context. This had been repeatedly acknowledged in GATT practice where, given equally plausible alternative interpretations, GATT panels had applied the interpretation that best maintained the intended balance of the agreement.181 That approach, reflecting a longstanding principle of public international law, had been affirmed by the Appellate Body in its Reports in the United States – Restrictions on Imports of Cotton and Man-made Fibre Underwear and United States – Measure Affecting Imports of Woven Wool Shirts and Blouses from India cases.182

3. The Agreement on Agriculture

(a) Outline

4.162.
New Zealand claimed that the case at issue was about subsidies provided to exporters of dairy products who were granted access to milk for processing into products for export at prices lower than those charged for milk sold for processing into products destined for the domestic market. New Zealand emphasized that the subsidy was financed, not by way of a rebate funded by a direct levy on producers, but under a scheme that compelled milk producers to accept a lower price for milk designated for that purpose. Producers were the source of the financing of the subsidy, but the subsidy itself was provided to exporters. Both New Zealand and the United States claimed that the Special Milk Classes Scheme were export subsidy practices listed in Article 9.1(a) and (c). As such, these practices were subject to reduction commitments under the Agreement on Agriculture.183
4.163.
The United States noted that the Agreement on Agriculture, Article 1(e), defined export subsidies as "subsidies contingent on export performance, including export subsidies listed in Article 9 of this Agreement." Thus, two elements had to be shown to establish an export subsidy: (i) that a subsidy existed and (ii) that receipt of that subsidy was contingent on export performance. The United States claimed that the Special Milk Classes Scheme was a subsidy because it was a government mandated and controlled system that provided processors with milk at prices well below the comparable price for milk destined for the domestic market. In turn, these low prices allowed the processors to make export sales that would otherwise not be made and to earn "assured margins" on such sales pursuant to the CDC’s calculation of the net return to the dairy producer. The subsidy was contingent on export performance because the lower prices could only be obtained for export sales.
4.164.
Canada emphasized that Article 1 of the Agreement on Agriculturedefined "export subsidies" to be "subsidies contingent upon export performance, including the export subsidies listed in Article 9 of this Agreement." (emphasis added) Hence, if there were either an export subsidy listed in Article 9 or a subsidy contingent on export performance, there was an export subsidy for the purposes of the Agreement on Agriculture. However, as the sales of milk at differing prices for domestic and export markets, and, in particular, sales of milk under Special Classes 5(d) and (e), did not constitute a "subsidy" pursuant to the definition of the SCM Agreement, it followed that these sales could not constitute a subsidy for the purposes of the Agreement on Agriculture.184 Therefore, by definition, such sales could not constitute an "export subsidy" within the meaning of the definition in Article 1 of the Agreement on Agriculture. Canada further claimed that the practices at issue did not constitute an "export subsidy" within the meaning of any of the export subsidy practices described in Article 9.1 of the Agreement on Agriculture and, in particular, not within any of the practices cited by the Complainants.

(b) Article 9.1(a)

(i) The meaning of "direct subsidies, including payments-in-kind"

4.165.
New Zealand argued that, interpreted in accordance with the ordinary meaning of the terms of Article 9.1(a), in their context, and in the light of the object and purpose of the Agreement on Agriculture, Classes 5(d) and (e) of the Special Milk Classes Scheme constituted the provision by a government agency of a direct subsidy to an industry contingent upon export performance. This direct subsidy was provided through the foregoing of revenue or through a payment-in-kind within the meaning of Article 9.1(a) of the Agreement on Agriculture.
4.166.
New Zealand argued that Canada, in focussing on the word "subsidies" in Article 9.1(a) of the Agreement on Agriculture and interpreting it in isolation, removed it completely from its own context. The term used in fact in Article 9.1(a) was "direct subsidies, including payments-in-kind". There was no justification in the rules of treaty interpretation for taking words individually out of a phrase, giving them each a meaning and then reconstructing the phrase on the basis of those individual meanings. That was divorcing meaning from context completely.
4.167.
New Zealand argued that the normal usage of the term "payments-in-kind"was in respect of a payment in a form other than money, such as goods or services.185 The provision of a production input at no charge would clearly be a payment-in-kind. The provision of a production input (milk) at a reduced price was no less a payment‑in‑kind. New Zealand maintained that Article 9.1(a) referred specifically to "payments-in-kind" as included within the ambit of the concept of "direct subsidies". In the present case, government agencies made milk available to processors for export under Classes 5(d) and (e) at lower prices. This was the alternative Canada had chosen to providing a money sum to compensate processors for export for having to purchase milk at the higher domestic price. The benefit of access to lower‑priced milk was provided through the combined actions of the CDC and the provincial marketing agencies. Their actions made the provision of milk by producers at these lower prices mandatory. Through the operation of Classes 5(d) and (e), the government agency provided a subsidy through a "payment‑in‑kind" within the meaning of Article 9.1(a).
4.168.
New Zealand argued that the foregoing of revenue was well recognized as a form of subsidy,186 a common example being the foregoing of revenue through the remission of taxes. Thus, the ordinary meaning of the term "subsidy" in Article 9.1(a) included "revenue foregone". New Zealand noted that this was further illustrated in Article 9.2 which provided that in determining export subsidy commitment levels in Members’ Schedules "revenue foregone" was to be treated as a subsidy. Article 9.2 provided in sub‑paragraph (a)(i) that the "budgetary outlay reduction commitments" made by Members in respect of the subsidies listed in Article 9.1 shall in any year constitute the maximum level of expenditure for such subsidies. Article 1(c) of the Agreement on Agriculture defined "budgetary outlays" as including revenue foregone. Since revenue foregone was to be included in calculating levels of reduction commitments, it also had to be included in the concept of a subsidy for which reduction commitments were to be made. Thus, Article 9.2 made clear that the concept of "revenue foregone" was included within the scope of the subsidies listed in Article 9.1.
4.169.
New Zealand noted that this conclusion was confirmed by reference to the negotiating history of the export subsidy provisions of the Agreement on Agriculture. The de Zeeuw Text contemplated that states would table lists of "financial outlays and revenue foregone" in respect of subsidy practices.187 Similarly, the Modalities for the Establishment of Specific Binding Commitments under the Reform Program (the "Modalities document")188 in the context of export subsidy reduction commitments stated that "the expressions ‘outlays’ or ‘expenditure’ shall, unless the context otherwise requires, be taken to include ‘revenue foregone’."189 There had therefore been no doubt that revenue foregone was contemplated as a subsidy that would be subject to export subsidy disciplines.
4.170.
New Zealand maintained that its understanding of the interpretation of Article 9.1(a) was confirmed by reference to the preparatory work in the negotiation of that article. From the outset it had been understood that mechanisms to shield exporters from having to pay high domestic prices would be regarded as export subsidies. The text of the "Generic Criteria" produced as a basis for considering export competition issues in the negotiations spoke of any form of subsidy which resulted "in the sale of such products for export at a price lower than the comparable price charged for like products to buyers in the domestic market."190 The "Illustrative List of Export Subsidy Practices", set out to give specific content to that Generic Criteria, included what subsequently became Article 9.1(a).
4.171.
New Zealand argued that negotiating history made it clear that providing inputs at a lower price for export constituted a direct subsidy contingent upon export. The means by which the subsidised input was provided was not material. It could be through a transfer of money or it could be through revenue foregone. It could be viewed simply as the foregoing of revenue or it could be viewed as a "payment-in-kind". Regardless of the characterisation given, it constituted a direct subsidy captured by the terms of Article 9.1(a).
4.172.
New Zealand argued that the ordinary meaning of the terms of Article 9.1(a), read in the particular context of Article 9 as well as in the broader context of the Agreement on Agriculture and the WTO subsidies regime as a whole, was that a government‑mandated scheme whereby milk was made available by a government agency for the production of dairy products for export at prices that were lower than the prices for milk from the same agency for the production of comparable domestic products constituted a "direct subsidy" that was "contingent on export performance". Thus, the foregoing of revenue or the provision of a "payment‑in‑kind" by government agencies on milk provided to processors under Classes 5(d) and (e) was a subsidy within the meaning of Article 9.1(a) of the Agreement on Agriculture.
4.173.
New Zealand argued that it was the lack of choice between supplying the domestic or the export markets which lead producers to forego revenue (addressed in paragraph 4.93 and following). The decision to place milk in one "market" rather than the other was not made by the producer. The decision that the domestic market was satisfied and that, accordingly, milk had to be classified into Special Class 5(d) or 5(e) was made by government. Rational, profit-seeking producers, however, would - if they had the choice - supply their milk to the higher-priced domestic market (even although the influx of more milk into that market may ultimately have the effect of lowering prices). But the decision that they may not do so was made for them. Hence, the distinction between domestic and export markets was government-created: it was, indeed, a legal fiction created by Canada. With regard to in-quota milk, producers collectively forewent revenue by being forced to accept a lower in-quota price by virtue of export sales being pooled with higher-priced domestic sales. With regard to over-quota milk, individual producers forewent revenue by having no choice other than to accept the export price for that portion of their production which was ultimately deemed to be over-quota. Under the Special Milk Classes Scheme, the government of Canada obliged milk producers to forego the revenue they would otherwise have received from sales of milk at domestic prices in order to create an economic incentive for exporters to export. Producers were compelled to forego revenue and the benefit of this revenue foregone was passed on to exporters. Since both in-quota and over-quota milk were allocated to Class 5(e), the foregoing of revenue under Class 5(e) applied as much to over-quota as it did to in-quota milk.
4.174.
New Zealand noted that under the old producer levy-based system, producers received the same gross price for all milk produced (both in-quota and over-quota) but were forced, by government regulation, to forego revenue by virtue of a levy to subsidise the cost of exports. The situation was little different under the Special Milk Classes Scheme. Producers were forced to accept a lower price for milk that was subsequently exported. In the case of in-quota milk the revenue received by a producer was reduced by pooling. In the case of over-quota milk, revenue was not pooled and the price was determined by the CDC and the provincial milk marketing agency on the basis of world prices. New Zealand pointed out that the fact that world prices were the benchmark should not obscure the fact that the decision to place milk in one "market" rather than the other was not made by the producer.
4.175.
The United States argued that the Government of Canada, whether through the CDC or through the provincial governments, played a clear role in the establishment and administration of the Special Milk Classes Scheme. It was through the CDC that processors obtained a permit for preferentially priced milk for dairy products for export. In the absence of the federal and provincial government authority and legislation for the Special Classes, the processors would be paying the full price for milk.191 The processors would not be receiving milk at an artificially reduced price tailor-made by the CDC to allow them to make export sales.192 Thus the requirement under Article 9.1(a) that direct subsidies were provided by governments or their agencies was met.
4.176.
The United States argued that Canada’s construction of Paragraph 9.1(a) would make the reference to payments-in-kind meaningless. The Canadian argument was contrary to the principles of interpretation under customary international law and, in particular, those that required that the terms of an agreement be given effect, and that they be interpreted in good faith, in context and in light of their object and purpose.193 Article 9.1(a) covered "direct subsidies, including payments-in-kind". The ordinary meaning of the phrase "payments-in-kind" in the context of Article 9.1(a) was that the provision of artificially low-priced goods was to be regarded in the same way as straight cash. Accordingly, as Classes 5(d) and (e) provided milk at a reduced price contingent on the export of the manufactured product, the measure fell within Article 9.1(a) of the Agreement on Agriculture. It would be inconsistent with the ordinary meaning of the phrase "payment-in-kind" to suggest that the provision of goods without payment would be a subsidy, but that any level of payment, even though less than adequate remuneration, would not be an export subsidy. Moreover, the frequent statements by both industry leaders and Members of Parliament that the Special Classes would allow milk producers to share the "costs" of exports confirm that indeed the government was transferring value from the milk producers to dairy processors.
4.177.
The United States argued, in respect of revenue foregone, that the Canadian position was premised on the idea that under Canada’s milk marketing system producers could not sell milk into the domestic market that was designated for export as surplus milk. Yet, this division of markets into domestic and export segments was an artificial one and completely a construct of Canada. As New Zealand had stated, there were separate "markets" only because Canada had created a "special milk class" scheme and assigned the export of dairy products to Class 5. The United States recalled that milk was declared surplus to the domestic market in Canada pursuant to the discretion of the Canadian Dairy Commission. This was not a determination based on the operation of a free market. For instance, there was no determination of demand elasticities at different price levels. Instead, prices were maintained at rigid levels in the domestic market and if the market could not be cleared at a particular price level, there was not much latitude to reduce price to sell additional product domestically. Moreover, the declaration of a milk surplus was generally based on conditions within a province, usually without consideration of conditions in other provinces despite the fact that there was considerable movement of milk across provincial boundaries. The United States argued that Canadian consumers would use more milk if domestic prices were lower. Special Classes 5(a) through (c), offering lower priced milk to compete with certain imports, implicitly recognized this market principle. Milk sold at those lower Special Class prices allowed Canadian producers to capture additional sales. If that milk had instead been exported at the still lower Special Class 5(e) prices, there was no question but that the total revenue received by the milk producer would have been less, resulting in "revenue foregone".
4.178.
The United States noted that the same principle was demonstrated by the controversy over the substitution of imported butter-oil for butterfat in products such as ice cream. The high domestic milk prices in Canada had caused processors to look for alternative products for inputs in fat-rich products such as ice cream. If milk were to be sold at lower prices in Canada, milk would retain such product markets. It was for this reason that a proposal had been made to create an additional Special Class to allow Canadian milk producers to be more price competitive with imports of butter-oil. The Canadian International Trade Tribunal, in its report relating to butter-oil imports, considered the possibility of Canadian milk producers simply selling the milk that was displaced by imports of butter-oil into world markets. In conducting an analysis of the impact of such action on Canadian milk producers, the Canadian Tribunal described the effects in terms of "revenue foregone" by the Canadian industry.194 Thus, this concept was not unfamiliar to the CITT, with its considerable knowledge of the Canadian milk marketing system.
4.179.
Canada reiterated that sales of milk at differing prices did not constitute a "subsidy" as it was defined in Article 1 of the SCM Agreement. As the Agreement on Agriculture and the SCM Agreement had to be taken together, the definition of "subsidy" in the SCM Agreement was applicable to the term "subsidy" as it was found in Article 9.1(a).195 It followed that there was no "subsidy", whether direct or not, for the purposes of Article 9.1(a). (Canada's detailed arguments on the application of the SCM Agreement definition of "subsidy" to the measures in question are summarized beginning at paragraph 4,309.)
4.180.
In respect of the matter of payments-in-kind, Canada noted that the United States argued that Canada’s interpretation of Paragraph (a) would make the reference to payments-in-kind meaningless and New Zealand characterized a sale of dairy inputs at a reduced price as a payment-in-kind. Canada submitted that the Complainants had not clearly articulated what amounted to a payment-in-kind. Canada’s position did not seek to make that expression meaningless but rather to give the expression its ordinary meaning. A payment-in-kind arose when a debt was satisfied by the provision of a good or a service rather than being paid for in money. For example, a government could impose a 5 per cent royalty with respect to a concession to drill for oil. If the government permitted that obligation to be discharged by the delivery to it of one barrel of oil for every twenty barrels extracted, that would be a payment-in-kind. In the case of the Canadian dairy system, payment was made for milk in the ordinary sense of the word. A market-based differential between payments was qualitatively different from a payment-in-kind.196
4.181.
Canada noted that the Complainants sought to find a subsidy in Article 9.1(a) by claiming that there was "revenue foregone" (paragraphs 4,173 and 4,177). Canada submitted that even if the word "payments" were held to include "revenue foregone", there was no revenue for the producers to forego with respect to sales of milk for export use under Special Classes 5(d) and (e). In the context of commercial sales, revenue was foregone when the vendor chose to sell the product at a price lower than the price at which the vendor could have otherwise sold it. In other words, if a vendor chose to forego a sale into a higher-priced market in favour of a sale into a lower-priced market, then the vendor had chosen to forego revenue.197
4.182.
Canada argued that "revenue foregone" implied a choice of markets, a choice foregone. Under the Canadian milk marketing system, milk could not be sold in the market for export uses if it was required for Canadian domestic requirements. Thus, sales of milk for export purposes at prices based on world market prices could not be made until there was no opportunity to sell milk into domestic markets at the higher domestic prices. This was a basic and rational approach in any commercial operation. Any milk sold for export uses was additional to the domestic demand. The sale of milk at world market prices - not always lower than domestic prices - therefore took place when milk producers in Canada could no longer place their products on the domestic market, and had therefore to try to obtain the highest prices possible in the alternative market, i.e., the market in Canada for milk for use in exports. Suggestions that Canadian milk producers somehow "forewent" revenue otherwise available to them in the domestic market demonstrated a serious misunderstanding of the Canadian milk marketing system.
4.183.
Canada argued that rather than representing "revenue foregone", such sales represented revenue enhancement. Canadian producers could choose not to produce any milk in addition to domestic requirements. Accordingly, they could choose not to produce milk for use in export sales. They could choose to limit their revenues to the returns they would get from the domestic markets. Instead, by choosing to produce milk in addition to domestic needs, the returns from which would be based on world market prices, producers chose to try to enhance their total revenues. As a result, the only reasonable approach was indeed the approach adopted by the milk producers in Canada; to find a market in which the highest return was found for milk at any particular time. Thus, in the sale of milk at world market prices, revenue was not foregone but rather enhanced.
4.184.
Canada argued that the fact that there was a domestic market and an export market was not a fiction but a fact of life. Canada rejected New Zealand’s characterization that producers were "forced" by government to forego their own revenue. The record showed that the producers had collectively and individually chosen to market their product in the manner reflected by the present regime. New Zealand failed to explain how producers making these collective and individual decisions were "foregoing revenue" in the sense understood by Canada and the United States. In addition, it remained Canada’s position that there was no revenue foregone, even to producers, when a person willingly sold a product in a market for the price that the market would bear for that product.
4.185.
Canada noted that New Zealand built arguments on the basis of the Uruguay Round negotiating document entitled: Modalities for the Establishment of Specific Binding Commitments under the Reform Program (paragraph 4,169). Canada pointed out that the following injunction was found on the cover sheet of the document:

"The revised text is being re-issued on the understanding of participants in the Uruguay Round that these negotiating modalities shall not be used as a basis for dispute settlement proceedings under the WTO Agreement".

4.186.
New Zealand argued – in respect of Canada's argument that the Special Milk Classes Scheme did not constitute a "payment-in-kind" within the meaning Article 9.1(a), because a payment-in-kind arose when a debt was satisfied by the provision of a good or a service rather than being paid for in money (paragraph 4,194) – that the suggestion that the existence of a debt was a prerequisite to any notion of payment-in-kind would render that concept, in the context of subsidies, completely redundant. If a debt was owed, the payment of it, or its discharge through a payment-in-kind, could not constitute a subsidy. It would simply be the re-payment of a debt. To the extent that Canada was suggesting that a payment-in-kind could never be a subsidy and could only be used in the context of the satisfaction of a debt, it was seeking to rewrite Article 9.1(a) to exclude the concept of "payment-in-kind" completely.
4.187.
Since Canada’s arguments that the Special Milk Classes Scheme did not meet the definition of the term "subsidy" and hence could not be a subsidy within the meaning of Article 9.1(a) of the Agreement on Agriculture were unfounded (paragraph 4,179), New Zealand’s arguments on the applicability of Article 9.1(a) remained unanswered by Canada.

(ii) The meaning of the term "direct"

4.188.
New Zealand argued that the term "direct" in relation to subsidization was used in a variety of senses and its meaning in any particular case had to be derived from the context in which it was used. In the context of subsidies under the GATT 1947, the word "direct" bore a particular meaning which was articulated in the negotiation of Article XVI of the GATT 1947. The term "indirectly" in Article XVI, it was explained, made it clear that "subsidization" could "not be interpreted as being confined to subsidies operating directly to affect trade in the production under consideration."198 In other words, a "direct" subsidy was one affecting trade in the product directly rather than one affecting trade incidentally or indirectly. In the present case, the subsidy provided was clearly "direct". Revenue was being foregone for the very purpose of affecting the trade in question - indeed of permitting its very existence. Without this foregoing of revenue no export trade would exist.199 The objective of this foregoing of revenue was to secure export performance. It was not a subsidy that had other objectives with export performance as an incidental effect; it was a subsidy that provided only for products which were destined for export. It was a direct subsidy contingent on export performance, and thus fell within the ordinary meaning of the words of Article 9.1(a).
4.189.
New Zealand argued that in Article 9.1(a) the term subsidies had to be interpreted in the light of the fact that this provision was referring to "direct" subsidies and in view of the fact that the category of direct subsidies had to be large enough to include "payments-in-kind." This alone was sufficient to demonstrate that a simple reliance on the definition of the term "subsidy" in Article 1 of the SCM Agreement was inadequate for the interpretation of Article 9.1(a). As this was ignored by Canada, Canada had failed to interpret Article 9.1(a) properly and thus had failed to show that the Special Milk Classes Scheme did not constitute an export subsidy within the meaning of Article 9.1(a). Accordingly, Canada had not discharged the burden of proof placed on it under Article 10.3 of the Agreement on Agriculture (paragraphs 4,290 and following refer).
4.190.
Canada refuted the broad interpretation of the term "direct" in Article 9.1(a) given by the Complainants. Canada rejected any necessity to enter into such an exercise and referred the Panel to Article 1.1(a)(1) of the SCM Agreementwhere "direct" was used to qualify a "subsidy" implying a transfer of funds by a government. Canada reiterated that no direct export subsidy existed in Canada for dairy products.

(c) Article 9:1(c)

(i) The meaning of the term "payment"

4.191.
New Zealand noted that the word "payment" was defined in the Oxford English Dictionary as: "1. the action, or an act of, paying; the remuneration of a person with money or its equivalent … 2. a sum of money (or other thing) paid;... ".200 The Dictionary of Canadian Law defined "payment" as "remuneration in any form."201 The term "remunerate" was defined in the Oxford English Dictionary as "1. To repay, requite, make some return for (services etc) … 2. To reward (a person) … to pay (one) for services rendered … 3. to give as compensation... ".202 New Zealand argued that the term "payments" covered both payments-in-kind and revenue foregone. The ordinary meaning of the term payment included a "payment-in-kind". Adding "in-kind" to the word "payment" simply described the form in which a payment was made. In respect of Article 9.1(a), the provision to exporters of lower-priced milk was a "payment-in-kind". It was the delivery of something of value in a form other than by way of a money transfer.
4.192.
The United States argued that the common meaning of the word "payment" was "the action, or an act, of paying; the remuneration of a person with money or its equivalent; the giving of money, etc. in return for something in discharge of a debt". The word was also defined as "a sum of money (or other thing) paid; pay, wages; or price".203 The verb "to pay," from which the noun "payment" was derived, was variously defined as "to give what is due, as for goods received; remunerate; recompense; to give or return as for goods, or services; to give or offer." However, "to pay" had also been construed as meaning "to give money or other equivalent value for; to hand over the price of a (thing); to bear the cost of; to be sufficient to buy or defray the cost of".204Thus, although the word payment often connoted an exchange of value for the provision of goods or services, or the provision of value on the occasion of a particular event or condition, it could also encompass bearing the cost. This latter meaning was perhaps the most consistent with the word’s use in the context of a provision defining subsidies. Thus, the term payment could be used in Article 9.1 consistent with the concept of conferring a benefit through the bearing of a cost.205
4.193.
In respect of the term "payment", Canada refuted that its ordinary meaning included "revenue foregone". The ordinary meaning of the term "payments" was straightforward – it meant "a sum of money."206 Canada noted that, pursuant to Article 33 of the Vienna Convention, the French language version of the text of this provision provided additional support for its interpretation of the word "payment". The term used in the French text was "versement" which meant literally to remit money.207
4.194.
Canada recalled that while the United States acknowledged that the common meaning of payment was "the action, or act of paying; the remuneration of a person with money or its equivalent; the giving of money, etc. in return for something in discharge of a debt" (paragraph 4,192), wishing to argue that the ordinary meaning of the word payment included "revenue forgone" and realizing that this "common meaning" was not supportive of their position, the United States had attempted to find an alternative meaning through the verb "pay", pointing to the eleventh listed definition which includes the phrase "bear the cost". This was a frail argument when it was considered that this was found within the eleventh definition of a related word, and did not appear at all the most recent New Shorter Oxford English Dictionary.208"Payments" had to be interpreted in its "ordinary meaning": i.e. to reflect an action of paying something of value.
4.195.
Canada noted that New Zealand as well encountered difficulty in establishing that the "ordinary meaning" of the word "payment" included "revenue forgone". New Zealand noted that the Oxford English Dictionary definition of "payment" included "remuneration of a person with money or its equivalent" and suggested that the reference to "remunerate" indicated that the term was to be read broadly. Canada argued that there was a major difference between suggesting that payments could be made with a wide variety of items of value, i.e., from actual cash to payment-in-kind, to concluding that the ordinary meaning of "payment" included an indirect result such as "revenue foregone".
4.196.
In New Zealand's view the dictionary definitions referred to above, indicated that the concept of "payment" had a wide ambit. New Zealand noted that in order to determine how the term payment was being used in the specific case of Article 9.1(c), reference had be made to the context in which the term was used and the object and purpose of the Agreement on Agriculture as a whole.209 New Zealand noted that Article 9.1(c) was contained in a provision that identified the mechanisms that states had used to provide export subsidies and which were subject to reduction commitments. One mechanism the negotiators of Article 9.1(c) had in mind was the use of producer levies to fund payments to exporters to compensate for the high cost of a product purchased at domestic rather than at world prices. Such payments were referred to specifically as being included in the definition of "payments on the export of an agricultural product." The wording of Article 9.1(c) made clear that it had not been intended that its provisions be limited only to money paid from the proceeds of a producer levy.210 An examination of the context in which the word "payments" in Article 9.1(c) appeared confirmed that in the light of the object and purpose of the Agreement on Agriculture as a whole, the term "payments" covered both revenue foregone and payments-in-kind.
4.197.
New Zealand emphasized that as Article 9.2 included revenue foregone within the determination of budgetary outlay commitments to be made with regard to the subsidies listed in Article 9.1, the concept of "payments", in Article 9.1(c), had to include "revenue foregone" as they had to be quantified under the heading of "budgetary outlays", and this included, explicitly, revenue foregone. There was no need to provide specifically that "payments" included revenue foregone because the definition of "budgetary outlays" already carried that implication.
4.198.
The meaning of the phrase "payments on the export of an agricultural product" in Article 9.1(c) of the Agreement on Agriculture must, however, be distinguished from the term "the payment of subsidies exclusively to domestic producers" under Article III:8(b) of the GATT 1994. In Canada - Certain Measures Concerning Periodicals the Appellate Body stated that:

" … an examination of the text, context, and object and purpose of Article III.8(b) suggested that it was intended to exempt from the obligations of Article III only the payment of subsidies which involves the expenditure of revenue by a government."211

4.199.
Accordingly, New Zealand noted that the Appellate Body concluded that a reduction in postal rates did not constitute a "payment of a subsidy exclusively to a domestic producer" within the meaning of GATT Article III:8(b). In reaching this conclusion the Appellate Body was influenced by the fact that Article III:8(b) was an exception to the national treatment obligation. In its view, it was never intended that exceptions to national treatment by way of subsidies based on tax reductions or other forms of revenue foregone were to be permitted under Article III:8(b).
4.200.
In fact, New Zealand maintained that the rationale of the Canada - Periodicals decision reinforced the conclusion that the term "payments on the export of an agricultural product" in Article 9.1(c) had to include revenue foregone. The objective of the list in Article 9.1 was to bring agricultural export subsidies under WTO disciplines. An interpretation of Article 9.1(c) that narrowed the scope of the subsidies included therein to direct money transfers would defeat rather than serve the object and purpose of Article 9. It would expand the opportunity for Members to avoid their WTO obligations - precisely what the Appellate Body in Canada - Periodicals was seeking to avoid. Such a conclusion was strengthened when the object and purpose of the export competition provisions of the Agreement on Agriculture as a whole were considered. Revenue foregone, which was simply an alternative way of securing a benefit that could be obtained through the direct transfer of money, had to be included in the concept of "payment" under Article 9.1(c) if the progressive reduction of export subsidies through reduction commitments was to be successful.
4.201.
New Zealand noted that under the old producer levy system, the CDC, acting in concert with provincial milk marketing boards or agencies, transferred money to exporters to compensate for the cost of processors purchasing milk for products for export at domestic rather than at world prices. Under Classes 5(d) and (e) of Special Milk Classes, processors were permitted to purchase milk for products for export at world rather than at domestic prices. The difference between the two approaches was one of form only. In each case, the processor for export was being shielded from the high domestic cost of milk. In each case, the processor for export was being provided with a subsidy that was captured by the phrase "payments on the export of an agricultural product" in Article 9.1(c) of the Agreement on Agriculture. Under Classes 5(d) and (e) revenue was foregone by provincial milk marketing boards or agencies providing access to milk from producers to processors at "special class" prices. The provincial milk marketing board or agency forewent the revenue that it would have received if the milk had been sold at domestic prices.
4.202.
New Zealand contended that although the provincial milk marketing board or agency was the vehicle for providing the special milk classes subsidy, it was the producer who bore the financial cost. The role of the provincial board or agency was one of a conduit - to pass on to the producer through the pooling arrangements the revenue that results from Special Milk Class sales. In fact, what the agency passed on to the producer were the losses that resulted from sales of milk at world market prices. It was the producer who forewent the revenue that would have been received if all milk was sold to processors at domestic prices. The Special Milk Classes Scheme shrouded in complexity the obvious fact that it was the producer who made the "payments on the export of an agricultural product" that brought the scheme within Article 9.1(c).
4.203.
New Zealand claimed that, in substance, the revenue foregone by producers "on the export of an agricultural product" was the equivalent of a subsidy provided to such processors by a direct money transfer financed from the proceeds of a levy on "an agricultural product from which the exported product is derived" - a form of subsidy that Article 9.1(c) expressly enjoined. Hence, the Special Milk Classes Scheme involved an elaborate structure for a very simple subsidy. It consisted of the foregoing of revenue on the export of an agricultural product. That foregoing of revenue could be viewed as a foregoing by a provincial milk marketing board or agency, or it could be viewed as the foregoing of revenue by producers. The difference between the two was largely a matter of accounting. The form could differ, but the substance remains the same.
4.204.
As noted under Article 9.1(a), New Zealand argued that Classes 5(d) and (e) of the Special Milk Classes Scheme could also be characterised as providing payments‑in‑kind - something that was equally encompassed in the ordinary meaning of the term "payment". A payment‑in‑kind involved remuneration through something other than, or something equivalent to, money. The provision of goods (milk) at a reduced price, instead of providing a money sum to compensate for the higher price that would be paid for milk for processing into products destined for the domestic market, was a payment‑in‑kind. It was a direct substitute for a payment by way of money transfer. And that, of course, was its intent. It should be seen as a substitute for the old "money-transfer" subsidies paid by Canada from producer levies. It was a "payment-in-kind" that fell within the concept of "payments on the export of an agricultural product" under Article 9.1(c).
4.205.
The United States claimed that like the producer levy programme it replaced, the Special Milk Classes Scheme was an export subsidy within the meaning of Article 9.1(c) of the Agreement on Agriculture. By the express terms of that sub-paragraph, there was no requirement that an export subsidy be a charge on the public account. By way of example, the Article specified that payments subject to its coverage may be "financed from proceeds of a levy imposed on the agricultural product concerned or on an agricultural product from which the export product is derived." As argued under Article 9.1(a), the United States claimed that the term "payment" was broad enough to include instances in which value was given by some other means than the actual transfer of funds (paragraph 4,192). If this were the case, then a fortiori "payment" included situations where value was given to another by means such as a product, in this case milk, at less than the market price.
4.206.
The United States noted that the term payment was used twice in Article 9.1(c), but was not defined there or elsewhere in the Agreement on Agriculture.212 The United States noted that the Vienna Convention counselled that the ordinary meaning of a term was to be given in light of its context, which, in this case, was the Agreement on Agriculture and, more specifically, that Agreement’s provisions governing export subsidy disciplines. The meaning of the term "payment" had also to be fixed by considering the object and purpose of the pertinent treaty. Article 9.1, as a whole, broadly identified the subsidies, including payments within Article 9.1(c), that were to be taken into account in calculating the maximum level of expenditure for export subsidies that a Member could incur in a given year consistent with its Schedule of Concessions and Commitments and Articles 3 and 8 of the Agreement on Agriculture. This level constituted the Member’s budgetary outlay reduction commitment. The term "budgetary outlays" was defined in Article 1(e) of the Agreement on Agriculture to include "revenue foregone." Thus, for example, not only direct subsidies, such as those described in Article 9.1(a), but also reduced charges, i.e., revenue foregone, as described in Paragraph 1(c), counted toward the budgetary outlays that were subject to the reduction commitments, whether or not charged to the public account.
4.207.
The United States argued that in this context, and given that the purpose of Part V of the Agreement was to impose discipline on export subsidies, the term "payment" had to be construed consistently with the broad meaning given to budgetary outlays. If such outlays, and thus the applicable reduction commitments overall were expressly defined to include revenue foregone, then one consistent construction was to construe "payments" in a similar manner. Such an interpretation was consonant with the purpose of the Agreement to bring export subsidies under the transitional disciplines established by the reduction commitments. Considered from this perspective, if the total expenditures subject to reduction commitments were defined as total budgetary outlays and revenue foregone, then the individual expenditures had logically to comprise all payments, including price reductions that had the same economic effect as an export rebate. By mandating the sale of industrial milk at a discount, the Government of Canada was conferring a benefit to milk processors equivalent in its trade distorting effect to an export rebate. Moreover, the principle that an export subsidy could be accorded in the form of sales of products at a loss, or by offering goods or services for export at a more advantageous price than when offered for sales for domestic consumption, was not only inherent in the Agreement’s definition of budgetary outlays, but was also reflected in the provisions of Article 9.1(a), (b) and (e) which defined "payments-in-kind," sales of non-commercial stocks, and discounted transport and freight charges as forms of subsidization.213
4.208.
In addition, the United States noted that the context of Article 9.1(c) also included the related subsidy reduction commitments contained in Articles 3, 8, and the remainder of 9 of the Agreement on Agriculture. These provisions were collectively intended to impose meaningful disciplines on the use of export subsidies. The Agreement on Agriculture thus narrowed the universe, and amount, of potential subsidies in several respects. First, Article 3 specified that no export subsidies were to be provided "in respect of any agricultural product not specified" in a Member’s schedule. A Member could not introduce subsidies for products that were not identified in its schedule and which had not been subsidized during the pertinent base period. Furthermore, Article 9.1 referenced a broad cross-section of subsidy practices that were intended to capture the agricultural subsidies used by the Members at the time of negotiation of the Agreement. Such subsidies were made subject to reduction commitments pursuant to Articles 3, 8 and 9 of the Agreement, with significant reductions required in both subsidy outlays and the quantity of exports that benefitted from subsidies to occur during the transition period. In addition, Article 10, to protect Article 9 disciplines on export subsidies, prohibited the introduction of any subsidies not listed in Article 9 that either "results in, or which threatens to lead, to circumvention of export subsidy commitments... " And finally, Article 3 of the SCM Agreement, when read in conjunction with Article 13 of the Agreement on Agriculture, provided that if a Member did not comply with the export subsidy disciplines contained in the Agreement on Agriculture, any offending export subsidy was to be subject to the terms of the SCM Agreement and its prohibition on export subsidies.214
4.209.
Hence, the United States argued that given the Agreement on Agriculture’s comprehensive treatment of export subsidies, which revealed the Members’ intent to establish real and effective disciplines respecting export subsidies, a narrow construction of the term "payment" that would result in a weakening of the export subsidy reduction commitments and disciplines, would be contrary to the over-arching objective and purposes of the relevant treaty provisions as a whole.
4.210.
In this regard, the United States noted that Professor Tangermann also compared the pernicious effect of producer-financed export subsidies with price pooling by state export agencies and concluded that the latter should be subject to the same export competition disciplines:

"Where a state agency sells domestically at a price above the price charged for exports, while domestic producers are paid the average price, exports are implicitly subsidized. To see why, it is best to compare this policy to one of producer-financed export subsidies. In the latter case, a levy is charged on the domestic sales, and the proceeds are then used to finance export subsidies. Under a price pooling regime, the same prices can result, and it is only the technical nature of financial flows which is different, but not the economic result. Hence, price pooling differs from producer-financed export subsidies only in form, not in substance. Countries should, therefore, not be allowed to escape their export subsidy commitments by using a price pooling regime."215

4.211.
The United States argued that in construing the text of Article 9.1(c), it was clear that the reference to the levy financed export rebates was made for purposes of illustration, and did not limit the scope of Article 9. Thus, the language in Article 9.1(c) relating to the "payments that are financed from the proceeds of a levy" commenced with the introductory word "including", indicating that payments so financed constituted only one example of the types of payments which fell within the scope of the Article 9 subsidy disciplines. There was nothing in the text of the paragraph to provide a basis for concluding that other types of producer-financed funding for export payments, such as discounted prices, would be excluded from the subsidy constraints imposed by Article 9.1(c). To the contrary, subsidies that were the functional equivalent of producer-financed levies had to be assumed to be included in Article 9.1(c) disciplines.
4.212.
The United States argued that Canada’s Special Milk Classes Scheme differed from the producer levy programme that preceded it only in form, not substance. Revenue foregone (on export sales under the Special Classes) by the milk producers, which translated into discounted prices for dairy product manufacturers, was equivalent to the export rebates paid to such manufacturers under the levy system. To exclude such discounted milk from the coverage of Article 9.1(c) because the benefit or payment received by the dairy product exporter was in the form of a lower milk price, rather than in the form of an export rebate contingent on export of a product in which the milk had been used, would elevate form over substance. Hence, the Special Milk Classes Scheme, like the producer levy programme that it replaced, qualified as "payments on the export of an agricultural product" within the meaning of Article 9.1(c). First, the discounted milk prices provided for in Special Milk Classes 5(d) and (e) were available only in connection with the production of dairy products for export and, thus, were provided "on the export of an agricultural product". Second, the lower prices extended to milk processors, contingent on the use of the milk for production of dairy products for export, were the same both in substance and economic effect as the earlier levy-financed export rebates and, therefore, were likewise a "payment" within the meaning of that term as used in Article 9.1(c) of the Agreement.216
4.213.
Canada submitted that the sales of milk under Special Classes (d) and (e) did not constitute an export subsidy within the meaning of Article 9.1(c) as it could not be shown that "payments" were made on the export of products from Canada.
4.214.
Canada argued that, in accordance with the Vienna Convention, the interpretation of the text of WTO agreements had to proceed on the following basis: (i) the starting point was always the actual text and its ordinary meaning; (ii) the ordinary meaning of the terms of the agreement was to be read in their context; (iii) "context" had to comprise the text of the agreement, including other contemporaneous agreements reached by all the parties, as outlined in Article 31.2 of the Vienna Convention, and (iv) recourse could only be had to supplementary means of interpretation, such as negotiating history, in the event of ambiguity or to confirm a meaning.
4.215.
Canada noted that New Zealand argued that support for the proposition that payment should include revenue forgone could be found the Canada - Periodicals case (paragraph 4,198 and following). In that case, the Appellate Body had ruled that the term "payment" as it appeared in Article III:8(b) of the GATT 1994 did not include a reduction of postal rates since Article III:8(b) was an exception and should be interpreted narrowly. New Zealand suggested that on the basis that Article 9.1 was a positive obligation the reverse should hold true: its terms should be interpreted broadly. Canada submitted that this misconstrued the proper interpretive approach. The fundamental rule was that of "ordinary meaning" under Article 31 of the Vienna Convention. In the case of an exception, the exceptional approach of a narrow approach was applied. Absent an exception, the interpreter had to revert to the fundamental rule of "ordinary meaning".
4.216.
Canada argued that it was also noteworthy that in Article 1 of the SCM Agreement, the drafters were careful to add a provision covering "government revenue that is otherwise due is forgone", they had realized that the ordinary meaning of the preceding provision relating to "direct transfers of funds" would not be sufficient to cover "revenue forgone".
4.217.
Canada refuted the allegations that the interaction between Article 9.2 and Article 9.1 implied that the defined meaning of "budgetary outlays", i.e., including "revenue forgone" was applicable to "payment" in Article 9.1(c). This would be to set aside the ordinary meaning of a word based on an indirect inference and notwithstanding the clear contextual confirmation provided by Annex 2.
4.218.
Canada noted that in United States - Reformulated Gasoline, the Appellate Body undertook a comparison of the words used in different paragraphs of Article XX to determine the meaning of the test set out in Article XX(g). The Appellate Body noted that different words were used in the different paragraphs so as to properly describe the required relationship or degree of connection between the objectives in question and the measures implemented: "necessary" in certain paragraphs, "relating to" in certain others, "for the protection of" in another, and so on. Accordingly, the Appellate Body held that:

"It does not seem reasonable to suppose that the WTO Members intended to require, in respect of each and every category, the same kind of degree of connection or relationship between the measure under appraisal and the state interest or policy sought to be promoted or realised."217

4.219.
Hence, Canada argued that it could be presumed that specific words carried specific meanings and that the use of different words in a treaty text meant that the parties to the treaty intended different meanings to be applied to those differing terms. The approach to the interpretation of the term "payments" in Article 9.1(c) in the context of Article 9.1 as a whole should be no less rigorous. Canada argued that the choice of different terminology in each provision of Article 9.1 clearly indicated that negotiators had very precise and distinct concepts in mind with respect to each provision. In the light of the terms used in the other provisions, the selection of the word "payment" in Article 9.1(c) indicated an intention to apply a precise and limited ambit to the application of Article 9.1(c). Accordingly, the term "payment" had to be given its ordinary meaning and not be construed so as to encompass practices that could be covered by the use of broader terms such as "subsidies".
4.220.
Canadaargued that the negotiating history of Article 9.1 demonstrated that with respect to Article 9.1(c) in particular, the course of the agriculture negotiations had been to narrow the scope of that provision. Thus, the essential issue was what the parties agreed to at the end of the negotiations. In the context of the Agreement on Agriculture, this very point was made in one of the exhibits relied upon by the United States (paragraph 4,210).218 Therein, Professor Tangermann identified production quotas with over-quota output sold at world market prices and price pooling arrangements as matters not presently covered under the Agreement on Agricultureand thus were subjects to be pursued in future negotiations.219 Professor Tangermann recognized that the proper way to reconcile the divergent economic theories was by multilateral negotiation, not negotiation through litigation. Canada argued that it was apparent that Professor Tangermann viewed the results of the Uruguay Round in agriculture as having an unsatisfactory economic result. While Professor Tangermann was entitled to his opinion as to the outcome he would consider to be desirable220, the Agreement on Agriculture was not concerned with any particular economic theory but rather with specific legal obligations that were agreed to by Members.
4.221.
Canada argued that if, indeed, the intention of the negotiators had been to stretch the meaning of "payment" beyond its ordinary meaning, this would have been provided for specifically. For example, the terms "budgetary outlay" and "outlay" had been defined in Article 1 of the Agreement on Agriculture as including "revenue foregone". Payment, however, which did not incorporate "revenue forgone" in its ordinary meaning, was not similarly defined. Confirmation of this could be found in the treatment of the word "payment" as it was used elsewhere in the Agreement on Agriculture. "Payment" appeared prominently in Annex 2 with respect to domestic support. In Article 5 of Annex 2 the reference was to "payments (or revenue forgone, including payments in kind)". Clearly, the drafters were aware that the ordinary meaning of payment did not include "revenue forgone" and specific provision for its inclusion would be required. The absence of any counterpart in Article 9.1 could only lead to the conclusion that in the case of Article 9.1(c) the drafters intended the ordinary meaning of "payment" to stand.
4.222.
Canada further claimed that the negotiating history of the Agreement on Agriculture also supported Canada's submission that the term "payment" had to be construed precisely. Article 32 of the Vienna Convention permitted recourse to supplementary means, including the travaux préparatoires, to support the interpretation arrived at under Article 31.
4.223.
Canada noted that Annex 7 to the Draft Dunkel Working Papers dated 21 November 1991 provided a proposed list of measures that would be deemed to be "export subsidies" for the purposes of reduction commitments.221 As such, this list, which reflected earlier draft texts circulated in the summer of 1991222, was a precursor to the eventual Article 9.1 in the Agreement on Agriculture. In particular, Article 3(k) of this draft was an early version of the text that would ultimately become Article 9.1(c). This paragraph referred to "subsidies", not "payments". The text of Paragraph 1(c) continued to refer to "subsidies" in the text of the 12 December Dunkel draft.223 On 17 December 1991, Canada submitted to Arthur Dunkel a number of specific redrafting proposals on the agriculture text. Amongst these was an amendment to Paragraph 1(c) so as to substitute the word "payments" for the word "subsidies".224 The text of Article 9(1) of the "Draft Final Act" of 20 December 1991 reflected this change and referred to "payments".225 This drafting history confirmed that the term "payments" was specifically and deliberately selected in place of a broader term. Accordingly, the word "payment" in Article 9(1)(c) had to be interpreted strictly in accordance with its ordinary meaning.
4.224.
Moreover, Canada recalled its argument as set out under Article 9.1(a) that even if the word "payments" were held to include "revenue forgone", there was no revenue for the producers to forego with respect to sales of milk for export use under Special Classes 5(d) and (e) (paragraph 4,181 and following).