B. Article 6 of the SPS Agreement

106. The next issue we address is whether the Panel exceeded its terms of reference by extending the scope of its examination of Article 5 to include Article 6 of the SPS Agreement.

107. In the request for the establishment of a panel, Canada requested that the Panel consider and find that:

(a) The Australian Measures are Inconsistent with:

(b) The Application of the Australian Measures Nullifies or Impairs the Benefits Accruing to Canada Pursuant to the Agreement Establishing the World Trade Organization.⁵¹

108. In the context of examining the third element of Article 5.5, i.e., the existence of discrimination or a disguised restriction on international trade resulting from the arbitrary or unjustifiable distinction in the levels of sanitary protection in different situations, the Panel took into account three "warning signals" and three "additional factors".⁵² The third of these "additional factors" concerns the Panel's doubts on whether Australia applies "similarly strict" sanitary standards to the internal movement of salmon within Australia as it does on imports of salmon.⁵³ In the context of its brief and inconclusive discussion on Australia's internal controls on the movement of salmon, the Panel noted:

109. In a footnote to this paragraph, the Panel referred to Article 6.1 of the SPS Agreement and stated:

110. Canada's request for the establishment of a panel did not include a claim of violation of Article 6 of the SPS Agreement. The Panel's terms of reference are determined by Canada's request for the establishment of a panel. We, therefore, agree with Australia that Article 6 of the SPS Agreement is not within the terms of reference of the Panel. However, we disagree with Australia that the Panel exceeded its terms of reference in quoting Article 6.1 in a footnote, attached to a paragraph in which the Panel examined a violation of Article 5.5. More precisely, we reject Australia's contention that the Panel, by merely referring to Article 6.1 in a footnote, made an implied finding of inconsistency with Article 6. In our view, the statement of the Panel with regard to Article 6, in footnote 430 of its Report, is similar in character to the statement of the panel in United States - Shirts and Blouses, with regard to the powers of the Textile Monitoring Body ("TMB"). India appealed from this statement, but we found it to be "purely a descriptive and gratuitous comment providing background concerning the Panel's understanding of how the TMB functions".⁵⁶ We did not consider that statement to be "a legal finding or conclusion" which the Appellate Body "may uphold, modify or reverse". Likewise, we consider that in this case, the Panel's statement in footnote 430 of its Report regarding Article 6.1 of the SPS Agreement is a purely gratuitous comment and not "a legal finding or conclusion". By making such a comment, the Panel did not exceed its terms of reference.

V. The SPS Agreement

A. Preliminary Observation

111. The Panel made separate analyses and findings on ocean-caught Pacific salmon and on other Canadian salmon. With respect to other Canadian salmon, the Panel found that Australia, by maintaining the measure at issue, without even purporting to conduct or rely on a risk assessment, has acted inconsistently with Articles 5.1 and 2.2 of the SPS Agreement.⁵⁷ This finding is not appealed. The Panel limited its further examination and findings of violations of Articles 5.1, 5.5, 5.6, and the consequential violations of Articles 2.2 and 2.3 of the SPS Agreement, to ocean-caught Pacific salmon.⁵⁸ The following three sections (B, C and D) of our Report, which examine Australia's appeal against these findings, are therefore also confined to ocean-caught Pacific salmon. Canada's appeal from the Panel's lack of findings relating to its claims under Articles 5.5 and 5.6 with respect to other Canadian salmon is examined in Section E. Canada's appeal from the Panel's lack of a finding on its independent claim under the first sentence of Article 2.3, for both categories of salmon, is examined in Section F.

B. Article 5.1 of the SPS Agreement

112. We first address the question of whether the Panel erred in finding that the measure at issue, as it applies to ocean-caught Pacific salmon, is not based on a risk assessment and that Australia, therefore, has acted inconsistently with Article 5.1 and, by implication, Article 2.2 of the SPS Agreement.

113. The Panel first addressed the issue of whether the 1996 Final Report is a risk assessment within the meaning of Article 5.1 of the SPS Agreement. It did not make a finding on this issue but assumed, for the purpose of its further examination, that the 1996 Final Report meets the requirements of a risk assessment. It subsequently examined whether the SPS measure at issue, which, in its view, is the heat-treatment requirement, is based on the 1996 Final Report, "the only risk assessment forwarded by Australia".⁵⁹ In paragraph 8.99 of its Report, the Panel came to the conclusion that "the measure at issue - in so far as it applies to ocean-caught Pacific salmon - is not 'based on' a risk assessment," and that, therefore, Australia has acted inconsistently with Article 5.1 and, by implication, Article 2.2 of the SPS Agreement.⁶⁰

114. Australia appeals from the Panel's finding that the measure in dispute was not based on a risk assessment. In the view of Australia, this finding is vitiated by the Panel's conclusion that the SPS measure in dispute is the heat-treatment requirement rather than the import prohibition.

115. We recall that we have found that the SPS measure at issue is the import prohibition on fresh, chilled or frozen salmon, not the heat-treatment requirement.⁶¹ The import prohibition is, therefore, the measure which must be examined under the SPS Agreement. More particularly, under Article 5.1, the Panel should have determined whether the import prohibition on fresh, chilled or frozen salmon was based on a risk assessment, instead of examining whether the heat-treatment requirement was based on a risk assessment. We believe this to be an error of law and we therefore reverse the Panel's finding in paragraph 8.99 that the SPS measure, as it applies to ocean-caught Pacific salmon, is not based on a risk assessment, and that Australia has acted inconsistently with Article 5.1 and, by implication, Article 2.2 of the SPS Agreement.

116. The Panel did not examine whether the import prohibition on fresh, chilled or frozen salmon, the proper SPS measure at issue in this dispute, is based on a risk assessment, and there is no finding on this question in the Panel Report.

117. We are cognizant of the provisions of Article 17 of the DSU that state our jurisdiction and our mandate. Article 17.6 of the DSU provides: "An appeal shall be limited to issues of law covered in the panel report and legal interpretations developed by the panel." Article 17.13 of the DSU states: "The Appellate Body may uphold, modify or reverse the legal findings and conclusions of the panel." In certain appeals, when we reverse a panel's finding on a legal issue, we may examine and decide an issue that was not specifically addressed by the panel, in order to complete the legal analysis and resolve the dispute between the parties. This occurred, for example, in the appeals in United States - Gasoline⁶², Canada - Certain Measures Concerning Periodicals⁶³, European Communities - Measures Affecting the Importation of Certain Poultry Products ("European Communities - Poultry")⁶⁴, and United States - Import Prohibition of Certain Shrimp and Shrimp Products.⁶⁵

118. As we have reversed the Panel's finding that the SPS measure at issue, erroneously identified as the heat-treatment requirement, is not based on a risk assessment, we believe that -- to the extent possible on the basis of the factual findings of the Panel and/or of undisputed facts in the Panel record -- we should complete the legal analysis and determine whether the actual SPS measure at issue, i.e., Australia's import prohibition on fresh, chilled or frozen ocean-caught Pacific salmon, is based on a risk assessment.

119. In examining whether Australia's import prohibition is consistent with Article 5.1, we first have to address whether the 1996 Final Report is, indeed, as Australia contends, a risk assessment within the meaning of Article 5.1 of the SPS Agreement. After a detailed analysis of the requirements of the SPS Agreement relating to risk assessments, and a detailed analysis on whether the 1996 Final Report fulfils these requirements, the Panel assumed -- without making a finding on this issue -- "that the 1996 Final Report meets the requirements of a risk assessment set out in Articles 5.1 and 5.2".⁶⁶ We do not believe it appropriate to base our examination of Article 5.1 on this assumption made by the Panel that the 1996 Final Report is a proper risk assessment. We must, therefore, address this question ourselves. This is possible, because the Panel made all the factual findings necessary to enable us to examine whether the 1996 Final Report meets the requirements of a risk assessment under Article 5.1 of the SPS Agreement.

120. Paragraph 4 of Annex A of the SPS Agreement defines two types of "risk assessment". We agree with the Panel that the type of risk assessment which is required in this case is the type defined in the first part of paragraph 4 of Annex A⁶⁷, which reads as follows:

121. On the basis of this definition, we consider that, in this case, a risk assessment within the meaning of Article 5.1 must:

122. We note that the Panel identified the same three requirements for a risk assessment, within the meaning of Article 5.1 and the first part of paragraph 4 of Annex A of the SPS Agreement.⁶⁸

123. Before taking up the question of whether the 1996 Final Report satisfies these requirements, we note that the first definition in paragraph 4 of Annex A speaks about the evaluation of "likelihood."⁶⁹ In our report in European Communities - Hormones, we referred to the dictionary meaning of "probability" as "degrees of likelihood" and "a thing that is judged likely to be true", for the purpose of distinguishing the terms "potential" and "probability".⁷⁰ For the present purpose, we refer in the same manner to the ordinary meaning of "likelihood", and we consider that it has the same meaning as "probability".⁷¹ On this basis, as well as on the basis of the definition of "risk" and "risk assessment" developed by the Office international des épizooties ("OIE")⁷² and the OIE Guidelines for Risk Assessment⁷³, we maintain that for a risk assessment to fall within the meaning of Article 5.1 and the first definition in paragraph 4 of Annex A, it is not sufficient that a risk assessment conclude that there is a possibility of entry, establishment or spread of diseases and associated biological and economic consequences. A proper risk assessment of this type must evaluate the "likelihood", i.e., the "probability", of entry, establishment or spread of diseases and associated biological and economic consequences as well as the "likelihood", i.e., "probability", of entry, establishment or spread of diseases according to the SPS measures which might be applied.

124. We note that, although the Panel stated that the definition of a risk assessment for this type of measure requires an valuation of the likelihood", for the purpose of satisfying the second and third requirements⁷⁴, it subsequently was hesitant in applying these requirements, by stating or suggesting in paragraphs 8.80, 8.83, 8.89 and 8.91, that some evaluation of the likelihood or probability would suffice. We consider this hesitation unfortunate. We do not agree with the Panel that a risk assessment of this type needs only some evaluation of the likelihood or probability. The definition of this type of risk assessment in paragraph 4 of Annex A refers to "the evaluation of the likelihood" and not to some evaluation of the likelihood. We agree, however, with the Panel's statements in paragraph 8.80 that the SPS Agreement does not require that the evaluation of the likelihood needs to be done quantitatively. The likelihood may be expressed either quantitatively or qualitatively. Furthermore, we recall, as does the Panel⁷⁵, that we stated in European Communities - Hormones that there is no requirement for a risk assessment to establish a certain magnitude or threshold level of degree of risk.⁷⁶

To continue with Article 5.1 of the SPS Agreement

⁵¹WT/DS18/2, 10 March 1997.

⁵²Panel Report, para. 8.19.

⁵³Panel Report, para. 8.155.

⁵⁴Panel Report, para. 8.156.

⁵⁵Panel Report., footnote 430.

⁵⁶Adopted 23 May 1997, WT/DS33/AB/R, p. 17.

⁵⁷Panel Report, para. 8.59.

⁵⁸Panel Report, para. 8.60.

⁵⁹Panel Report, para. 8.96.

⁶⁰It appears that the Panel did not find it necessary to make a definitive finding on whether the 1996 Final Report constituted a "risk assessment", after having come to its conclusion that the measure at issue was not based on a risk assessment.

⁶¹Supra., para. 105.

⁶²Adopted 20 May 1996, WT/DS2/AB/R, pages 19 ff.

⁶³Adopted 30 July 1997, WT/DS31/AB/R, pages 23 ff.

⁶⁴Adopted 23 July 1998, WT/DS69/AB/R, paras. 154 ff.

⁶⁵WT/DS58/AB/R, dated 12 October 1998, paras. 123 ff.

⁶⁶Panel Report, paras. 8.91 and 8.92.

⁶⁷The SPS measure in dispute is a measure to protect animal life or health from risks arising from the entry, establishment or spread of diseases, rather than from risks arising from additives, contaminants, toxins or disease-causing organisms in foodstuffs. Therefore, the type of risk assessment required is the type defined in the first part, rather than the second part, of paragraph 4 of Annex A of the SPS Agreement.

⁶⁸However, the test developed by the Panel seems to suggest that the Member imposing the SPS measure would itself have to conduct the risk assessment (Panel Report, para. 8.72). In that respect the Panel is mistaken. We recall that in paragraph 190 of our report in European Communities - Hormones, we stated that "Article 5.1 does not insist that a Member that adopts a sanitary measure shall have carried out its own risk assessment. It only requires that the SPS measures be 'based on an assessment, as appropriate for the circumstances ...'. The SPS measure might well find its objective justification in a risk assessment carried out by another Member, or an international organization."

⁶⁹We note that the first type of risk assessment in paragraph 4 of Annex A is substantially different from the second type of risk assessment contained in the same paragraph. While the second requires only the evaluation of the potential for adverse effects on human or animal health, the first type of risk assessment demands an evaluation of the likelihood of entry, establishment or spread of a disease, and of the associated potential biological and economic consequences. In view of the very different language used in paragraph 4 of Annex A for the two types of risk assessment, we do not believe that it is correct to diminish the substantial differences between these two types of risk assessments, as the European Communities seems to suggest when it argues that "the object, purpose and context of the SPS Agreement indicate that no greater level of probability can have been intended for the first type of risk assessment than for the second type, [as b]oth types can apply both to human life or health and to animal or plant life or health". (Third participant's submission of the European Communities, para. 7).

⁷⁰Adopted 13 February 1998, WT/DS26/AB/R, WT/DS48/AB/R, para. 184.

⁷¹"Likelihood: probability", The Concise Oxford Dictionary of Current English (9th ed., Clarendon Press).

⁷²OIE Code, Section 1.1, Definitions, p. 13. We note that Article 5.1 provides that risk assessments should take "into account risk assessment techniques developed by the relevant international organizations". In the context of this case, the relevant international organization is the OIE.

⁷³OIE Code, Chapter 1.4.2, p. 33 and 34.

⁷⁴Panel Report, para. 8.72.

⁷⁵Panel Report, para. 8.80.

⁷⁶Adopted 13 February 1998, WT/DS26/AB/R, WT/DS48/AB/R, para. 186.