AUSTRALIA - MEASURES AFFECTING IMPORTATION OF SALMON
- RECOURSE TO ARTICLE 21.5 BY CANADA -

C. "THE EXISTENCE � OF MEASURES TAKEN TO COMPLY WITH THE RECOMMENDATIONS AND RULINGS" OF THE DSB IN THE SENSE OF ARTICLE 21.5 OF THE DSU

7.23 Canada claims that Australia has not implemented all of the measures required for it to comply with the recommendations and rulings of the DSB. On that ground, Canada submits that no measures to implement the recommendations and rulings of the DSB "exist" in the sense of Article 21.5 of the DSU.

7.24 At the DSB meeting of 27-28 July 1999, Australia announced that its "Quarantine and Inspections Service Decision of 19 July had brought Australia into full conformity with its WTO obligations".¹⁵¹ The decision referred to is AQPM 1999/51.¹⁵² This decision sets out new policies in respect of salmonids, non-salmonids and live ornamental finfish.

7.25 Subsequent to the 19 July 1999 decision, seven additional AQPM's were published. These additional AQPM's set out in more detail the new policies announced in AQPM 1999/51.

7.26 Canada contests that this series of new measures are consistent with the SPS Agreement. Australia, in contrast, is of the view that they ensure full implementation of the DSB recommendations and rulings and are fully consistent with the SPS Agreement.

7.27 We do not, at this stage, need to examine questions of consistency. Of importance here, is whether the new measures - according to Australia fully implementing DSB recommendations and rulings -- "exist".

7.28 In our view, a new regime of implementing measures can be said to "exist" when this regime sets out all requirements and criteria under which the product concerned can enter the market of the implementing Member. For products to be able to enter the market, the new measures setting out these requirements and criteria also have to be in force. We do not consider a framework regulation setting out the basic - but not all - requirements and criteria to be sufficient for a new regime to "exist". On the other hand, we do not consider it necessary that product has actually entered the market. In our view, of decisive importance is whether under the new regime trade opportunities effectively exist; not whether they will occur in the future, nor whether they have actually given rise to specific transactions in the past.

7.29 Examining the measures at issue here, we find that taken together they define all requirements and criteria for relevant product to be able to enter the Australian market under the new quarantine regime.¹⁵³

7.30 However, the date of entry into force of the new measures varies according to the products covered. In all cases, the entry into force - and thus the "existence" of the measures taken to comply - occurred subsequent to 6 July 1999, the date of expiry of the reasonable period of time given to Australia to implement the DSB recommendations and rulings. Since, in this case, Australia was under an obligation to implement the DSB recommendations and rulings by the end of the reasonable period of time¹⁵⁴, we find that for the period of time that the new measures did not and will not apply subsequent to 6 July 1999, no measures taken to comply existed or will exist in the sense of Article 21.5.

7.31 For salmonids, including fresh chilled or frozen salmon from Canada at issue here, the basic framework was laid down on 19 July 1999 (AQPM 1999/51). The necessary supplements to this basic framework -- "conditions which clarify arrangements for the importation of uncanned salmonid product in accordance with the policies announced in AQPM 1999/51" -- were published on 20 October 1999.¹⁵⁵ Thus, only on 20 October 1999 -- almost three and a half months late - did measures taken to comply in respect of Canadian fresh chilled or frozen salmon "exist".

7.32 For non-salmonids -- including herring for use as bait which was one of the situations compared to Canadian salmon under Article 5.5 in the original dispute -- the new measures entered into force on 1 December 1999.¹⁵⁶ Thus, the inconsistency with Article 5.5 found in the original dispute in respect of Canadian salmon vis-�-vis herring for use as bait was maintained, at least in part, until 1 December 1999, i.e. until almost five months after the end of the reasonable period of time.

7.33 Finally, for live ornamental finfish -- the other situation compared to Canadian salmon under Article 5.5 in the original dispute -- implementation of new requirements will be staged over a period of time.¹⁵⁷ On 1 December 1999, the requirements relating to import permits entered into force. From 1 February 2000, all new requirements relating to exporters and exporting countries will apply. From 1 May 2000, all importers must fully comply with new post-arrival quarantine requirements. Thus, although the inconsistency with Article 5.5 found in the original dispute in respect of Canadian salmon vis-�-vis live ornamental finfish may be gradually alleviated, all requirements and criteria for product to enter Australia under the new regime - a regime that according to Australia will achieve compliance - will only apply from 1 May 2000. This particular inconsistency with Article 5.5 will thus - as Australian quarantine policy now stands - be maintained, at least in part, for almost 10 months subsequent to the expiry of the reasonable period of time.

7.34 In its oral statement, Canada also refers to 1 January 2002 as the date of entry into force of certain disease testing requirements that apply to imports of goldfish. We note, however, that the requirement referred to relates to statements of freedom from specified disease agents based on a testing programme that demonstrates absence of the disease agents in the source population over a period of at least two years. The very nature of this requirement makes it difficult to impose the requirement immediately. If this were done, imports from countries where so far no testing programmes were carried out would be banned for two years. On that ground, a staged implementation of disease testing requirements, as the one imposed by Australia, is, in our view, justifiable and does, we believe, not prevent the new regime on live ornamental finfish from "existing" in the sense referred to earlier.¹⁵⁸

7.35 Consequently, for the periods of time specified above, no measures taken to comply "existed" in the sense of Article 21.5. As a result, during those periods of time, Australia failed to bring its measure into compliance with the SPS Agreement as called for in the DSB recommendation, in the sense referred to in Article 22.6 of the DSU.

7.36 Under its claim that Australia has not taken measures to comply, Canada also submits that "not all of the requirements Australia imposes are necessarily listed in the AQPMs".¹⁵⁹ Canada refers, in particular, to a requirement imposed on Canada in order to obtain a health certificate that "fish do not come from waters within 10 kilometers or one tidal interchange of an [ISA] infected farm, whichever is greater". Canada argues that this requirement is specified neither in the AQPMs nor mentioned in the 1999 Report. However, since Canada does not make any claim of inconsistency with the SPS Agreement, nor provided any documentary evidence, in respect of this requirement, we are not called upon, nor in a position, to make any findings on this requirement.

D. SANITARY MEASURES BASED ON A RISK ASSESSMENT PURSUANT TO ARTICLE 5.1 OF THE SPS AGREEMENT

7.37 The previous section of our Report addresses the existence of certain measures. We now turn to the second part of Canada's claims, those relating to the consistency of the new measures with certain provisions of the SPS Agreement. We note, generally, that Australia has, indeed, lifted the import prohibition on Canadian fresh chilled or frozen salmon and taken steps to facilitate access of Canadian product, albeit subject to certain conditions. We recall also that the burden of proof rests on Canada to demonstrate that these conditions are inconsistent with the provisions of the SPS Agreement.

7.38 We start our examination with Canada's claims under Article 5.1.

Article 5.1 reads as follows:

"Members shall ensure that their sanitary � measures are based on an assessment, as appropriate to the circumstances, of the risks to human, animal or plant life or health, taking into account risk assessment techniques developed by the relevant international organizations".

7.39 An examination of whether sanitary measures are based on a risk assessment in accordance with Article 5.1 involves two steps:

(1) does the study put forward as a risk assessment meet the requirements of a risk assessment set forth in Article 5.1 and Annex A of the SPS Agreement?;

(2) if so, are the sanitary measures finally selected based on this risk assessment as required in Article 5.1?

1. The three requirements of a risk assessment in accordance with the SPS Agreement

7.40 Risk assessment for purposes of the SPS Agreement and, in particular, Article 5.1 thereof, is defined in paragraph 4 of Annex A as

"The evaluation of the likelihood of entry, establishment or spread of a � disease within the territory of an importing Member according to the sanitary � measures which might be applied, and of the associated potential biological and economic consequences".

7.41 This definition contains a three-pronged test. Consequently, the 1999 IRA¹⁶⁰, the only study Australia puts forward as a risk assessment in support of its measures, needs to

"(1) identify the diseases whose entry, establishment or spread a Member wants to prevent within its territory, as well as the potential biological and economic consequences associated with the entry, establishment or spread of these diseases;

(2) evaluate the likelihood of entry, establishment or spread of these diseases, as well as the associated potential biological and economic consequences; and

(3) evaluate the likelihood of entry, establishment or spread of these diseases according to the SPS measures which might be applied".¹⁶¹

7.42 Canada does not contest that the 1999 IRA meets the first requirement. At issue here is whether the 1999 IRA evaluates the likelihood of entry, establishment or spread of the diseases identified by Australia (second requirement) and whether it does so according to the sanitary measures which might be applied (third requirement).

7.43 The definition of risk assessment in Annex A is to be read and applied in the context of the general obligation in Article 5.1 to base sanitary measures on a risk assessment as well as in light of the specific factors a risk assessment has to take into account pursuant to Article 5.1¹⁶², Article 5.2¹⁶³ and Article 5.3.¹⁶⁴ Finally, also the basic obligations in Article 2.2 impart meaning to the definition of risk assessment.¹⁶⁵

(a) The second requirement: "The evaluation of the likelihood of entry, establishment or spread of � disease"

7.44 The context we referred to in the previous paragraph is of particular importance when examining Canada's claim that the 1999 IRA does not adequately evaluate likelihood and evaluates likelihood in a highly subjective way.

7.45 In the original dispute Canada claimed, and the Appellate Body agreed¹⁶⁶, that the 1996 Final Report¹⁶⁷ evaluated possibility -- instead of likelihood or probability -- of disease entry, establishment or spread. On that basis, the Appellate Body found that the 1996 Final Report did not meet the second requirement of a risk assessment. "Some" evaluation of the likelihood - on the basis of which the original panel had continued its examination without making a finding on the issue¹⁶⁸ - was found to be insufficient. What is required, according to the Appellate Body, is "the evaluation of the likelihood", without there being a need for this evaluation to be done quantitatively.¹⁶⁹

7.46 In this case Canada agrees that the 1999 IRA constitutes progress vis-�-vis the 1996 Final Report and addresses likelihood or probability. However, Canada is of the view that the 1999 IRA does not appropriately, adequately or objectively evaluate such likelihood in accordance with the second requirement of risk assessment. Although Canada submits in its answer to a Panel Question that the 1999 IRA "cannot be said to have taken [the factors referred to in Articles 5.1 to 5.3] into account in an appropriate manner"¹⁷⁰, Canada does not make any specific claims of inconsistency in this respect.

7.47 Canada's claim raises the question of where to put the threshold of an evaluation of likelihood consistent with the SPS Agreement. On the one hand, we find it difficult to read into the summary definition of risk assessment set out in paragraph 4 of Annex A - which only refers to "the evaluation of the likelihood"¹⁷¹ - specific requirements such that minor flaws or misconceptions at a detailed level would preclude a study from falling within the SPS definition of risk assessment.¹⁷² As agreed by all parties and experts involved in this dispute, risk assessment, in particular a qualitative risk assessment like the 1999 IRA, inevitably includes subjective elements.¹⁷³ On the other hand, we realize that there may be studies that are flawed or biased to such extent that they cannot be said to meet any standard of objectivity. We do not think that such studies should pass the test of a risk assessment in accordance with the SPS Agreement.

7.48 In the absence of an explicit, textual threshold in paragraph 4 of Annex A itself, we turn to the context outlined in paragraph 7.43 above. The reference made there to a series of objective factors such as "risk assessment techniques developed by the relevant international organizations"¹⁷⁴, "available scientific evidence"¹⁷⁵, "scientific principles" and "sufficient scientific evidence"¹⁷⁶, strengthens our view that the evaluation of likelihood needs to achieve a certain level of objectivity.

7.49 We find further support in Article 5.7 of the SPS Agreement. This provision allows Members to take provisional sanitary measures when relevant scientific evidence is insufficient pending a search for the additional information "necessary for a more objective assessment of risk" (emphasis added). This implies that except for provisional measures - not at issue here - a risk assessment has to meet a certain level of objectivity.¹⁷⁷

7.50 We also considered Section 1.4 of the 1997 OIE International Aquatic Animal Health Code on "Import Risk Analysis" which includes techniques that a risk assessment in the area of aquatic animal health must take into account pursuant to Article 5.1. This OIE Code states that "[t]he principal aim of import risk analysis is to provide importing countries with an objective and defensible method of assessing the disease risks associated with the importation of aquatic animals � Import risk analysis is preferable to a zero-risk approach because it provides a more objective decision" (emphasis added).¹⁷⁸ This, as well, supports the view that the evaluation of likelihood needs to achieve a certain level of objectivity.

7.51 With this in mind, we hold the view that the level of objectivity to be achieved in a risk assessment must be such that one can have reasonable confidence in the evaluation made, in particular in the levels of risk assigned.

7.52 Applying this standard to the 1999 IRA, and after careful consideration of all the arguments and evidence submitted to us by the parties and the experts advising the Panel, we are of the view that the 1999 IRA meets the required level of objectivity.

7.53 The 1999 IRA first identifies the diseases of concern to Australia in respect of salmonids, applying certain criteria, e.g. whether the disease is infectious, exotic, OIE listed, etc. It sub-divides diseases of concern into "higher priority" and "lower priority" diseases.

7.54 For each of the 15 "higher priority" diseases of concern, it then determines the probability of the disease entering and becoming established in Australia through imports of eviscerated salmonids, making separate release and exposure assessments. For each of these diseases it also determines the expected impact or significance of disease establishment (consequence assessment). To each of these two elements, a qualitative scale is then attributed specifying, respectively, the probability of the disease becoming established (ranging from high, moderate, low, very low, extremely low to negligible) and the severity of the impact (ranging from catastrophic, high, moderate, low to negligible). This is done, again, disease-by-disease.

7.55 Following a standard risk evaluation matrix that applies in respect of all diseases, it is then determined whether, for each disease, the risk of disease establishment and its impact, related to imports of eviscerated salmonids, is acceptable in light of Australia's appropriate level of protection (ALOP). For example, a high probability of establishment with negligible consequences is tolerated, whereas a moderate risk of establishment with low consequences is not. On that basis seven of the 15 "high priority" diseases are found to represent a risk that is not acceptable.¹⁷⁹

7.56 We note that two of the three experts we appointed - Drs. Br�ckner and McVicar - are of the view that the 1999 IRA does appropriately evaluate the likelihood of entry, establishment and spread of diseases into Australia. The third expert - Dr. Wooldridge - concludes the opposite on the basis of certain flaws she detected in the 1999 IRA.¹⁸⁰ These flaws - also referred to by Canada -- centre around the exposure assessment not taking fully into account all information available, the possibility of bias in the release assessment due to the way certain information was presented and the complexity of the qualitative terms used.¹⁸¹

7.57 Neither Canada nor the experts advising us, refers to scientific or other information that was not taken into account in the 1999 IRA. Moreover, whereas Canada and Dr. Wooldridge do point out certain methodological flaws and alleged inconsistencies in the 1999 IRA that - if absent - might have led to a lower level of assessed risk, we have not been convinced that this would be so, at least not to such extent that we would no longer have reasonable confidence in the levels of risk currently assigned in the 1999 IRA. In summary, we believe that the flaws identified are not so serious as to prevent us from having reasonable confidence in the evaluation made and the levels of risk assigned.¹⁸²

7.58 Recalling that Canada bears the burden of demonstrating that the 1999 IRA does not fulfil the requirements of risk assessment, we thus conclude that the 1999 IRA evaluates the likelihood of disease entry, establishment or spread in accordance with the second requirement of a risk assessment.

(b) The third requirement: "The evaluation of the likelihood � according to the sanitary � measures that might be applied"

7.59 The one specific criterion that is textually referred to in paragraph 4 of Annex A for an "evaluation of the likelihood" to be consistent with the SPS Agreement is contained in the third requirement of the definition of risk assessment: the evaluation of the likelihood of entry, establishment or spread has to be made "according to the sanitary � measures that might be applied".

7.60 In the original dispute, the Appellate Body found that the 1996 Final Report does not meet this third requirement, considering, again, that "some evaluation of the likelihood is not enough".¹⁸³ It did so on the basis of factual findings made by the original panel :

"132. � We note that the Panel observed that the 1996 Final Report examines a large number of different risk reduction factors for each of the 24 diseases of concern, and we note that the Panel came to the following factual finding¹⁸⁴:

For most of these risk reduction factors, the 1996 Final Report provides some evaluation of the extent to which these factors could reduce risk. (emphasis added)

133. With regard to the quarantine policy options considered to reduce the total risk associated with all diseases of concern, the Panel, arrived at these factual findings:

� that the 1996 Final Report does not substantively evaluate the relative risks associated with these different options. Even though the definition of risk assessment requires an 'evaluation ... according to the sanitary ... measures which might be applied', the 1996 Final Report identifies such measures but does not, in any substantial way, evaluate or assess their relative effectiveness in reducing the overall disease risk. (emphasis added)".¹⁸⁵

7.61 In this case we have to examine whether the 1999 IRA meets the test which the 1996 Final Report failed. For the reasons explained below, we are of the view that it does.

7.62 Having identified the seven "higher priority" diseases in respect of which imports of eviscerated salmonids would not meet Australia's ALOP, the 1999 IRA next considers whether so-called risk management or risk reduction measures could be implemented to reduce the risk to a level that would meet Australia's ALOP.

7.63 The 1999 IRA first identifies available quarantine measures, both pre-export requirements applying to the country of origin and post-import measures applying in Australia. Thereafter, for each of the seven diseases found to represent a risk that is not acceptable, "Key Risk Factors" are pointed out, such as the kind of control measures in case of disease establishment, the type of salmon with the highest prevalence, most infected tissues, survival rate and risk related to waste.

7.64 Subsequently, for each disease - not only for some diseases as in the 1996 Final Report -- a series of "Risk Management Measures" that might be applied and that would reduce the risk associated with the particular disease are identified and discussed, such as control of health status through health surveillance and monitoring, restrictions as to the age of the fish, inspection and grading, processing, export certification and controls on waste disposal.

7.65 Critically -- and contrary to what was done in the 1996 Final Report -- the discussion provided in the 1999 IRA for each of the "Risk Management Measures" is made in light of the effect these measures would have on the "Key Risk Factors" previously identified. On the basis of these discussions - which we consider to be evaluations -- certain conclusions are made and a list of pre-export and/or post-import requirements is adopted for each specific disease in order to achieve Australia's ALOP.¹⁸⁶

7.66 On the basis of this disease-by-disease assessment, the 1999 IRA concludes that importation of eviscerated salmonids from any country should be permitted subject to a series of measures - a combination of all "Risk Management Measures" identified for the seven diseases - that would have the effect of reducing the overall risk related to imports of salmonids to a level that is acceptable to Australia.¹⁸⁷

7.67 We note that two of the three experts advising the Panel - Drs. Br�ckner and McVicar -- are of the view that the 1999 IRA evaluates the likelihood of disease entry, establishment or spread according to the sanitary measures that might be applied.¹⁸⁸ Dr. Wooldridge, in contrast, is "unable to find any indication that the probability of any individual (or indeed any combination) of measures has actually been assessed specifically with regard to the likelihood of bringing the assessed risk below Australia's ALOP" (emphasis added).¹⁸⁹ She does agree, however, that "[f]or each disease which does not meet the ALOP criteria, risk factors have been identified, and a list of possible risk management measures described. In addition, the particular risk factor which each measure would address is indicated".¹⁹⁰

7.68 Canada's claim, as well as Dr. Wooldridge's opinion, raises the question of whether the definition of risk assessment as such, requiring Members to assess risk "according to the [sanitary] measures which might be applied", can be construed so as to include the obligation to make the link between the assessment, the measures finally selected and the necessity to use these measures in order to achieve the ALOP. We find it difficult to read such a requirement into paragraph 4 of Annex A.

7.69 In our view, the rights and obligations in respect of these linkages are set out not in the definition of risk assessment itself - which logically precedes the selection of measures -- but, inter alia, in the obligation to base sanitary measures on a risk assessment in Article 5.1 and to ensure that sanitary measures are not more trade-restrictive than required to achieve the ALOP in the sense of Article 5.6. To examine these questions of relationship between the risk assessment, the measures selected and the ALOP under the definition of risk assessment - as Canada and Dr. Wooldridge seem to do -- would, in our view, run the risk of adding to or diminishing the more specific rights and obligations of Members set out in other SPS obligations, contrary to Article 19.2 of the DSU.

7.70 We examine these questions of relationship between the measures at issue and the risk assessment below.¹⁹¹ Even if we were to find there that some of the Australian measures in question are not based on the 1999 IRA on the ground, for example, that the 1999 IRA does not explain or assess these measures, we would not be precluded from finding here that the 1999 IRA meets the definition of risk assessment. Indeed, the fact that the 1999 IRA would not evaluate the likelihood according to all sanitary measures which may be applied, including some of those that were actually selected, does not, in our view, preclude that the 1999 IRA taken separately meets the definition of risk assessment. Paragraph 4 of Annex A refers to an evaluation "according to the sanitary � measures which might be applied" tout court. It does not require that all possible measures (of which there could be a very great number) be evaluated nor specify precisely which measures need to be evaluated. In any event, we prefer to address this question of relationship between the measures selected and the risk assessment under the obligation to base measures on a risk assessment pursuant to Article 5.1 rather than under the very definition of risk assessment referred to in the same provision.

7.71 For all the reasons above, after careful examination of all the arguments and evidence submitted to us by the parties and the experts advising the Panel and recalling that Canada bears the burden of demonstrating that the 1999 IRA does not fulfil the requirements of a risk assessment, we conclude that the 1999 IRA evaluates the likelihood of disease entry, establishment or spread "according to the sanitary � measures which might be applied" consistently with the third requirement of a risk assessment.

191 See paras. 7.72 ff. and 7.115 ff.

To continue with 2. Sanitary measures based on a risk assessment

Return to Table of Contents