Australia - Measures Affecting the Importation of Apples from New Zealand (AB-2010-2 / DS367)

Opening Statement of Australia

October 11, 2010

As delivered


Thank you Madam Zhang,

  1. In opening, I propose to say nothing of New Zealand's other appeal and to say nothing of ground (a) of Australia's appeal. I want to concentrate on the real and substantial issues raised by grounds (b), (c) and (d) of Australia's appeal, particularly in relation to fire blight, a very serious disease of apples, of which Australia is free but which has long been endemic in New Zealand.
  2. To aid that concentration, let me clear away some of the distractions that emerge from New Zealand's appellee submissions.
  3. The first distraction lies in the allegation that Australia's IRA process was, in some unspecified but apparently prejudicial way, tainted by what New Zealand describes as a "highly politicised context"1 and because one of the six members of the risk assessment team was an owner and manager of an apple production business.2 The allegation is the subject of no relevant findings of fact by the Panel. It relies on strongly disputed inferences sought selectively to be drawn from a manifestly incomplete evidentiary record. It was rightly ignored by the Panel. It ought be ignored by the Appellate Body.
  4. The second distraction lies in the repeated assertion that Australia sought in its IRA process to avoid its obligations under the SPS Agreement and that Australia sought before the Panel and seeks again in this appeal to "shield" the IRA from review.3 That assertion, too, was rightly ignored by the Panel, and ought be ignored by the Appellate Body. The Panel recorded uncritically that the revised methodology of the IRA was adopted by Australia in 2004 for the express purpose of reinforcing the transparency and objectivity of the analysis.4 Whatever other criticisms might legitimately be levelled against it, the IRA, a very lengthy document, leaves no doubt about: the potential biological pathways it identifies, the scientific evidence it takes into account, the methodology it applies and the conclusions it reaches at every step in the application of that methodology. Those conclusions were generated in part through the application of a computer probability distribution model given to New Zealand in 2006.5 New Zealand could, if it so chose, as part of its case before the Panel have re-run that computer model for itself. New Zealand could, if it so chose, have sought to demonstrate using Australia's own model the precise significance of the errors and exaggerations it alleged. What Australia sought before the Panel and seeks again in this appeal is nothing more than to ensure that Australia's own obligations under the SPS Agreement are properly interpreted and applied, that the functions of a Panel under the DSU are properly performed and that New Zealand is properly held to its burden of proof.
  5. The third distraction lies in the suggestion that the IRA process involved nothing more than a re-run by Australia, on the same old evidence, of the same old arguments, considered and rejected by two Panels in the Japan — Apples dispute.6 The suggestion is wrong: the evidence was not the same and the arguments were not the same.7
  6. The fourth and quite significant distraction lies in the global assertion that the scientific evidence in support of the IRA's conclusions either did not exist or went in the opposite direction from the conclusions actually reached in the IRA.8 The assertion is wrong. The two fire blight experts were Dr Deckers and Dr Paulin. Thosfe experts agreed there was a scientific basis for the proposition that there was some likelihood of mature, symptomless apples imported into Australia being epiphytically (externally) infested with the fire blight bacteria.9 The question was entirely one as to the degree of that likelihood. The IRA estimated a most likely value of 3.9%.10 Dr Paulin did not think this figure was credible.11 Dr Deckers said that it might have been overestimated but explained later that he saw no exaggeration at all.12 The experts agreed and the Panel accepted that there was also scientific support for the hypothesis that the fire blight bacteria can be transferred by browsing insects from epiphytically-infested apples to host plants in Australia.13 They accepted that the likelihood of that occurring was incapable of experimentation but that it was not simply theoretical; they both saw the likelihood as very low.14The question again was one of degree: just how low? The IRA estimated a probability range of zero to one in a million with a uniform distribution. Dr Paulin evidently thought that was too high.15 Dr Deckers said quite directly and explicitly that he thought it was right.16
  7. The final distraction lies in attempts by New Zealand, throughout its appellee submissions, to supplement the Panel's findings of fact by evidence to which the Panel makes no reference in its analysis17 and to bolster the reasons actually given by the Panel to justify its findings and recommendations by reference either to arguments put by New Zealand18 or to inferences sought to be drawn from questions asked by members of the Panel that nowhere get taken up by the Panel in its report.19 The Panel had an obligation under Article 12.7 of the DSU to set out the totality of the findings of fact it made and to set out the underlying reasoning or justification20 behind each of its findings and recommendations. The Panel has produced a very detailed 548-page report. The Panel can be taken in that report to have said all that it meant and meant all that it said.

Ground (b)

  1. Turning to ground (b), I limit myself to issues of principle and seek only to dispel three misconceptions about Australia's submission that run through the submissions of New Zealand.
  2. Australia does not propose that a Panel reviewing a risk assessment for compliance with Article 5.1 of the SPS Agreement must engage in de novo review. A Panel does not engage in de novo review — does not itself stand in the shoes of or perform the role of a risk assessor — merely by asking whether a particular conclusion reached by a risk assessor has or has not been shown to fall outside a legitimate range or merely by asking whether any particular error on the part of a risk assessor has or has not been shown to be material to the overall assessment of risk.
  3. Nor does Australia propose some "new test" that departs from the third criterion articulated by the Appellate Body in Continued Suspension. It is useful to recall precisely how that criterion was formulated. The Appellate Body said this:21

    A panel should also assess whether the reasoning articulated on the basis of the scientific evidence is objective and coherent. In other words, a panel should review whether the particular conclusions drawn by the Member assessing the risk find sufficient support in the scientific evidence relied upon.

  4. There are three observations to make about that criterion. The first is that it takes as given the totality of the scientific evidence that is available within the meaning of Article 5.2 and that comes from a respected and qualified source as required by the second criterion in Continued Suspension. The evidence will be what it will be: perhaps general or specific, perhaps sparse or dense, perhaps incomplete or conflicting, perhaps capable of direct application or perhaps capable of only indirect and analogical application. The second observation is that there is nothing to suggest that the reasoning articulated on the basis of the scientific evidence is to be anything other than objective and coherent in the sense explained in EC — Hormones and endorsed at an earlier stage of the analysis in Continued Suspension: that is to say, it is to be the product of "systematic, disciplined and objective inquiry and analysis".22 The third and most important observation flows from the link between the two sentences by which the criterion is formulated. The two sentences are expressed as different ways of saying the same thing. The key to the first sentence lies in the restatement that occurs in the second sentence. That restatement is directly linked to the language of the basic obligation set out in Article 2.2 of which Article 5.1 is a specific application.23 The second sentence makes clear that the criterion is not directed to assessing the quality of the reasoning as an end in itself. It is directed to determining whether the particular conclusions drawn by the Member find sufficient support in the scientific evidence. Whatever level of specificity might be involved in the notion of a particular conclusion, a risk assessment cannot fail at the third criterion merely because of a perceived shortcoming in the quality of the reasoning to a particular conclusion if the particular conclusion itself finds sufficient support in the scientific evidence relied upon. It is not the task of a Panel to assess whether more elaborate reasoning might have been desirable: the task of the Panel is to assess whether the particular conclusion finds sufficient support in the evidence. Sufficiency of support lies in the existence of a rational or objective relationship between the evidence and the conclusion: the question being whether the conclusion lies inside or outside a range considered legitimate according to the standards of the scientific community.
  5. Australia does not seek in any way to avoid review of particular conclusions drawn at intermediate steps in the performance of a risk assessment.24 The point is that the nature of the inquiry into the conclusions drawn at those intermediate steps is informed by the place of those intermediate steps within the overall assessment of risk. If a particular intermediate conclusion is found to be outside a range considered legitimate according to the standards of the scientific community, the inquiry must go further to assess whether the error is material: whether the error, alone or in combination with other errors, is "so serious" as to undermine "reasonable confidence" in the risk that has ultimately been assessed.25 New Zealand here appears to have missed the point. Throughout its submission, New Zealand attempts to demonstrate the materiality of errors found by the Panel by reference to its own submissions before the Panel as to the significance of particular steps in the risk assessment methodology.26 Not only does this illegitimately seek to supplement the Panel's own reasons, but, more fundamentally, it is not the significance of the step in the methodology that is, for present purposes, important: what is important is the significance of any error in the judgment or conclusion reached at that step. Even on the most significant of steps, a minor error in the conclusion may be immaterial, in the sense that it is not so serious as to undermine reasonable confidence in the risk that has ultimately been assessed.27

Ground (c)

  1. Ground (c) complements ground (b) in this sense. If a Panel reviewing a risk assessment for compliance with Article 5.1 of the SPS Agreement fails to consider or engage with relevant evidence, the failure of the Panel can indicate one or both of two things. More probably, it indicates a failure of the Panel properly to understand and apply the legal criterion that makes the evidence relevant. But additionally, or alternatively, it might indicate a failure of the Panel properly to discharge the legal obligation imposed by Article 11 of the DSU to make an objective assessment of the facts. The second is not lightly to be inferred.28 But nor is it to be treated lightly where it occurs. In Continued Suspension, the Appellate Body found to be irreconcilable with the duty to conduct an objective assessment the Panel's mere reproduction of testimony without assessment of its significance.29
  2. In the present case, Australia complains of even more serious failures by the Panel: in some cases, complete oversight of, and, in other cases, bare references in footnotes, in peremptory and dismissive terms, to significant evidence favourable to Australia's case.30
  3. The failure is most serious when regard is had to the way in which the Panel dealt — or, more correctly, failed to deal — with the evidence of Dr Deckers, one of the two fire blight experts. When the written and oral testimony of Dr Deckers is fairly read as a whole, the conclusion — if not inescapable — is at least very strongly open that, notwithstanding some criticism of matters of detail: Dr Deckers agreed with the conclusion reached in the IRA as to the quantitative likelihood of importation of the fire blight bacteria;31 he agreed with the conclusion reached in the IRA as to the quantitative likelihood of exposure;32 he agreed with the conclusion reached in the IRA as to the qualitative label to be attached to the potential consequences;33 and he considered that restricting imports to mature symptomless apples would be incapable of meeting Australia's appropriate level of protection.34 None of the testimony in this respect has been properly or fairly addressed and the thrust of what he truly had to say has been substantially ignored by the Panel.35
  4. The failure is most stark when it comes to Dr Deckers' agreement with the conclusion reached in the IRA as to the quantitative likelihood of transfer of the fire blight bacteria from mature apple fruit imported from New Zealand to a host plant in Australia. In written answers to the Panel's written questions Dr Deckers repeatedly gave evidence to the effect that the likelihood of this epiphytic population being transferred to a host plant was "very small", or "very low", or "rather small", or "rather exceptional", or "rather low"36, terminology which he used deliberately (in contrast to an "extremely low" likelihood if measures were applied37).
  5. When it came to the Panel's meeting with the experts, Dr Deckers was asked a very direct question:38

    Dr Deckers: In your response to Question 35 you state that "The chance that the epiphytic bacteria will be transmitted to the susceptible organs of a host plant on the appropriate moment to realise an infection is rather small". How does this compare to the conclusion in the IRA which states that the exposure value for an individual apple should be in the range from 0 to 1 in a million?

    He gave a very direct answer:39

    This value between zero and ten to the sixth is also very low, so I think this is true.

  1. Together with Dr Paulin, Dr Deckers was ultimately asked another fairly direct question on the same general topic:40

    Do you think that the IRA reaches a rational logical inference that the probability of transmission via trade in apple fruit is not zero? [That question is important because Australia defines its appropriate level of protection as "very low but not zero".] In other words, on the basis of what evidence we have, is it rational to conclude that transmission via apple fruit could occur, albeit as a rare event?

    He gave another direct and consistent answer:41

    Here I would say yes, there is indeed a risk for the importation of infected fruits. It is clear.

  1. In the report of the Panel, both of those answers of Dr Deckers are ignored entirely.42 The omission is stark. And the omission becomes even more stark when it is recognised that the Panel in its report places a great deal of weight on the quite different answer to the second of those questions that was given by Dr Paulin.43

Ground (d)

  1. The point in ground (d) of the appeal can be stated quite simply: to find a breach of Article 5.6, the Panel was required to find as a fact that the alternative measures New Zealand proposed would achieve Australia's appropriate level of protection.
  2. The Panel failed to find that fact for either fire blight or apple leafcurling midge. In stark contrast to its findings in respect of Article 5.1, which it expressed simply and properly as "findings",44 the Panel in dealing with Article 5.6 eschewed the language of "findings", avoided even stating as a matter of substance what it considered the facts to be and retreated instead to expressing itself in terms of New Zealand having raised a "sufficiently convincing presumption, not successfully rebutted by Australia".45 In relation to apple leafcurling midge, the Panel went so far as to explain that it was making a "legal, not a scientific finding" in the sense that "subject to an objectively justifiable analysis [Australia] may conclude that the ALCM risk exceeds Australia's ALOP".46 The Panel was not required to make a legal finding or a scientific finding: it was required to make a factual finding, which it failed to do.
  3. In so far as the Panel made any finding at all it manifestly failed properly to address the question of whether the alternative measures proposed by New Zealand would achieve Australia's ALOP. Given that consequences are an integral part of Australia's ALOP, the question cannot be addressed at all without analysis of consequences: there was none. And the Panel expressed such findings as it made in terms that the alternative measures "could" or "might" achieve Australia's appropriate level of protection: that is a significantly lower standard than the correct legal standard, expressed in the text of Article 5.6, that the alternative measure "would" achieve Australia's appropriate level of protection.


  1. Those are the main issues before the Appellate Body in this appeal. Australia looks forward to answering the Appellate Body's questions.