So, let’s talk strategy: Should the EC appeal the panel ruling on GMOs?

For all we know, the GMO panel faulted the EC on two grounds.  First, EC approval procedures have been subject to “undue delay” (Annex C.1(a)).  Second, safeguard restrictions imposed by certain EC member states against GMOs that were approved at the EC level, are not based on a risk assessment (5.1), nor provisionally justified (5.7) and, consequently, not based on scientific principles and maintained without sufficient scientific evidence (2.2.).

Looked at purely from the power perspective of the EC Commission, one might be inclined to declare victory.  First, the panel agreed with the EC that the moratorium ended in 2003 and, as a result, declined to make a recommendation in this respect.  Second, as to the safeguards by Austria, France and others, all the panel did was call upon EC member states to step in line with … EC policy.  What else could the EC Commission have wished for?  Is Mandelson’s statement that this case is “largely of historical interest”, and the reaction by Greenpeace that the panel ruling is “irrelevant”, a prelude to the EC not appealing this case?  After all, the risk of an Appellate Body confirming the panel (and thereby elevating its weight to de facto precedent) might well outweigh the on-the-ground benefits of an (unlikely) reversal on appeal.  Was not a similar trade-off made in high profile cases such as Kodak-Fuji and Section 301?

That said, the pressure from EC member states is most likely to push the EC Commission to appeal anyhow.  And, indeed, when it comes to the safeguards there is something distinctly fishy about the panel’s ruling. 

What the panel is essentially saying is that since the EC was able to make a risk assessment, EC member states can no longer claim that there is insufficient scientific evidence.  Hence, they cannot impose provisional measures under 5.7.  As the EC and its member states do, after all, remain distinct sovereigns and distinct WTO members, this amounts to saying that because one WTO member (say, the US) was able to make a risk assessment, any other WTO member can no longer invoke 5.7.  Imagine what this would mean for the EC or, even more so, for developing countries? 

The panel’s ruling in this respect raises a crucial question on the relation between 2.2, 5.1 and 5.7.  Is 5.7 really tied to 5.1 (in the sense that as soon as a risk assessment is feasible, provisional measures can no longer be enacted), or is the better view that 5.7 is tied to 2.2 (even if a risk assessment is possible for some, for others who have an extremely high level of protection, available scientific evidence may still be “insufficient” so that provisional measures can be in order)? 

In Hormones (para. 124), the AB linked sufficiency of evidence to one’s level of protection and precaution (thereby implicitly linking 2.2 to 5.7).  In Agricultural Products (para. 80), as well, the AB stated that 5.7 is a “qualified exemption” from 2.2 (not 5.1).  However, in Apples (para. 179) the AB did link 5.7 to 5.1, finding -- as the GMO panel did here -- that once a risk assessment is possible, provisional measures under 5.7 become unavailable.  In that case, the AB explicitly rejected Japan’s reference – much like that by EC member states in GMOs -- to remaining “scientific uncertainty”.  Relatedly, in Salmon (para. 138), the AB did link risk assessment to sufficiency of evidence by finding that a violation of 5.1 necessarily implies a violation also of 2.2. 

There is, in other words, a clear tension between Hormones and Agricultural Products, on the one hand, and Apples and Salmon, on the other. Yet, I see three possible ways to distinguish GMOs from Apples. 

First, in Apples we were talking about one WTO member (Japan); in GMOs, we have the EC and its member states (a risk assessment by the former may, therefore, not necessarily preclude provisional measures by the latter; the ongoing panel on EC – Customs may shed more light on the relation between the EC and its member states).  Second, the importance of the alleged risks is different:  Apples was ‘only’ about preventing fire blight in fruit; GMOs is, according to the EC, about a more vital question of human life (hence, scientific uncertainty under 5.7, even with a risk assessment available, may play a greater role in GMOs than in Apples, see Hormones, para. 124).  Third, in Apples, the only concern was plant life or health (a concern fully covered in Japan’s studies); in GMOs, in contrast, there are also environmental and moral concerns at play which may not be fully covered in the EC risk assessment.  These additional concerns fall arguably even outside the SPS agreement altogether, raising the separate question of whether, for example, GATT Art. XX(a) on public morals (as indirectly interpreted in Gambling) might justify the safeguards.  SPS does, of course, prevail over GATT in the event of conflict (Ad Note), but if one is really talking about entirely distinct justifications for a certain conduct, can there still be conflict, or are we then talking about two distinct measures so that one valid justification (under GATT) could save the safeguard even if another justification (under SPS) cannot?

I would, therefore, consider it wise for the EC to appeal the panel ruling on GMOs, a ruling which is – from the perspective of EC member states – far from “purely historical” or “irrelevant”. 

Comments

[Sungjoon Cho]

These are very fine "legal" points, but I would like to focus more on the general desirability of litigating this type of case under the WTO dispute settlement system.

First, this case might be a perfect example of the titular “wrong case,” as the late Robert Hudec aptly coined, which eventually undermines the WTO tribunal’s judicial integrity and legitimacy on account of the case’s politically combustible nature. As in the domestic legal system, certain disputes should be addressed in a non-judicial mode, either by negotiation or by other types of deliberation. Article 3.7 of the WTO Dispute Settlement Understating (DSU) also provides that “before bringing a case, a Member shall exercise its judgment as to whether action under these procedures would be fruitful.” In this type of case, regulatory experts, not diplomats, from both sides should be given enough time to conduct collective professional deliberation in a workman-like fashion, having recourse to any relevant international standards available.

Second, even if fully adjudicated, it is highly unlikely that the WTO panel or the future Appellate Body decision would truly end the underlying genetically modified (GM) food dispute. The notorious DISC case regarding the U.S. tax incentives for its exporting companies was first undertaken by a GATT panel in the Seventies and nearly short-circuited the whole GATT panel system. Yet, the very same dispute was reincarnated after more than two decades in the FSC case under the WTO. Although the WTO’s reinforced retaliation mechanism may look like an improvement, it also tends to escalate political tensions throughout the course, rather than defusing them.

Third, the public in both sides should be more informed and educated in this dispute. No due place for the public can be found both in the legal debate and political fray on this case. How many public hearings or discussions have been held on the GM foods? How much do people in both sides of the Atlantic know about the GM foods? Without such enlightening process, the public remains a victim of ignorance or fear, and thus cannot deliver informed decisions. That’s why governments in both sides should redouble their efforts toward adequate risk communication on this issue.

In sum, the GM debate should not, and cannot, be a mere legal case under the WTO. The GM case now carries with it a thick layer of political attention whose burden might break the back of the WTO’s hitherto well-operating dispute settlement mechanism. Now, it behooves both parties in this dispute to rid the WTO of such burden.

[Simon Lester]

The relationship between SPS 5.1 and 5.7 is an interesting one (I did not recall the AB's statement in para. 179 of Apples before Joost pointed it out). What intrigues me is the sequencing of these two provisions. When a Member wishes to invoke 5.7, must it first attempt a risk assessment, and then only after finding insufficient available evidence go to 5.7? Or can it just go straight to 5.7 whenever it believes, based on a general sense of things, the evidence is insufficient? It's not clear to me, based on para. 179, what the AB would say on this point.

I'm not sure of the facts in GMOs, but if the answer is the first one, it may be problematic that the EC member states did not make enough of an effort to conduct an actual risk assessment before taking the measures here.

[Aaron Ostrovsky]

Prof. Pauwelyn's point regarding who conducts the risk assessment is similar to the problems within the European system currently. As it stands, under the Deliberate Release Directive, if one European member state decides that a GMO is "too risky" to release into its territory, the decision may move to the Community level (or to another member state if the releasing party attempts to apply in another state). In any event, one may be confronted with the scenario of the risk assessment of a risk-preferring Member State being thrust upon a risk-averse Member State - thereby creating the kind problem Prof. Pauwelyn discusses under Art. 5.7 SPS.

[Sungjoon Cho]

Regarding Simon's question, I think the threshold question for any regulating Member to ask in the first place is whether there is "sufficient" scientific evidence available out there. Logically, this inquiry should be done in a pre-risk assessment stage since the Member will any way need certain scientific evidence with which to conduct a risk assessment. If the Member finds enough data or studies, it should assess the risks by means of such data or studies and base their regulation (e.g., import restriction or ban) on that assessment (art.5.1) However, if the Member concludes that there are few data or studies available out there, it may have recourse to a provisional measure (e.g., import restriction or ban) (art. 5.7) Of course, this threshold inquiry on the availability of evidence is the Member's call which can be rejected later by a panel as in EC-Biotech case. If the Member turns out to have overlooked available evidence, it violates not only art. 5.7 but also art. 5.1 since it should have used such evidence to assess the risks. Finally, art.5.7 provides an extra requirement, i.e., an ex post risk assessment with additional information. Since an SPS measure under art.5.7 is a "temporary" one, the regualting Member should resume a risk assessment with new evidences. Japan was found to violate this portion of art. 5.7 in Agricultural Products (1999) case. Having said that, could we understand that art.5.1 is a general obligation and art.5.7 is an exception? If so, would the third and fourth prongs of art.5.7 requirement be something like the chapeau (as in GATT Article XX)?

[Joel P. Trachtman]

Thanks, Joost. Just a question re Aaron's interesting reference. As I understand the issue, it is not (in the WTO) an issue of one state's risk assessment being thrust on another. Rather, the issue is whether one state's (or the EC's) scientific evidence may be assessed in determining the sufficiency of scientific evidence in connection with another state's assertion of 5.7. This seems far less problematic than forcing one state's risk assessment on another. So, in the EC, is it risk assessments or evidence that may be transferred?

[Aaron Ostrovsky]

In response to Prof. Trachtman's question (and someone else please correct me if this is wrong), the European system works like this: the releasing party makes an application to a Member State; that Member State then submits an assessment report to the Commission (with a yes or no recommendation for release); the Commission notifies all other Member States who then must agree or disagree to the release. If one Member State objects, a committee chaired by the Commission and comprised of Member States is formed - this committee must agree by qualified majority to allow the release (comitology). If the committee cannot agree one way or the other, the Commission then takes sole control of the approval process and the original Member State's duty at this point is to simply "ratify" the Commission's decision.

In this process, it is the risk assessment that is in question: how much risk a state is willing to tolerate, not just what evidence the risk assessment is based on. In the "Greenpeace France" case in front of the ECJ, the question was whether France could change its assessment from positive to negative after the approval process had reached the Commission (community) stage. France had had a change in government during the process and the new government was much more risk averse than the old one - based on the same raw data, it had decided it no longer wanted to approve the release. France lost the case based on an interpretation of the Deliberate Release Directive - but the interesting side note is that France had no problem declaring sua sponte that it no longer thought the risk was managable although conditions had no really changed. The decision highlights how political or policy oriented a stance on risk assessment really is. And the result can be that one state's normative preferences for risk can indeed be transferred to another, regardless of the metrics (or data) used to assess the risk.

[Tracey Epps]

Aaron Ostrovsky has raised a critical point when he refers to how political or policy oriented a stance on risk assessment really is. I have yet to examine more closely how disputes panels and the Appellate Body have characterized the term 'risk assessment', but it seems to me that it is crucial that when we use the term, we remember that in almost any SPS dispute, we are not dealing with easily ascertainable scientific "truths". Rather, most risk assessments will be full of data gaps and uncertainties. They are not therefore statements of fact about impact on health but judgments about what evidence to believe. Based on the same data, different countries using different policy assumptions will come up with widely varying interpretations and consequently varying standards. This point often seems to be overlooked, but when we focus on it, the questions raised by Joost Pauwelyn regarding Article 5.7 become even more murky (in terms of determining when a risk assessment is "feasible"). It also seems to provide support for Sungjoon Cho's argument that the GM debate should not be a mere legal case under the WTO because if we look closely at the risk assessment process we see that in a situation where there are as many uncertainties as there are surrounding gm foods, the process is inevitably going to be laden with value judgments that would be best acknowledged and dealt with openly in other fora.