The Measure at Issue

First-generation, or conventional, biofuels are produced from food or feed. In addition to contributing to GHG emissions associated with land use change,Footnote ³ they may adversely affect food security by diverting land from food and feed production. These controversial effects of conventional biofuels are reflected in the EU biofuels regime.

The EU’s Renewable Energy Directive in effect at the time of filing of the panel requests (RED II) imposed on the EU member states an obligation to bring the share of renewable energy in the final consumption of energy in the transport sector to at least 14 percent by 2030.Footnote ⁴ (This number has been revised up in the currently effective RED III.)Footnote ⁵ However, for compliance purposes, energy from conventional biofuels cannot account for more than 7 percent of the totalFootnote ⁶ (the “7 percent maximum share”) and is subject to certain sustainability and GHG emissions saving criteria.Footnote ⁷ Within the entire category of biofuels produced from food and feed crops, RED II specifically singles out “high ILUC risk” biofuels. Their eligible share, initially also capped at 7 percent, must gradually decrease to 0 percent by December 31, 2030 (the “high ILUC risk cap and phase-out”), with an exception for individual batches that have been certified as “low ILUC risk.” RED II instructed the Commission to define, by February 1, 2019, the criteria for identifying specific high ILUC risk feedstock.Footnote ⁸

The Commission established two criteria. A feedstock is designated “high ILUC risk” if: (1) its global production area has increased by more than 1 percent and by more than 100,000 ha per year on average since 2008 (criterion 1); and (2) more than 10 percent of this expansion was into land with high-carbon stock, i.e., forests and wetlands (criterion 2).Footnote ⁹ A factor of 2.6 is applied to the share of wetlands to reflect their higher carbon content relative to forests. The resulting sum of the shares is divided by a productivity factor to reflect different energy yields for different feedstocks.Footnote ¹⁰

In 2019, the Commission determined that only palm oil, with an average expansion rate of 4.0 percent or 702,000 ha per year, met both criteria as its share of expansion into forests and wetlands was estimated at 45 percent and 23 percent, respectively.Footnote ¹¹ The productivity factor of 2.5 proved insufficient to make up for a major breakaway from soybeans, the runner-up with a 8 percent expansion into forests and no expansion into wetlands.Footnote ¹² On this basis, palm oil, alone, was designated “high ILUC risk” feedstock. Even though this does not entail a ban on palm biodiesel, the fuel’s ineligibility under RED II (and RED III) will eliminate demand for it because this demand exists only insofar as the directive requires the use of biofuels in the member states’ energy mix (Malaysia, paras. 7.324, 7.328; Indonesia, paras. 7.336, 7.340). There is the option to certify a specific batch of palm biodiesel as low ILUC risk as such batches remain eligible to be counted toward the target of renewables in the transport sector. This avenue, however, allows at best only limited opportunities for export.

The WTO Panel’s Findings

Article 2.1 of the TBT Agreement lays down a non-discrimination obligation. It requires that “products imported from the territory of any [WTO Member] … be accorded treatment no less favourable than that accorded to like products of national origin and to like products originating in any other country.”Footnote ¹³ The complainants alleged that the high ILUC risk cap and phase-out accords less favorable treatment to their palm oil-based biofuel vis-à-vis rapeseed- and soybean-based biofuels originating in the EU and other WTO members. The panel stated it would apply the legal test that the Appellate Body established in US—Clove Cigarettes. According to that test, less favorable treatment is found where (1) a technical regulation at issue modifies the conditions of competition to the detriment of the complainant’s group of products and (2) that detrimental impact does not stem exclusively from a legitimate regulatory distinction.

As the first step of its analysis, the panel set out to determine whether the product groups put forward by the complainants were like, using the usual four criteria for assessing likeness. The panel concluded that: (1) “a number of physical characteristics of [the three types of biofuel], while not identical, are similar” (Malaysia, para. 7.431; Indonesia, para. 7.452); (2) the products shared the same primary end use as biofuel mixed with diesel fuel and were “nearly perfect substitutes” in warmer climates, although palm biodiesel is more rarely used in colder environments as it crystalizes and plugs the fuel filter at a higher temperature (Malaysia, paras. 7.437, 7.446; Indonesia, paras. 7.458, 7.467); (3) the EU had failed to convince the panel of significant consumer preferences against palm oil-based biofuels; and (4) all three biodiesels are classified under the same six-digit subheading. On this basis, the panel found palm oil-, soybean-, and rapeseed-based biofuels to be like.

The panel then determined that in the EU market, demand for biofuels, which are and will remain in the foreseeable future not cost-competitive with fossil fuels, exists only insofar as the EU renewable energy policy requires the blending of biofuels with fossil fuels. As a reduction of demand for high ILUC risk biofuels is an intended consequence of the high ILUC risk phase-out, the panel had no difficulty finding that the measure at issue has a detrimental impact on the competitive opportunities for palm oil biodiesel in the EU.

The panel’s further analysis of whether this detrimental impact stemmed exclusively from a legitimate regulatory distinction centered on the question of whether the concepts of “ILUC” and “high ILUC risk” as well as the two high ILUC risk criteria were sufficiently grounded in science (Malaysia, para. 7.501; Indonesia, para. 7.516). The panel formulated the ultimate question before it as “whether, considering the entirety of the evidence, there [was] a reasonable basis for the regulatory distinction drawn by the high ILUC-risk cap and phase-out and the manner in which it [was] applied” (Malaysia, para. 7.504; Indonesia, para. 7.519).

The panel said it would determine first whether the regulatory distinction (that “between different types of biofuels based on whether the feedstock used for its production presents a high degree of ILUC risk” (Malaysia, para. 7.508; Indonesia, para. 7.523)) could a priori not be legitimate” (Malaysia, para. 7.511; Indonesia, para. 7.526). This question, according to the panel, concerned the “conceptual underpinnings” of the high ILUC risk classification (Malaysia, para. 7.515; Indonesia, para. 7.530). The panel eventually found that there was a reasonable basis for the regulatory distinction reflected in the measure and that the distinction was a priori legitimate “on a conceptual level” (Malaysia, para. 7.545; Indonesia, para. 7.561). The panel reasoned that imperfections and uncertainties inherent in the methodology underlying the measure did not rise to the level of rendering the regulatory distinction not a priori legitimate because “the methodology as a whole [bore] a rational connection to the measure’s objective” (Malaysia, para. 7.526; Indonesia, para. 7.541). “Accepting a contrary proposition,” stated the panel, “would de facto deprive WTO Members of the right to regulate risks that do not lend themselves to quantitative analysis, or which are not sufficiently studied, even though such risks constitute a genuine cause of concern” (Malaysia, para. 7.527; Indonesia, para. 7.542).

The panel then turned to the “application of the regulatory distinction at issue,” i.e., the various parameters and values of the share-of-expansion (Malaysia, p. 184; Indonesia, p. 183). It eventually found no fault in either the criteria as such, including the use of a relative share rather than an absolute area of expansion into carbon-rich lands and the 10 percent significant expansion threshold, or the methodologies the Commission applied to obtain the 45 percent and the 23 percent shares of expansion of oil palm plantations based on 2008–2016 data and the palm oil productivity factor of 2.5 (Malaysia, paras. 7.546–7.589; Indonesia, paras. 7.562–7.612). In response to the complainants’ argument that reliance on past data to calculate the shares of expansion ignored more recent developments and resulted in an overestimation of the ILUC risk, the panel stated that it considered later data and scientific studies “relevant,” but concluded that the evidence on the record did not “put in question the data relied on by the European Union with regard to the 2008–2016 reference period” (Malaysia, paras. 7.564, 7.565; Indonesia, paras. 7.580, 7.581).

Nonetheless, the panel took issue with the very fact that the EU had not reviewed the data in a timely manner and found the EU had thus not applied the measure even-handedly (Malaysia, paras. 7.566–7.572; Indonesia, paras. 7.582, 7.588). In addition, the panel determined that certain low ILUC risk certification criteria and requirements were “overly vague and ambiguous, as well as incomplete” and a ten-year limit on eligibility for low ILUC risk certification disadvantaged palm oil (the only perennial crop among those listed in the delegated regulation), which takes several years to bear fruit (Malaysia, para. 7.604, 7.624–7.626; Indonesia, para. 7.627, 7.648–7.650).Footnote ¹⁴ These findings led the panel to conclude that the EU “ha[d] administered the high ILUC-risk cap and phase-out inconsistently with Article 2.1 of the TBT Agreement” (Malaysia, para. 7.636; Indonesia, para. 7.660).

In a separate opinion, one panel memberFootnote ¹⁵ disagreed with the majority’s conclusion that, save in these administration aspects, the measure was otherwise non-discriminatory. That member reckoned there was “an element of arbitrariness in singling out palm oil-based biofuel for the high ILUC-risk cap and phase-out when other types of oils, notably soybean, appear to pose the same alleged risk” (Malaysia, para. 7.1443; Indonesia, para. 7.1447).

* * * *

The Appellate Body’s interpretation of Article 2.1 of the TBT Agreement has long been subject of criticism in academic literature and meetings of the WTO Dispute Settlement Body. Critics argued that in transposing its jurisprudence under the General Agreement on Tariffs and Trade (GATT) into the TBT domain, the Appellate Body had misconstrued the TBT Agreement.Footnote ¹⁶ One expression of this misconception is the like products analysis which is exclusively concerned with the competitive relationship between products rather than their relevance to the policy objective of the technical regulation at issue.Footnote ¹⁷ More consequential, however, is the Appellate Body’s requirement that any detrimental impact on imported products stem “exclusively” from a legitimate regulatory distinction—a standard described as “very demanding,” “quite unwarranted,”Footnote ¹⁸ and “rais[ing] the bar very high, imposing a potentially impossible practical if not legal burden of proof on the defendant.”Footnote ¹⁹ Claiming to remain within the Appellate Body’s analytical framework, the Palm Oil panel applied the GATT-style likeness analysis, but appears to have abandoned the exclusivity standard for a laxer “reasonable basis” test.

In the early days of the WTO, the Appellate Body decided that in the non-discrimination provisions of the GATT 1994, product likeness would largely be determined by the products’ substitutability in the eyes of the consumer, so that, for example, “a determination of ‘likeness’ under [GATT] Article III:4 is, fundamentally, a determination about the nature and extent of a competitive relationship between and among products.”Footnote ²⁰ For this reason, regulatory considerations underlying a difference in treatment, unless they are reflected in consumer preferences, will normallyFootnote ²¹ only be taken into account under GATT Article XX “General Exceptions”Footnote ²² or another affirmative defense.

Petros Mavroidis has argued that a competition-based approach to likeness does not make sense under the TBT Agreement, as governments issue technical regulations, i.e., establish mandatory requirements for products, precisely because they want to exclude from the market some of the products (those not meeting the requirements) that consumers would otherwise be willing to buy.Footnote ²³ In other words, under the TBT Agreement, products should be deemed like so long as they have like effects on the achievement of the regulatory objective behind a measure (“policy likeness”).Footnote ²⁴ This, in fact, was the approach adopted by the panel in US—Clove Cigarettes. It assessed likeness of menthol cigarettes and clove cigarettes through the lens of their effect on smoking uptake by youngsters which the measure at issue sought to reduce.Footnote ²⁵ The Appellate Body, however, then rejected this approach in favor of competition-based likeness.Footnote ²⁶ Under the Appellate Body’s approach, a finding of likeness between and among products that meet the requirements of a technical regulation and those that do not is virtually automatic. If consumers are willing to substitute one for the other (if they were not, the government would not have to intervene in the first placeFootnote ²⁷), the products’ physical properties and end uses are apparently similar if not identical. The likeness step of analysis under Article 2.1 of the TBT Agreement then becomes a superfluous exercise that serves no purpose.

In the Palm Oil disputes, the futility of competition-based likeness test under the TBT Agreement is especially pronounced. The panel spills ink discussing the cold-filter plugging point and the cetane number of different types of biodiesel when the whole point of the technical regulation at issue is to distinguish biofuels based on their ILUC-related GHG emissions that have nothing to do with the physical properties of biofuels (Malaysia, paras. 7.427–7.446; Indonesia, paras. 7.447–7.467). The finding of competition-based likeness is useless for deciding the question before the panel—whether the measure accords less favorable treatment to Malaysia’s and Indonesia’s imported palm biodiesel. Under the Appellate Body’s approach, this discussion is unnecessarily deferred to the second step of the test, i.e., whether the detrimental impact “stems exclusively from a legitimate regulatory distinction.”

Having rejected the panel’s policy likeness test in US—Clove Cigarettes, the Appellate Body had to find a way to integrate the measure’s regulatory objectives into its discrimination analysis. As the TBT Agreement does not contain a general exceptions provision similar to GATT Article XX, the Appellate Body found a way out in declaring that Article 2.1 does not prohibit detrimental impact so long as it “stems exclusively from legitimate regulatory distinctions.”Footnote ²⁸ The stringency of this standard fully played out in US—Tuna II. It took the United States two amendments of its dolphin-safe labelling measure and two compliance proceedings for the panel and the Appellate Body to find that the measure was finally “calibrated” to risks to dolphins from different fishing methods in different parts of the ocean and thus met the non-discrimination standard.Footnote ²⁹ Among other things, the Appellate Body was then criticized for allowing a de novo assessment of the measure during the first compliance proceedings, which resulted in a finding of a lack of calibration to risks that had not been reviewed in the original dispute.Footnote ³⁰

The United States has persistently objected to the Appellate Body’s standard.Footnote ³¹ It has argued, including as a third party in the Palm Oil disputes, that the right test under Article 2.1 is whether the detrimental impact experienced by imports is explained by factors unrelated to the foreign origin of the product: “A measure does not discriminate based on origin simply because an adjudicator considers it could have balanced costs and benefits better or designed a more effective or perfect measure.”Footnote ³² Indeed, it is doubtful that (1) the intention of the WTO members was to prohibit every bit of inadvertent (yet unavoidable) detrimental impact on competition; and (2) adjudicators are in a position to assess whether this standard, if taken seriously and to the letter, is met.Footnote ³³

The Palm Oil panel found its way around the issue. Noting the lack of objection by the parties, the panel declared it would use the Appellate Body’s test (Malaysia, para. 7.406; Indonesia, para. 7.426). However, the panel never actually applied the test as it never engaged in the Tuna II-style “calibration,” or balancing, analysis by asking, for instance, whether the measure (or the regulatory distinction) at issue was calibrated to the ILUC risk of various biofuel crops. Instead, the panel stated it would determine “whether, considering the entirety of the evidence, there [was] a reasonable basis for the regulatory distinction drawn by the high ILUC-risk cap and phase-out and the manner in which it [was] applied” (Malaysia, para. 7.504; Indonesia, para. 7.519). Thus the panel’s focus was on palm oil and the existence of a “reasonable basis” for designating it as a high ILUC risk feedstock rather than on whether the measure’s treatment of the different biofuel crops was commensurate with their different ILUC risks.

Thus, the actual test for less favorable treatment that the Palm Oil panel applied was not much different from the one proffered by the United States. Judging by the panel’s reasoning, the question it answered was whether the detrimental impact on the competitive opportunities of palm oil-based biodiesel could be explained by (or stemmed, but not necessarily exclusively, from) a legitimate regulatory distinction. A legitimate distinction, in turn, is that for which scientific evidence provides a reasonable basis.

The panel’s approach was also novel in one other respect. It conducted a separate inquiry into the “a priori legitimacy” of the regulatory distinction as the initial step of its analysis. Previous panel and Appellate Body reports did not specifically review the legitimacy of a regulatory distinction—rather, adjudicators determined whether the measure was even-handed in its application, and if not, concluded that its detrimental impact did not stem from a legitimate regulatory distinction. The Appellate Body stated that “where a regulatory distinction is not designed and applied in an even-handed manner—because, for example, it is designed or applied in a manner that constitutes a means of arbitrary or unjustifiable discrimination—that distinction cannot be considered ‘legitimate.’”Footnote ³⁴

The fundamental purpose of the a priori legitimacy step of analysis was apparently not only to review the “conceptual underpinnings” of the measure but also to lay out the panel’s own conceptual precepts that would guide its further reasoning. One of them is the de facto pronouncement that in situations of scientific uncertainty, WTO members’ right to regulate prevails over the obligation to ground technical regulations in methodologically flawless science (Malaysia, para. 7.527; Indonesia, para. 7.542).

Finally, the panel’s conceptualization of ILUC risk as unrelated to the quantity of ILUC-associated GHG emissions merits a brief comment. The panel stated at the outset of its regulatory distinction analysis that “it is the high degree of ILUC risk that lies at the heart of the regulatory distinction at issue rather than, for example, quantification of the respective volumes of GHG emissions linked to ILUC caused by production of any particular biofuel feedstock” (Malaysia, para. 7.508; Indonesia, paras. 7.523). The panel repeats this thesis multiple times throughout its analysis (and italicizes the word “risk” twice (Malaysia, paras. 7.306, 7.509; Indonesia, paras. 7.315, 7.524)) to reject several of the complainants’ arguments (Malaysia, paras. 7.508–7.509, 7.528, 7.555; Indonesia, paras. 7.523–7.524, 7.543, 7.571). It never explains, however, how the quantity of GHG emissions may be irrelevant to an analysis of a measure whose objective, as the panel itself determined, was to “limit the risk of ILUC-related GHG emissions associated with crop-based biofuels” (Malaysia, paras. 7.218, 7.267; Indonesia, paras. 7.218, 7.276).

* * * *

To summarize, the Palm Oil panel could have saved itself time and effort (and do justice to the object and purpose of the TBT Agreement) by adopting the policy-based likeness test. The assessment of the scientific evidence and of its interpretation by the Commission, which the panel undertook as part of its “reasonable basis for the regulatory distinction” analysis, would have become the core of the panel’s likeness determination. The panel would presumably have found that, due to its particularly high ILUC risk, palm oil-based biofuel, on the one hand, and rapeseed- and soybean-based biofuel, on the other, were not like products, and the analysis under Article 2.1 would have ended there. The issues the panel identified with the administration of the measure (the lack of a timely review of the underlying data, the deficiencies in the ILUC certification criteria and requirements) could then be addressed separately under other provisions of the TBT Agreement (e.g., Article 2.2, which prohibits technical regulations that are more trade restrictive than necessary to fulfill a legitimate objective).

As the panel adopted the Appellate Body’s competition-based likeness test, it found the products were close competitors and therefore like, which necessitated a separate analysis of whether the detrimental impact on the competitive opportunities for PME stemmed exclusively from a legitimate regulatory distinction. The panel, however, applied a modified version of this test by tacitly abandoning the demanding exclusivity requirement and framing its assessment in terms of whether there existed a reasonable basis for the regulatory distinction between palm biodiesel and the other biofuels. The panel also introduced an additional initial step of analysis—an inquiry into the “a priori legitimacy” of the regulatory distinction, which it used to formulate a couple of foundational precepts that would inform significant aspects of the panel’s further assessment of the regulatory distinction at issue. It remains to be seen if the panel’s approach takes root in future disputes.