The WTO Appellate Body Report on the WTO compatibility of the EU’s anti-dumping measures on Biodiesel from Argentina was released on 6 October 2016. The Appellate Body confirms the findings of the dispute settlement panel from 29 March 2016.
In May 2013 the EU had imposed anti-dumping duties on imports of biodiesel originating in Argentina. The normal value of biodiesel in Argentina was constructed on the basis of Article 2(5) of the EU’s anti-dumping Regulation (Regulation 1225/2009).
When Normal Value is constructed investigating authorities have to find the different costs to make up the cost of production in the country of origin. In Argentina Biodiesel is made from soya. During the period of investigation Argentina imposed an export tax on soya. For this reason, the EU Commission did not use the Argentinian price, i.e. the cost of soya entered into the accounts of the producers on the basis that:
the domestic prices of the main raw material used by the biodiesel producers in Argentina were …. lower than the international prices due to the distortion created by the Argentinian export tax system and, consequently, the costs of the main raw material were not reasonably reflected in the records kept by the Argentinian producers under investigation in the meaning of Article 2(5)…
Rather than use the cost on the accounts of the producers the Commission used the reference price used by the Argentinian government for the calculation of the export tax as this price reflected international prices. In other words, the Commission replaced the cost actually paid with a cost that reflected the international price.
Article 2(5) of the Basic Regulation allows the Commission to replace costs when constructing the normal value. The question before the AB was whether this was compatible with WTO law and in particular with the terms of GATT Article VI and Article 2.2 and Article 188.8.131.52 of the WTO Anti-Dumping Agreement.
The Appellate Body found that the EU did not comply with the provisions of WTO law in determining the level of the the anti-dumping duty. Three aspects of the case stand out.
i) the meaning of the words: ‘reasonably reflect’;
ii) the use of information from other markets;
iii) the fair comparison between the normal value and the export price.
The meaning of the term: ‘reasonably reflect’
Article 2 of the Anti-Dumping Agreements defines dumping. It is when a product is introduced into the export market at a price less than the normal value in the set in the ordinary course of trade.
Article 2.2 of the Anti-Dumping Agreement allows for the construction of the normal value when using the price in the country of origin is no appropriate. It provides:
1. When there are no sales of the like product in the ordinary course of trade in the domestic market or when, because of a particular market situation or the low volume of the sales in the domestic market of the exporting country, such sales do not permit a proper comparison, the margin of dumping shall be determined by comparison with a comparable price of the like product when exported to an appropriate third country, provided that this price is representative, or with the cost of production in the country of origin plus a reasonable amount for administrative, selling and general costs and for profits.
Article 184.108.40.206 of the Anti-Dumping Agreement provides (in part):
For the purposes of paragraph 2 [which allows the construction of normal value], costs shall normally be calculated on the basis of records kept by the exporter or producer under investigation, provided that such records are in accordance with the generally accepted accounting principles of the exporting country and reasonably reflect the costs associated with the production and sale of the product under consideration.
The question that the AB had to address was the correct interpretation of the phrase ‘and reasonably reflect the costs associated with the production and sales of the product under consideration’.
The EU argued that the phrase ‘reasonably reflect’ incorporated the idea that the cost in question must be ‘reasonable’ and not in some way distorted. The Panel and the AB rejected this approach and found that the adverb ‘reasonably’ modifies or clarifies the word ‘reflect’ and not the word ‘cost’.
In doing so the AB rejected the argument that Article 2 of the Anti-Dumping Agreement is imbued with a general standard of ‘reasonableness’ which endows the investigating authority with discretion, within the terms of Article 220.127.116.11 to disregard the records kept by the exporter or producer when the authority considers that the costs recorded in the accounts are not reasonable.
The core of the AB’s reasoning is that Article 2.2 determines when normal value can be constructed (in other words when it is permissible not to use the price in the market of origin as the normal value). The construction of the normal value must arrive at a cost of production in the country of origin. Any interpretation of Article 2 or of Article 18.104.22.168 must be seen in that light. The binding obligation is to arrive at a cost of production in the country of origin.
There are a number of broad reflections on the AB’s approach.
It is clear that the AB saw its role in determining whether or not the Panel had erred in its interpretation of the law, even if the AB did undertake its own analysis. However, that analysis was framed by the analysis of the Panel. On this basis the AB did not address the significance of the word ‘normally’ in the first line of Article 22.214.171.124.
Nor does the AB address the definition of dumping in Article 2.1 which clearly provides that the normal value or the comparable price, must be set in the ‘ordinary course of trade’. The ‘ordinary course of trade’ test for determining the quality of the normal value was not addressed by the AB. A question that remains to be addressed it the nature of the ‘ordinary course of trade’ test and whether it is of a higher normative value that the detailed provisions of Article 126.96.36.199.
That being said it seems reasonable to consider that a normal value that is not determined on the basis of costs which result from the ordinary course of trade does not comply with the definition of dumping in Article 2.1. While the Panel and the AB may well have been correct in their finding that the word ‘reasonably’ applies to the verb ‘reflect’, this is to ignore the higher norm to ensure that the normal value is set in the ordinary course of trade.
Paragraph 6.41 of the AB report. It is worth considering this paragraph as a whole:
We do not subscribe to the European Union's reading of the Panel Report. We note that, to the Panel, the second condition in the first sentence of Article 188.8.131.52 of the AntiDumping Agreement requires a comparison between the costs in the producer's or exporter's records and the costs incurred by such producer or exporter. The Panel emphasized that "the object of the comparison is to establish whether the records reasonably reflect the costs actually incurred, and not whether they reasonably reflect some hypothetical costs that might have been incurred under a different set of conditions or circumstances and which the investigating authority considers more 'reasonable' than the costs actually incurred." In this connection, the Panel explained that its understanding of this condition does not imply that "whatever is recorded in the records of the producer or exporter must be automatically accepted.” To the Panel, an investigating authority is "certainly free to examine the reliability and accuracy of the costs recorded in the records of the producers/exporters" to determine, in particular, whether all costs incurred are captured; whether the costs incurred have been over- or understated; and whether non-arms-length transactions or other practices affect the reliability of the reported costs. The Panel further stated that "Article 184.108.40.206 does not involve an examination of the 'reasonableness' of the reported costs themselves, when the actual costs recorded in the records of the producer or exporter are otherwise found, within acceptable limits, to be accurate and faithful." In light of these statements, we consider the Panel's interpretation of Article 220.127.116.11 to be more nuanced than the European Union's argument suggests.
In this paragraph the AB is signalling that even if there is no reasonableness test in Article 2.2 it is still possible to reject costs on the records (whether reasonable or not) where not to do so would be to use costs which were in someway distorted or would not allow for the proper construction of normal value in line with the ordinary course of trade norm. However, the AB has not given investigating authorities much guidance as to how to apply this opening.
Information from other markets
Argentina argued that Article 2.2 of the Anti-Dumping Agreement does not allow any information to be used in determining the normal value other than the producers’ costs in the country of origin. The Panel found that neither Article 2.2 of the AntiDumping Agreement nor Article VI:1(b)(ii) of the GATT do not limit the sources of information that may be used in establishing the cost of production. What is required is that the constructed normal value reflects the cost of production in the country of origin. The AB agreed with this proposition.
The AB is probably right on this approach particularly in the light of Article 16.6(ii) of the Anti-Dumping Agreement which provides that where the law admits of more than one interpretation a Panel should not second guess the interpretation of an investigating authority so long as it rests on one of the possible interpretations.
That being said the result of the AB’s reasoning is not clear. The second sentence of Article 18.104.22.168 provides that that the authorities shall consider all available evidence on the proper allocation of costs. The AB considers that this sentence suggests that the evidence used to establish a cost can be different from the cost itself. It is not at all clear what this means.
In paragraph 6.73 the AB finds:
This suggests that, in such circumstances, the authority is not prohibited from relying on information other than that contained in the records kept by the exporter or producer, including in-country and out-of-country evidence. This does not mean that an investigating authority may simply substitute the costs from outside the country of origin for the costs of production on the country of origin.
How should this finding be interpreted? In the subsequent section of the Report the AB examines whether the replacement of the domestic soya price with a price reflecting the international price was legitimate. The AB found that it was not.
In conclusion the AB found that:
When relying on out-of-country information to determine the “cost of production in the country of origin” under Article 2.2, an investigating authority has to ensure that such information is used to arrive at the “cost of production in the country of origin”, and thus may require the investigating authority to adapt that information.
The AB gives no guidance to the competent authority as to how to ‘adapt’ the information so as to determine a ‘cost’ for the purposes of determining the normal value.
The approach of the AB in EU-Biodiesel is not significantly different from the approach in EC-Fasteners (21.5) even if the former deals with the costs used to construct normal value and the latter addresses adjustments to the normal value itself and the one was dealing with a market economy and the other with a non market economy methodology.
In EU-Biodiesel, the AB is saying that there is the possibility of not using the costs on the accounts on the producers (even if they are reasonably reflected) and that information can be based on out-of-country costs but that the out-of-country information needs to be adjusted before being used to construct a value that reflects the cost of production in the country of origin.
In EC-Fasteners (21.5) the AB found that adjustments to the normal value must be made if they are claimed and if they are valid. In that case adjustments were claimed for access to raw materials (import tax in India, none applicable in China), efficiency in electricity consumption, productivity per employee (see paragraph 5.236 of the Report). However, the AB did put a limit on the adjustments to be made. Adjustments should not be made where the effect of the adjustment would be to bring the cost back to the costs in China which were distorted in the first place. In other words, the adjustment should not be a means to defeat the purpose of using an analogue country to determine normal value.
The question arises whether the AB trying to align the concept of adjustments whether the adjustment is to a cost or the resulting normal value and whether the methodology is for a market or a non-market economy? Is the AB anticipating the correct interpretation of the consequences of changes in text of Article 15 of China’s Protocol of Accession?
A fair comparison
In paragraphs 6.47 and 6.48 the AB is very keen to draw a distinction between the manner in which the normal value must be constructed and the provisions of Article 2.4 on the fair comparison between the normal value and the export price.
What the AB rejects is the idea that if a tax (in this case and export tax) can be taken into consideration within the scope of Article 2.4 it can also be taken into consideration within the terms of Article 22.214.171.124. This seems a reasonable approach. Article 126.96.36.199 is about constructing the normal value not about comparing that value with the export price for the purposes of ensuring that the comparison is fair.
The AB returns to the issue Article 2.4 in section 188.8.131.52 of the Report. In this section of the Report the AB was addressing the argument raised by Argentina that the EU’s methodology for determining normal value (now found to be incompatible with Article 184.108.40.206) meant that the comparison was not fair.
Here the AB strongly affirms the separation of the two provisions (Articles 220.127.116.11 and 2.4) and holds that ‘The text of Article 2.4 makes clear that “[due] allowance shall be made in each case on its merits”.
In paragraph 6.89 the AB states that as it has already found a WTO inconsistency there is no need to opine on Article 2.4.
Given these findings, and notwithstanding our reservations about certain aspects of the Panel’s analysis under Article 2.4 of the Anti-Dumping Agreement, we do not consider it fruitful, in the particular circumstances of this dispute, to examine further whether the EU authorities also failed to conduct a “fair comparison” in comparing the constructed normal value to the export price.
One wonders how the AB would have analysed Article 2.4 in the light of the circumstances of this case. The AB has ruled on ‘reasonably reflect’ but has left open the idea that the costs used to construct normal value need not necessarily be the costs on the account (even apparently these costs reasonably reflect the costs incurred). The AB has held that information from outside the country of origin can inform the analysis of the choice of costs to be used to construct the normal value. And the AB has refused to apply Article 2.4. Is the AB leaving this paragraph of the Anti-Dumping Agreement as a vehicle for addressing situations where there is a feature of the market of origin which is not present in the market of export and thus rendering the comparison not fair.
In EC-Bed Linen, the AB when examining the provisions of ADA Article 2.4.2, argued:
Article 2.4 sets forth a general obligation to make a "fair comparison" between export price and normal value. This is a general obligation that, in our view, informs all of Article 2….
In US-anti-dumping measures on hot rolled steel from Japan, the AB interpreted Article 2.4 widely:
Article 2.4 of the Anti-Dumping Agreement provides that, where there are "differences" between export price and normal value, which affect the "comparability" of these prices, "[d]ue allowance shall be made" for those differences. The text of that provision gives certain examples of factors which may affect the comparability of prices: "differences in conditions and terms of sale, taxation, levels of trade, quantities, physical characteristics, and any other differences". However, Article 2.4 expressly requires that "allowances" be made for "any other differences which are also demonstrated to affect price comparability" [emphasis added]. There are, therefore, no differences "affect[ing] price comparability" which are precluded, as such, from being the object of an "allowance".
It is clear that competent authorities are allowed to make allowances for any factor that makes the comparison unfair. The overriding objective is to ensure fairness in the comparison between the two dumping comparators. Would the AB have made an allowance for the fact that the soya price was distorted in Argentina and not in the EU?
The ruling of the WTO Appellate Body in EU-Biodiesel will have a profound impact on how countries construct normal value in anti-dumping cases in the future. It will also colour the debate in the EU on the changes to normal value calculation methodologies consequent on the expiry of part of Article 15 of China’s protocol of accession to the WTO. The Appellate Body has closed the door in relation to current EU practice but appears to have opened the way for different approaches. This is the second ruling from the AB in the course of this year in relation to adjustments to normal value. Both judgements seem to go in the same direction. Investigating authorities must pay greater attention when adjusting the normal value itself and the prices and costs that can be used to construct the normal value.