2. What did the panel address?

The salient issues, in my view, that the panel addressed begin with the inconsistency (as such and as applied) with Art 9.3 of the ADA and Art VI: 2 of the GATT. This restricts the amount of the AD to being less than or equal to the dumping margin.

Secondly, Articles 6.10 and 9.2 of the ADA are considered. Article 6.10 concerns appropriate notification of interested parties, including the opportunity to respond under assurance of confidentiality. Article 9.2 stipulates the non-discriminatory collection of duties in appropriate amounts. A specific concern is the rebuttable proposition under US law in which all companies in an NME are presumed to belong to a single entity.

The panel also revisited the decision of the panel in US‒Shrimp I (Viet Nam) regarding Article 6.8 and Annex II(1)of the ADA. This pertains to the use of facts available (FA) when an ‘interested party’ declines to provide ‘necessary information within a reasonable period or significantly impedes the investigation’.

The panel also considered claims under Articles 9.4, 11.1, 11.2, 11.3, and 18.1 of the ADA. Additionally, claims under Articles 3.8 and 19.1 of the Dispute Settlement Understanding were addressed. Section 129 (c)(1) of the Uruguay Round Agreements Act of the US in 1994 also was considered. These either were discussed thoroughly by Broude and Moore, or I have nothing of interest to introduce.

3. A few pertinent facts

The original investigation named Brazil, China, Ecuador, India, Thailand, and Vietnam as producers of subject merchandise, and commenced on 31 December 2003. The final report of the US International Trade Commission (USITC) was issued in January of 2005.Footnote ⁹ The USITC determined that approximately 95% of US production was wild-caught and the remainder was farmed. That farming was not used more extensively in the US was reported to be due to a less suitable climate, higher prices of coastal land, and more restrictive environmental regulations than in the named countries. The harvesting season extended from May to December in the US.

There were 125 firms identified as domestic processors of shrimp, and 190 firms reported to be shrimp fisherman. Fresh, frozen, and canned warmwater shrimp were deemed to constitute the domestic industry. That is, the interests of the fisherman and the processors were coincident. There were 36 purchasers of processed warmwater shrimp identified in the US market. Thus there was no monopsony power amongst buyers.

There were 10 product categories identified in the investigation. Subject imports appeared to be differentiated primarily by processing (canned or frozen) and size. Within each category, the good appeared to be approximately homogeneous.Footnote ¹⁰

Thirteen processing firms were identified in Brazil, 28 in China, 9 in Ecuador, 83 in India, 37 in Thailand, and 36 in Viet Nam. These constituted from 46.7% of exports to the US in Brazil to 97.1% in Viet Nam. Unlike the US industry, the ‘vast’ majority of subject imports was farmed. The production of aquaculture was deemed to be ‘less seasonal’, and to have a longer harvest period. While the harvest seasons in the named countries were not coincident, the USITC reported that an imported product category was generally available from a named source at all times.

4. Final modification for reviews and the final rule

Because of the publication of the Final Modification, the panel found that Viet Nam had failed to establish that the use of zeroing by the DOC was ‘as such’ inconsistent with Article 9.3 of the ADA and Article VI:2 of the GATT 1994. The legal standard for an ‘as such’ determination is ‘a measure of general and prospective application’. This is irrespective of practice, as there cannot be a circumstance in which the policy is not implemented. The panel appeared to believe that disuse of zeroing would become the ‘general and prospective application’.

In the Final Modification, the DOC refers to retaining ‘considerable discretion in selecting an appropriate comparison methodology in reviews’. It also states that ‘It is not necessary, appropriate, or desirable’ to adopt a ‘total prohibition of zeroing’, as the Final Modification satisfies ‘the WTO settlement reports at issue’. This seems eerily similar to the incrementalism of the compliance by the US in the litany of disputes in which it has been a respondent. Importantly, the Final Modification discloses that the DOC ‘will continue to work closely and actively with USTR with a view towards clarifying that the AD Agreement should not be read to require WTO members to provide offsets for non-dumped comparisons’. There is no surrender here.

In response to a request to include regulation of targeted dumping in the Final Modification, the DOC dismissed it as being ‘unrelated’.Footnote ¹¹ This view was reiterated in the Final Rule issued by the DOC on 22 April 2014.Footnote ¹² Under this Rule, targeted dumping, which is defined as exhibiting a pattern of significant differences in export prices among purchasers, regions, or time periods, permits use of zeroing.Footnote ¹³ In its justification for this Rule, the DOC refers to the need for ‘discretion’, the benefits to ‘refinement’ of its approach through ‘experience’, and the necessity of ‘improving its ability to identify and address masked dumping’.

This is worrisome, as Ahn and Messerlin (Reference Ahn and Messerlin2014) have noted that the DOC recently has utilized targeted dumping in investigations against ‘Bottom Mount Combination Refrigerator-Freezers’ from KoreaFootnote ¹⁴ (final determination on 26 March 2012), and against ‘Large Residential Washers’ from KoreaFootnote ¹⁵ (final determination on 26 December 2012). Thus I concur with Ahn and Messerlin (Reference Ahn and Messerlin2014) that the definition of targeted dumping lacks discipline, and add that so does ‘necessary’, ‘appropriate’, and ‘desirable’.

That the panel did agree with Viet Nam that the use of zeroing by the DOC was ‘as applied’ inconsistent with Article 9.3 of the ADA, and Article VI:2 of the GATT 1994 is of limited consolation. While disagreeing with the ‘as such’ allegation may not be of operational significance in the context of this dispute, it is of substantial significance with regard to the deterrence of future use of this method of computing AD margins. Thus the panel left the DOC with too much room to breathe.

5. What can economics contribute?

Although there is no statutory requirement in the ADA to determine the reason(s) a firm may have engaged in third degree price discriminationFootnote ¹⁶ or may have charged a price below that of unit costs, it is instructive to consider this in the context of the reasonableness of the DOC's AD margins.

Given that the shrimp market is approximately perfectly competitive, third degree price discrimination (Viner, Reference Viner1923), is not relevant. This requires imperfectly own price elastic demand functions that are reflective of market power, which is arguably absent in this industry. The reciprocal dumping explanation of Brander and Krugman (Reference Brander and Krugman1983) requires international shipping costs and strategic interaction (oligopoly) of firms. It results in an export price net of production costs and international shipping costs that is less than the domestic price net of production costs. This is not relevant here.

Predatory dumping (Bernhofen, Reference Bernhofen1996; Hartigan, Reference Hartigan1994, Reference Hartigan1996) requires high entry and re-entry costs. This may provide a foreign firm with an incentive to set a price below the unit costs of domestic competitors so that a monopoly can be established. It requires a firm that can absorb a temporary sacrifice of profits, and is not relevant here either.

In the presence of cyclical demand, excess capacity, and high fixed costs as a percentage of total production costs, a firm may engage in episodic dumping to manage a recession in its (primary) domestic market (Ethier, Reference Ethier1982; Staiger and Wolak, Reference Staiger and Wolak1992). As the USITC reports, capacity utilization for the period of investigation (January 2001‒June 2004) averaged around 90%. The exports of shrimp constituted about 85% of total shipments of the foreign producers. Thus this does not appear to provide the motive for dumping.

It is plausible that a firm would introduce ‘an experience good’ to a new market through a low initial offer (Bagwell and Staiger, Reference Bagwell and Staiger1989).Footnote ¹⁷ The low introductory offer would provide inducement to consumers to try a product in lieu of others with established (positive) reputations. It does not seem to work with shrimp.

The explanation of dumping in agricultural and aquacultural markets by Hartigan (Reference Hartigan2000) seems instructive. This entails a perfectly competitive market in which expected profits over the course of a growing and harvesting period are zero. Actual profits could be positive or negative in the presence of an adverse or beneficial stochastic (probabilistic) shock. Firms contract with itinerant labor at the beginning of the planting season, so that all inputs essentially are fixed for the current season. Since variable costs would be very low, it would be rational to sell at a low price.Footnote ¹⁸ Production is zero when out of season. Production is also zero in the growing season.Footnote ¹⁹ During the harvest season, production gradually increases, reaches a peak, and gradually declines. If profits are expected to be zero over the harvest and demand is constant, a loss would incur at peak harvest due to higher supply relative to demand, thus inducing lower prices.

The average-to-transaction (A-T) method of calculating AD margins is particularly destructive in this context. Under this method, a weighted average of the normal value of foreign market prices is compared to individual export prices. As prices vary over the course of a harvest, this method inevitably yields dumping, and may produce high margins. This is a distinctive aspect of agricultural and aquacultural markets.

Article 2.2.1 of the ADA stipulates that prices below unit costs should ‘provide for the recovery of all costs within a reasonable period of time’. This article defines the ‘ordinary course of trade’ to be from six to 12 months. Even if the A-T method is eschewed, dumping can be found if a stochastic shock generates ex post losses despite the expectation of zero profits. The recovery of the losses would not be plausible until the harvest of the succeeding year. Article 2.2.1 provides that sales made below the unit cost of production (fixed and variable) plus average selling, general, and administrative expenses (SG&A) may, under certain circumstances, be considered ‘as sales that are not in the ordinary course of trade’.

This holds only if sales had been made ‘within an extended period of time’ (normally one year, but in no case less than six months), ‘in substantial quantities’ (if the weighted average of selling prices of the transactions under consideration is below the weighted average of per unit costs, or if the volume of sales at a loss represents at least 20 per cent of the volume of transactions), and ‘at prices which do not allow for the recovery of all costs within a reasonable period of time’. Sales made at prices that are below per unit costs at the time of sale, but above the weighted average per unit costs for the period of investigation, shall be taken into account as means for the ‘recovery of costs within a reasonable period of time’.

The wording of Article 2.2.1 ADA does not suggest that what is described above is the only case of sales that are not encompassed by the ordinary course of trade. Case law has provided additional examples of sales outside the ordinary course of trade.Footnote ²⁰

Must trade that is unfair be a matter of volition? Article 2.1 of the ADA refers to the introduction ‘into the commerce of another country at less than its normal value’. Article 8 of the ADA addresses the matter of price undertakings which permit suspension or termination of an investigation upon receipt of ‘satisfactory’ assurances from ‘any exporter’ that export prices have been revised to eliminate injurious effect to the domestic industry. However, Article 8.3 states that undertakings ‘need not be accepted’, such as when ‘the number of actual or potential exporters is too great’. This suggests that the prices of exports that ostensibly are a cause of injury to the domestic industry can be a matter of competitive pressures, and do not have to be intentionally unfair to qualify as dumping. If an exporter cannot commit to an undertaking because of competition, it is arguable that its ‘dumping’ is a response to that competition. Nonetheless, the seasonality of prices exhibited in an agricultural or aquacultural market should be regarded as normal and unintentional. Unless a harvest is unusually bountiful, resulting in prices that are below the norm, dumping should not have been deemed to have occurred from mere seasonality.

Because of the prevalence of aquaculture in the named exporting countries, and its limited profile in the US, the seasonality of production and attendant decline in prices, to the extent that it existed, was caused more plausibly by US wild-caught production. Although the prices at peak harvest might benefit the domestic processors at the expense of domestic fisherman, higher prices at other phases of the production cycle would be to their detriment. Since fisherman sell to local processors, direct competition is between the US processing industry and those of the named countries. Consistent with Hartigan (Reference Hartigan2000), the USITC reports that US processors historically have retained frozen supply for sale during the January to April non-fishing season at higher prices. However, the presence of farmed imported shrimp during this period has mitigated the value of their inventories. Nonetheless, the presence of imports, fairly or allegedly unfairly traded, would have this effect.

In accordance with Article 3.1 of the ADA, material injury must arise in the domestic market.Footnote ²¹ That is, it must be incurred by domestic producers of subject merchandise. Article 3.2 references a ‘significant increase in dumped imports’, which can be measured absolutely or relatively. Article 3.5 requires ‘a causal relationship between the dumped imports and the injury to the domestic industry’. If the wild-caught harvest of the US industry exhibited greater seasonality than imported production, and this was the source of the decline in domestic prices over which (perfectly competitive) firms have no discretion, it is arguable that a basis of Viet Nam's complaint should have included an Article 3 violation by the US.

6. Is market structure an available fact?

The US conducts AD investigations in accordance with Title VII of the Tariff Act of 1930. This act permits the characterization of economies as market or NMEs by US investigative authorities. It is not part of GATT VI or the ADA. It is US practice.

The prices in NMEs are deemed to be unreflective of scarcity, and, hence, fair value. The DOC's investigators are directed to utilize cost and price data from third countries in this instance. The DOC begins with a rebuttable proposition that all firms in an NME will receive the same economy-wide rate (EWR). Firms may qualify for a separate rate (SR) if they can establish to the satisfaction of the DOC that they are sufficiently independent of the government in exporting.

This matter arose in US‒Shrimp I (Viet Nam), and has been detailed and addressed compellingly by Broude and Moore (Reference Broude and Moore2013) with regard to sampling. However, given my introduction of market structure to the discussion, I have a few additional points to include.

The panel asserted that Viet Nam objected to the rebuttable proposition that in an NME all companies belong to a single entity. In this instance, the presumption must be disproven to the satisfaction of the DOC. Viet Nam also contested the manner by which the AD margins were computed, particularly the use of facts available (FA). The panel concurred with Viet Nam that the rebuttable proposition could be challenged ‘as applied’. However, it did not agree that the existence of a DOC practice for the calculation of AD margins for NMEs was established that was subject to an ‘as such’ challenge.

The panel did conclude that the NME-wide entity rate practice could be challenged ‘as such’ for inconsistency, specifically with regard to Articles 6.10 and 9.2 of the ADA as a matter of general and prospective application. In making this determination, the panel refers to the presumption of a single NME-wide entity. Would this conclusion have been reached in the absence of the presumption? That is, on the basis of the use of FA alone? Plausibly not. There has to be some method of calculating margins in instances of non-cooperation by named firms and/or the submission of inadequate data. Use of such a method is sanctioned by Article 6.8 of the ADA. However, the presumption is not. This method or methods may not have supported a conclusion that all exporters warranted the same AD margin in the absence of the rebuttable presumption.

So is market structure an available fact? Well no, not really. However, I contend that it may be discernible from the facts of the case, and, if so, should be utilized in assessment of the reasonableness of actions by an investigative authority. For example, the investigation will disclose the number of firms producing the named product, the existence of brands and warranties, and the extent of homogeneity of the subject good. The investigative reports of the USITC, for example, include results of a survey of purchasers of the subject import with regard to the importance of such matters as price, quality, interchangeability, safety, and availability. These are indicators of market structure.

Consider the EWR of 25.6%. This was established in the original investigation and has persisted through six administrative reviews. Given that the market is approximately perfectly competitive, it has to be prohibitive. It is unconceivable that the firms saddled with this impediment could compete in the US market. Indeed, even the 4.57% weighted average AD margin for mandatory respondents in the original investigation and the first three administrative reviews, and the 3.92% margin of the fourth administrative review are plausibly prohibitive in an approximately perfectly competitive market; that is, where goods are approximately homogeneous. While these separate rates (SR) may or may not be prohibitive, the EWR certainly is. This underscores the importance of the ‘individual margin’ referenced in Article 6.10, subject to the ‘impracticable’ constraint. The reference to ‘appropriate amounts’ in Article 9.2 is delimiting of the discretion of the DOC. It evidently is not a sufficient delimiter, as investigative authorities retain substantial discretion.

I concur with the panel's finding of ‘as such’ inconsistency with Articles 6.10 and 9.2 in the assignment of a single rate to all exporters of an NME under the presumption that all companies are members of an NME-wide entity. Moreover, I have provided additional justification for an ‘as applied’ determination of inconsistency. In this complaint, and in possible future complaints involving a large number of exporters of a reasonably homogeneous good, I recommend that an ‘as applied’ determination take market structure into account. It may, as a matter of policy, be an important distinction between agricultural (including aquacultural) and industrial production.Footnote ²² It is important, however, to be mindful of the extraordinary discretion that the ADA grants investigative authorities.

Of course, as market structure is contextual, this does not inform ‘as such’ judgments. As the panel asserted that its conclusions with regard to ‘as such’ inconsistency with Articles 6.10 and 9.2 ‘are applicable mutatis mutandis’ to a consideration of an ‘as applied’ allegation of inconsistency, one might wonder why I address this determination. My concern is prospective. Suppose that the US withdrew its presumption, or that all of the sources of its exports became market economies, as Viet Nam is doing. There may be future complaints for which an ‘as applied’ determination is relevant but an ‘as such’ verdict is not ‒ thus my discussion.

Is there a basis within the ADA for asserting that a panel could determine inconsistency of an investigation or review for not assessing the influence of market structure upon pricing and its effects? It does not state so explicitly, but the inference is supportable by its text. Article 2.2 refers to the ‘ordinary course of trade’. It is plausible that the ‘ordinary course of trade’ cannot be ascertained without consideration of market structure. Article 3.4 references ‘all relevant economic factors and indices having a bearing upon the state of the industry’. This should include market structure. Article 3.5 mentions ‘competition between domestic and foreign producers’. This occurs in the context of a market structure.Footnote ²³

The panels for US‒Shrimp I (Viet Nam) and US‒Shrimp II (Viet Nam) disagreed regarding interpretations of Article 6.8 of the ADA. This pertains to the use of FA when ‘necessary information’ is not provided ‘within a reasonable period’ or the investigation is ‘significantly’ impeded by ‘any interested party’. The panel for this dispute concluded that because the DOC applied the NME-wide rate computed in the original investigation to that entity in all six subsequent administrative reviews, it did not make any determination in reviews 4, 5, and 6, which were at issue in US–Shrimp II (Viet Nam). That is, FA were not invoked. The panel for US–Shrimp I (Viet Nam) concluded that ‘to fail to treat this rate as a facts available rate would elevate form over substance, and ignore the true factual circumstances surrounding the assignment of the rate’ for the earlier administrative reviews. I concur with the panel for US–Shrimp I (Viet Nam) and disagree with the panel for US–Shrimp II (Viet Nam). Deciding to utilize a rate from a preceding investigation is, in fact, a determination. What's more, it is more consistent with an objective of the promotion of trade and stare decisis,Footnote ²⁴ and leaves the beast with room to breathe.

7. Is this an AD case?

Although it would have been inappropriate for the panel to have considered whether or not an AD action was the proper venue for redress of the US shrimp industry's viability, it is instructive to speculate about the matter. However, the domain of the panel's analysis is the specific issues alleged by the complainant.

The US shrimp industry, at its discretion, could have filed a petition under Section 201 of the Tariff Act of 1974 instead of Section 731 of the Tariff Act of 1930. That is, it could have filed a safeguards (SG) petition rather than an AD complaint. The lesser injury standard (material vs. serious),Footnote ²⁵ the involvement of a sympathetic DOC, the absence of a provision for an allegation of nullification and impairment of expected benefits (compensation or retaliation), and the absence of a presidential decision may have been good reasons to file an AD petition.Footnote ²⁶

Consider, however, the purpose of SG. It is, plausibly, either of two objectives. One is to facilitate orderly exit from the production of subject merchandise. This would include reducing employment by attrition through extending the cessation of production over a longer period of time. The other is to facilitate restructuring. This would include plant closure, bankruptcy, and investment in new equipment (Bown, Reference Bown2002; Sykes, Reference Sykes2006; Hartigan, Reference Hartigan2015).

Recall the discussion of the domestic industry in Section 2 above. At the time of the original investigation, US shrimp were 95% wild-caught, and the ‘vast majority’ of subject imports were characterized as produced through aquaculture. Farming appeared to modulate seasonality and provide more consistent size and quality of harvest.Footnote ²⁷ That the US did not implement aquaculture to a greater extent was reputedly due to coastal land prices, climate, and environmental regulations. This appears to epitomize the need for orderly exit. Thus SG measures seemingly would have been more appropriate than AD duties. The facts did not portend a future globally competitive US shrimp industry.

However, this is changing, and in a way that underscores the appropriateness of SG as the petition that should have been filed. On 2 July 2014, the New York Times published an article about the emergence of the indoor shrimp aquacultural industry in the US.Footnote ²⁸ This evolving and innovative industry uses ‘fiberglass tanks fitted with advanced recirculation, filtration, and temperature control systems’. It relies upon tanks of sea water from the Atlantic Ocean. This water is ‘filtered and reused from harvest to harvest’. It does not utilize antibiotics, hormones, or pesticides. In the five years leading up to the article, the number of such farms has increased from two to at least 22 in the US. This suggests that the advantages in the technology of the US may overcome the disadvantages in labor costs, land prices, and regulatory requirements in sustaining a viable domestic industry. It would appear that there would be some churning, however, as the emerging businesses may differ in ownership and location from those that are contracting. This is a natural evolution of a market that arguably does not have much to do with dumping, but its adjustment costs may be mitigated by SG.

Given the magnitude of the AD margins for an NME-wide entity in the context of an approximately perfectly competitive market, I infer that an AD margin of 25.6 is indicative of sales at a loss or cost of dumping. That is, unless the NME entity was being subsidized. In such an instance, a petition under the ASCVM agreement would have been more appropriate, and these margins more defensible. However, a subsidy in an NME may not be countervailable, as there isn't any market to distort.Footnote ²⁹ As Feldman and Burke (Reference Feldman and Burke2013) discuss, the US began to initiate AD and countervailing duty investigations simultaneously against China in 2006, generating legal controversy at the US Court of International Trade and the US Court of Appeals for the Federal Circuit. Ramanujan and Sampath (Reference Ramanujan and Sampath2015) critique panel and AB reports over this pairing of measures. Thus a SG petition arguably was more appropriate than one alleging subsidization.Footnote ³⁰

In discussing the advantages of the new techniques of aquaculture, I note the absence of hormones, antibiotics, and pesticides. This continues with a discussion of food safety and the SPS Agreement. I note, in this regard, that the USITC report for the original investigation refers to the use of antibiotics in shrimp farming. That report also mentions that less than 2% of imports are inspected by the Federal Drug Administration (FDA), while domestic production is subject to ‘multiple US government regulations’. This includes state boards of health and the Hazardous Analysis Critical Control Points (HACCP). The USITC report contains allegations of US processors that the regulations are ‘effectively different’ between US and imported production. In addition to the investigation, there are other reports of concern in the EU, Japan, and the US over the use of antibiotics, disinfectants, and pesticides in shrimp farming in developing countries.Footnote ³¹

This would appear to be a matter for Article XX(b) of the GATT and the SPS agreement. Article 1 establishes the domain of the Agreement as SPS measures that ‘may, directly or indirectly, affect international trade’. Article 2(1) of the SPS refers to a right to protect ‘human, animal, or plant life or health’. Article 2(2) states that such protection should only be ‘to the extent necessary’. The chapeau of the SPS Agreement, Articles 5, 9, and 10, and Annexes A, B, and C refer to ‘appropriate’ levels of protection. Implementing increased inspection of imported shrimp plausibly could protect the domestic industry by introducing or raising compliance costs of foreign producers, thereby leading to higher domestic prices through reduced foreign supply. While AD duties impose a pure surplus loss upon domestic consumers, an SPS measure would provide them a benefit for paying higher prices. This arguably would be a safer product and/or lead to more confidence in the product's safety. The negative implication is that invocation of SPS in such an instance would be an example of what Bhagwati (Reference Bhagwati1988) has termed the ‘law of constant protectionism’. That is, curtailment of one instrument for protecting a domestic industry induces the emergence of an alternative measure.