2.1 The US – Semiconductors (China) Case and the Emerging Regulation of Dual-Use Items

On 15 December 2022, China initiated a trade dispute at the WTO against the US in relation to the latter's measuresFootnote ²⁰ restricting the sales of advanced computing semiconductor chips, supercomputer items, semiconductor manufacturing items and related services, and technologies.Footnote ²¹

Semiconductors belong to the ‘three families’ of technologies that were identified by the US administration as crucial for preserving the global technological leadership: (1) computing-related technologies, including microelectronics, quantum information systems, and artificial intelligence; (2) biotechnologies and biomanufacturing; and (3) clean energy technologies. The US National Security Advisor Jake Sullivan emphasized that leadership in each of these technologies is a ‘national security imperative’ and advocated for the so-called ‘small yard, high fence’ strategy to impose stringent controls on the pertinent items.Footnote ²² In the light of this strategy, the preservation of the US technological edge requires strategic use of export controls, foreign investment screening, and outbound investment controls that focus ‘on a narrow slice of technology and a small number of countries intent on challenging us militarily’.Footnote ²³

The Chinese authorities argued at the WTO that the US ‘implements export control on items for civilian use or on activities of commercial entities, with a view to weaken the scientific and technological development of other WTO Members and to preserve its technology edge’.Footnote ²⁴ The US export control lists so far include around 2,800 items exceeding by about 1,000 items the international export controls under the existing non-proliferation frameworks.Footnote ²⁵ In addition, the foreign direct product (FDP) rules of the US provide for extra-territorial enforcement of the export controls since these cover ‘foreign-produced items located outside the US … when they are a “direct product” of specified “technology” or “software,” or are produced by a complete plant or “major component” of a plant that itself is a “direct product” of specified “technology” or “software”’.Footnote ²⁶ Furthermore, the US Persons’ Activities Rules ‘restrict US persons from engaging in or facilitating activities supporting the development or production of certain ICs at fabs in China’.Footnote ²⁷ In 2023, the US moved to expand the coverage of the specified export controls and introduced additional circumvention prevention measures ‘to restrict the PRC's ability to both purchase and manufacture certain high-end chips critical for military advantage’.Footnote ²⁸

In its request for consultations, China alleged that the US export control measures constitute violations of Article I:1 (most favoured nation treatment), X:3 (discriminatory administration of the laws and regulations), XI:1 (quantitative restrictions) GATT, Article 28 TRIPS (restrictions on patent holder rights), and Article VI GATS (discriminatory administration of the laws and regulations). In relation to China's allegations, the US officials contended that semiconductors are game-changing dual-use technologies that serve China's military modernization.Footnote ²⁹ On this basis, the US invoked security exceptions under Article XXI GATT, Article XIVbis GATS, and Article 73 TRIPS and reiterated that ‘issues of national security are political matters not susceptible to review or capable of resolution by WTO dispute settlement’.Footnote ³⁰ At the time of writing, this case remains pending, but a wave of similar economic restrictions is underway.

First of all, the US is likely to maintain its strategy of blending national security and economic nationalism.Footnote ³¹ The US administration continues to frame its regulatory policies under the banner ‘economic security is national security’ based on the premise that the maintenance of an advanced military requires economic growth, resources, and technological innovation.Footnote ³² Other countries have followed suit. Similar export controls were introduced in Japan, another significant manufacturer of the lithography equipment.Footnote ³³ In 2023, in spite of the domestic manufacturers’ reservations,Footnote ³⁴ the Dutch government also imposed export controls on the advanced semiconductor manufacturing equipment.Footnote ³⁵ Spain expanded its national export controls list by adding quantum computing, additive manufacturing and other emerging technologies.Footnote ³⁶ At the end of 2023, Finland jumped on the bandwagon and launched a public consultation on the expansion of its national export controls to include emerging technologies such as integrated circuits, quantum computers, and additive manufacturing equipment.Footnote ³⁷ The European Commission has responded by publishing the national export control lists adopted by Netherlands and Spain so as to allow other EU Member States to impose authorization requirements on the items included therein.Footnote ³⁸ The expansion of the export control lists by the individual Member States was criticized by the Chinese authorities,Footnote ³⁹ and it was also raised within the EU that ‘the decision by the Dutch to implement export controls on DUV machines for reasons of national security could be challenged at the WTO by China’.Footnote ⁴⁰ Nevertheless, the EU officials appeared resolute to further align its export control policies towards China with those of the US. In its White Paper on Export Controls released in early 2024 as a part of the European Economic Security package, the Commission further called for better coordination of the national control lists so as to ensure that the EU would have a common voice on European security and the control of trade in dual-use items internationally.Footnote ⁴¹ As noted by Thierry Breton, the EU Commissioner for Internal Market, ‘We cannot allow China to access the most advanced technologies, be they in semiconductors, quantum, cloud, edge, AI, connectivity, and so on. I see a very strong alignment on this agenda between the EU and the U.S., even if we may sometimes differ on the methods, which is normal’.Footnote ⁴² As a result, the recent transatlantic coordination of the controls over emerging technologies, such as artificial intelligence, quantum technologies, and 6G wireless communication systems, took the central stage at the periodic meetings of the EU–US Trade and Technology Council.Footnote ⁴³

Further proliferation of export control measures in the US, the EU, and elsewhere would likely give rise to new trade disputes and retaliatory measures, which further underscores the urgency of clarifying the trade rules applicable to the dual-use items. The growing perception of economic and technological leadership through security lens leads to the increasing overlap between security policy and ordinary regulation of trade and investment activities, thus creating a formidable challenge for international economic law.Footnote ⁴⁴ However, while the examples mentioned in the present section concern primarily semi-conductor products and related strategic technologies, it must be emphasized that the family of dual-use items continues to grow and so do the implications and complexity of trade restrictions on these products, services, and technologies.

2.2 Expanding Family of Dual-Use Items

In the ordinary usage of the word, the term ‘dual-use’ refers to dual functionality of a product, activity, or technology that could be used for at least two distinct purposes. Traditionally, dual-use products were defined in terms of the civilian–military dichotomy.Footnote ⁴⁵ At one end of this spectrum lie weapons stricto sensu, while at the opposite extreme reside purely civilian items that scarcely lend themselves to military applications in any reasonable scenario. Everything in between could be subsumed under the rubric of dual-use items. In the past, countries used to regulate the trading of dual-use products in a bid to curb the proliferation of weapons that would enhance the military capability of their actual or potential rivals.Footnote ⁴⁶ However, in the current context of geopolitical confrontations, three trends have augmented the scope of dual-use products and hence posed new challenges for the regulation thereof: first, the ‘technologization’ of modern warfare and privatization of military supply chains; second, morals- or value-based control of the dual-use items; and third, deficiencies of the multilateral non-proliferation regimes. In what follows we set forth these three trends with some salient examples.

The first trend is underpinned by the increased ‘technologization’ of the modern warfare and privatization of military supply chains, which resulted in the accelerated expansion of the dual-use items capable of concomitant military and civilian applications. This phenomenon can be illustrated by the example of China's civilian–military fusion strategy. In 2014, President Xi called for the reform of the current approach towards military innovation, which should increasingly rely on the independent innovation generated by the private sector.Footnote ⁴⁷ In 2017, this approach was further articulated in the 13th Five-Year Special Plan for Science and Technology Military–Civil Fusion Development.Footnote ⁴⁸ While the plan advocates for the sharing of military and civilian science and technology infrastructure, such as laboratories, testing facilities, large scientific installations, and scientific equipment centers, the State Council's Opinions on promoting military–civilian integration provide for integration of the civilian research and development and manufacturing capacities into the military supply chains whereby private companies are encouraged to supply directly to the People's Liberation Army (PLA).Footnote ⁴⁹ The specifics of the ongoing Russia–Ukraine war also confirmed the importance of the civilian technologies that could be put to military use. Both Russian and Ukrainian combatants widely use civilian drones for gathering intelligence on the battlefield.Footnote ⁵⁰ General Zaluzhnyi of Ukraine, whose book collection includes the Chinese President Xi Jinping's ‘The Governance of China’,Footnote ⁵¹ asserted that the ‘central driver of this war is the development of unmanned weapons systems’.Footnote ⁵² Against this backdrop, the ‘risk of some high specification and high-performance civilian unmanned aerial vehicles being converted to military use’ was cited by China's Ministry of Commerce when adopting corresponding export control measures.Footnote ⁵³

In addition, commercialization and privatization of the military supply chains have further elevated the role of private contractors in the defense-related procurement. As has been observed, ‘national governments have found that higher quality and lower prices are available “off the shelf” in private markets’.Footnote ⁵⁴ Consequently, military products are often supplied and traded by private actors. These private traders may engage simultaneously in civilian manufacturing and transactions, resulting in the same type of products flowing through the military channels and civilian commerce in tandem. As a recent example, the importance of engaging private actors in the military supply chains was underscored by the shortages of ammunition experienced by the Ukrainian armed forces in the ongoing conflict with Russia. The former commander of Ukraine's armed forces, Valerii Zaluzhnyi acknowledged that ‘imperfection of the regulatory framework governing the military-industrial complex in our country, and partial monopolization of this industry lead to difficulties in the production of domestic ammunition, as result – the deepening of Ukraine's dependence on the supply of allies’.Footnote ⁵⁵ Owing to these trends, the family of products bearing both civilian and military uses is continuously expanding, so do the stakes of its regulation.

The second observable trend is the burgeoning identification of dual-use items according to more ethics- and value-dependent ‘benevolent’ and ‘malevolent’ purposes. Such purposes usually include anti-terrorism, anti-crime enforcement, cybersecurity, and human rights protection.Footnote ⁵⁶ The identification of dual-use items is, in essence, a subjective choice. As elucidated by scholars, ‘a technology's dual usage is influenced not only by its technical characteristics, but also by international norms, by how it is conceived, and by how knowledge is mediated by socio-scientific agents’.Footnote ⁵⁷ Yet, this subjective nature of identifying dual-use items has become more noticeable over the past few years, as countries now scrutinize not only products employed for military purposes, but also for purposes that are deemed malevolent. The US, for example, applies a pro-active and forward-looking approach identifying certain ‘emerging technologies’ that currently may have no military applications, but could potentially create security risks due to their disruptive nature.Footnote ⁵⁸ Under this expanding range of the desirable and undesirable uses, the ‘dual-use’ rubric extended to the generative artificial intelligence such as ChatGPT, which could be used to automate the production of fake news and spread disinformation on the social networks.Footnote ⁵⁹ The EU, taking a step further, currently applies a broad ‘human security approach’Footnote ⁶⁰ that goes beyond the traditional civilian–military dichotomy.Footnote ⁶¹ This approach led to the expansion of the definition of dual-use items to cover uses related to human rights violations. In this regard, the EU Regulation 2021/821 (EU Dual-Use Regulation) covers ‘traditional’ dual-use items,Footnote ⁶² ‘cyber-surveillance items’,Footnote ⁶³ and a catch-all category of items ‘if the exporter has been informed by the competent authority that the items in question are or may be intended, in their entirety or in part, for use in connection with internal repression and/or the commission of serious violations of human rights and international humanitarian law’.Footnote ⁶⁴ The increasingly ‘value-dependent’ basis of identifying dual-use items makes it challenging to provide an objective standard to guide the resulting trade restrictions.

The third factor that stands behind the prevalence of unilateral dual-use export controls is the deficiencies in the multilateral export control regimes. Thus far, export control of dual-use items has been regulated under the multilateral Wassenaar Arrangement on Export Controls for Conventional Arms and Dual-Use Goods and Technologies.Footnote ⁶⁵ The Wassenaar Arrangement was created in the 1990s to restrict the transfer of arms and dual-use technologies to communist countries during the Cold War.Footnote ⁶⁶ It provides two lists of the controlled items: the Munitions List and the Dual-Use Goods and Technologies List. Signatory countries agree to impose controls on the exportation of the items that are thus enumerated. Between the two types of controlled goods, the dual-use items were defined as ‘those which are major or key elements for the indigenous development, production, use or enhancement of military capabilities’.Footnote ⁶⁷ However, as the traditional security foundation at the basis of the Wassenaar Arrangement has transmogrified,Footnote ⁶⁸ the list has been deemed no longer capable of accommodating new security concerns and thus increasingly obsolete. Despite several attempts to expand the list,Footnote ⁶⁹ due to its consensus-based decision-making and reliance on the voluntary compliance by the participating members, its overhaul has come to naught amid geopolitical tensions between Russia and the West.Footnote ⁷⁰ The brewing pessimistic sentiments of the multilateral regime have eventually paved the way for the emergence of unilateral responses. It also denotes the lack of a multilateral platform for discussing emerging security issues, resulting in an inevitable absence of multilateral solutions.

The trends set forth above have led to the expansion of dual-use products and the fragmented unilateral control thereof. Thus, the regulation of dual-use items in the domain of trade law entails an increased degree of complexity. The dynamic interplay between technological innovation, security interests, and international trade demands a nuanced understanding of how trade rules deal with trade restrictions on dual-use products. In this regard, the following questions assume vital importance: First, to which extent does WTO law permit trade control of dual-use items? Second, is WTO law adequate to accommodate the emerging security concerns? Third, is WTO law capable of striking a balance between protecting legitimate security concerns and precluding protectionism? The next section provides insights to these questions within the confine of Article XXI(b)(ii) GATT.

3.1 Interpretation of ‘Self-Judging’ Elements in Security Exceptions

Article XXI GATT constitutes a legitimate derogation from WTO commitments. It reads:

Noting in this Agreement shall be construed

1. (a) to require any contracting party to furnish any information the disclosure of which it considers contrary to its essential security interests; or

2. (b) to prevent any contracting party from taking any action which it considers necessary for the protection of its essential security interests

   1. (i) relating to fissionable materials or the materials from which they are derived;

   2. (ii) relating to the traffic in arms, ammunition and implements of war and to such traffic in other goods and materials as is carried directly or indirectly for the purpose of supplying a military establishment;

   3. (iii) taken in time of war or other emergency in international relations.

In what follows, we shall interpret this provision in a bid to ascertain the connection between the sub-paragraph (b)(ii) and the regulation of dual-use items.Footnote ⁷³ Importantly, this analysis seeks to identify the conditions under which the sub-paragraph (b)(ii) permits trade restrictions on dual-use products.

As mandated by the Vienna Convention on the Law of Treaties (VCLT), a ‘treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose’.Footnote ⁷⁴ Following these rules on treaty interpretation the analysis first proceeds with ascertaining the ordinary meaning of Article XXI GATT. Further, pursuant to the directives of the VCLT, our analysis shall use travaux preparatoires as supplementary means of interpretation if the primary method of interpretation results in an obscure meaning of the terms or ‘leads to a result which is manifestly absurd or unreasonable’.Footnote ⁷⁵ Thus, our analysis will factor in the drafting history of the respective agreements with a view to ascertaining the meaning of several ambiguous terms included in Article XXI(b)(ii) GATT.

To begin with, Article XXI provides two grounds for WTO Members to legitimize their trade restrictions, both concern essential security interests. Paragraph (a) permits states to withhold information which they consider contrary to the essential security interests, whereas paragraph (b) permits trade-restrictive measures that have been introduced to protect essential security interests under certain circumstances.

Paragraph (b) contains two components. The first is a chapeau which provides that a contracting party is permitted to take any measures ‘which it considers necessary’ for the purpose of protecting essential security interests. The second specifies several circumstances under which the countries are allowed to do so. The enumeration in the latter part is exhaustive, thereby allowing WTO Members to activate this provision only when both the chapeau and one or several of the requisite conditions are satisfied.

In terms of the chapeau of paragraph (b), there has been a long-standing debate regarding whether it is self-judging, arising from the ambiguous term ‘it considers’. One group of countries and some scholars supported the unrestricted discretion of sovereign states to define their security interests and actions necessary to protect such interests.Footnote ⁷⁶ The opponents of this approach were concerned with the unchecked invocation of security exceptions and called upon the WTO panels to develop appropriate legal standards that could be used to review such invocations.Footnote ⁷⁷

A strand of recent cases has clarified this puzzle, wherein the WTO panels were required to entertain the interpretation of the chapeau and the sub-paragraph (b)(iii) of the provision, which permits WTO Members to protect their essential security interests in times of war or other emergencies in international relations.Footnote ⁷⁸ Although none of the respective panel reports has been scrutinized by the Appellate Body due to the latter's dysfunctional status, such reports provide much warranted certainty regarding the utilization of the security exceptions in times of severe disruptions in international relations.Footnote ⁷⁹

Despite some nuances, in these cases, the panels almost consistently ruled that security exceptions include both subjective and objective considerations. First of all, the panels acknowledged that the chapeau of Article XXI(b) GATT contains elements that can be determined subjectively by the invoking state.Footnote ⁸⁰ Overall, when invoking security exceptions, the WTO Members are allowed to determine what constitutes their essential security interests. This potentially encompasses any interests or threats that could be considered ‘essential’ to a state's existence, including the inter-state war, civil war, domestic riots, terrorism, or even human rights violations that transpire in territories other than that of the invoking country.Footnote ⁸¹ Yet, a WTO Member's latitude on this score is not unlimited. In order to substantiate the term ‘essential’, the panel in Russia–Traffic in Transit felt the need to underscore that the ‘re-labelling’ of trade interests as essential security interests represents a misuse of security exceptions.Footnote ⁸² At any rate, whatever the WTO Member considers its essential security interests to be, it must articulate these interests ‘sufficiently enough to demonstrate their veracity’Footnote ⁸³ so that the panel is satisfied that the invocation is made in good faith and not to circumvent its obligations under WTO agreements.Footnote ⁸⁴

At the same time, the panels posited that the elements inscribed in sub-paragraphs (i), (ii), and (iii) can and ought to be objectively verified by the panel.Footnote ⁸⁵ Thus, the panels repeatedly found that sub-paragraph (iii) of Article XXI(b) GATT contains conditions that have to be evidenced by the invoking members to the extent that allows the panel to review fulfillment of these requisite conditions.Footnote ⁸⁶ In this regard, the panels held, with the notable exceptions of the Russia–Ukraine warFootnote ⁸⁷ and Qatar blockade,Footnote ⁸⁸ that the mere existence of geopolitical tensions is not sufficient for the invocation of Article XXI(b)(iii) GATT.

Based on these cases, one could argue that the panels read the phrase ‘it considers’ in the introductory part as having two main controlling functions for interpretation. Firstly, it allows WTO Members to have substantial discretion in deciding what security interests they seek to safeguard and the manner in which they do so. Secondly, it weakens the standard of proof required to determine the specific circumstances mentioned in the sub-paragraphs, even though these circumstances are objective in nature. The subsequent component and its influence on the standard of proof will be examined in greater detail in Section 4.

3.2 Interpretation of Objective Elements in Sub-Paragraph (b)(ii) of Article XXI GATT

Building on the panels’ decisions, we shall proceed to interpret sub-paragraph (b)(ii) of Article XXI GATT, which did not feature in the recent WTO trade disputes. This sub-paragraph pertains to trade restrictive measures ‘relating to the traffic in arms, ammunition and implements of war and to such traffic in other goods and materials as is carried directly or indirectly for the purpose of supplying a military establishment’. At first glance, the provision straightforwardly offers a possibility to restrict the arms trade. Moreover, the positioning of the phrase ‘traffic in other goods and materials as is carried on for the purpose of supplying a military establishment’ in a separate sub-paragraph indicates that such traffic can be restricted even in times of peace as opposed to the security-related measures taken ‘in time of war or other emergency in international relations’ under sub-paragraph (b)(iii).

In terms of its structure, sub-paragraph (b)(ii) contains two distinct parts separated by the conjunction ‘and’. Since the items covered in the two parts are mutually exclusive, some scholars proposed to use the term ‘implements of war’ as a border line dividing the specific items covered therein.Footnote ⁸⁹ We concur to this view. Accordingly, part one concerns ‘traffic in arms, ammunition and implements of war’, while part two relates to ‘such traffic in other goods and materials as is carried on directly or indirectly for the purpose of supplying a military establishment’. In what follows we address these two parts in sequence.

To begin with, the word ‘traffic’ that appears in both parts of sub-paragraph (b)(ii) should be understood as ‘trade’ (the French and Spanish versions use the terms commerce and comercio respectively). The three items concerned are ‘arms’, ‘ammunition’, and ‘implements of war’. These should be understood in the ordinary sense of these words, unless there is evidence that suggests otherwise.Footnote ⁹⁰ One of the early usages of these terms dates back to the definition of ‘contraband of war’ in the Jay's Treaty of 1794 between Britain and the US, which covers ‘arms and implements serving for the purposes of war, by land or sea’.Footnote ⁹¹ The 1919 Convention on the Control of Trade in Arms and Ammunitions defining ‘arms’ and ‘ammunition’ refers to ‘artillery of all kinds, apparatus for the discharge of all kinds of projectiles, explosive or gas-diffusing, flame-throwers, bombs, grenades, machine-guns and rifled small-bore breech-loading weapons of all kinds, as well as the exportation of the ammunition for use with such arms’.Footnote ⁹² The 1925 Geneva Convention for the Supervision of the International Trade in Arms and Ammunition and in Implements of War when referring to ‘arms, ammunition and implements of war’ specified that these are ‘exclusively designed and intended for land, sea or aerial warfare, which are or shall be comprised in the armament of the armed forces of any State, or which, if they have been but are no longer comprised in such armament, are capable of military to the exclusion of any other use’.Footnote ⁹³ The same convention recognizes that some arms and ammunition can be used for both military and other purposes, such as sport or personal defence.Footnote ⁹⁴ The juxtaposition of ‘implements of war’ with the other two items suggest that the products mentioned are similarly intended to serve combat activities during an armed conflict. This definition should effectively exclude those items that are, or potentially could be, used by a military but are not specifically intended for combat, such as footwear, medicines, food, and other daily consumables.

However, it is contestable whether raw materials, technological products, and services that might contribute to the manufacturing and maintenance of weapons could be considered as ‘implements of war’.Footnote ⁹⁵ This obscurity potentially leaves the door open for WTO Members to impose restrictions on such products or services on the ground that they might be used for combat during the war. In this respect, it was suggested that only products that are earmarked for military use should fall under this rubric. For instance, a semiconductor specifically designed for military use would fall under the ‘implements of war’, while a high-capacity semiconductor that is not military-specific should fall under ‘other goods’ (the second part) that would require a verification whether it is intended for supply of a military establishment.Footnote ⁹⁶ Although this interpretation is practical, we fail to read this distinction from the original language of the provision. Yet, in any case, we concur that products designated for military use could be covered by part two of sub-paragraph (b)(ii), as will be set out below.

For one thing, as the negotiation history indicates, part two of sub-paragraph (b)(ii) was incorporated into the US proposal to precisely address military supplies. The US negotiators of the International Trade Organization (ITO) Charter have raised in their internal deliberations that, ‘export restrictions, such as those “designed to conserve domestic supplies of scarce materials necessary in war or designed to prevent supplies from reaching a possible US enemy,” were essential to U.S. security interests and appeared prohibited by the proposed Charter’.Footnote ⁹⁷ It was this concern that prompted a proposal to add the terms ‘and other supplies of the use of the military establishment’ after the reference to ‘arms, ammunition, implements of war’.Footnote ⁹⁸

Eventually, part two of sub-paragraph (b)(ii) GATT prescribes as follows: ‘traffic … carried on directly or indirectly for the purpose of supplying a military establishment’. The term ‘military establishment’ could be approximated to the definition of ‘armed forces’, which ‘consist of all organized armed forces, groups, and units which are under a command responsible to that party for the conduct of its subordinates’.Footnote ⁹⁹ At the same time, there are convincing reasons to include under the definition of ‘military establishment’ not only personnel directly involved in combat but a broader defense and military apparatus. Given that the US was the primary drafter of Article XXI GATT, one should consider the meaning of the term ‘military establishment’ as used in the National Security Act of 1947, the US legislation in force at the time of the ITO negotiations.Footnote ¹⁰⁰ This legislation signed by President Truman merged the Department of War, Department of Navy, and the newly created Air Force into a single organization referred to as National Military Establishment under the leadership of the Secretary of Defense. It also included the Joint Chiefs of Staff, the Munitions Board, and the Research and Development Board. However, already in 1949, to strengthen the civilian control over military,Footnote ¹⁰¹ the National Security Act was amended renaming the National Military Establishment as the Department of Defense. The US foreign policy scholarship equally referred to military establishment as a broader military bureaucracy or defense apparatus speaking of ‘civilians in the military establishment’.Footnote ¹⁰² It is also notable that the ITO negotiators, perhaps besides the US delegates, appeared in doubt as to the precise meaning of this term. The remarks of the Australian delegate are illustrative in this regard: ‘I do not know precisely what a military establishment is, but I doubt whether it would cover a factory which was engaged only or partly in the production of materials of war.’Footnote ¹⁰³

Further, the connection to the word ‘traffic’ suggests that the intended purpose of such ‘traffic’ or ‘trade’ is to supply a ‘military establishment’.Footnote ¹⁰⁴ The ‘purpose test’ is therefore a linchpin to this provision. It distinguishes between ‘intentional supply’ as required by the provision and situations where the dual-use items may eventually reach a military establishment since the items in questions are capable of being used by such an establishment. As a result, in order to bar a product under part two of sub-paragraph (b)(ii), it is insufficient to merely show that the products could be potentially used by a military establishment. Instead, it must be proven that particular ‘traffic’ or ‘trade’ is carried out with a view to supplying a military establishment, directly or indirectly. In addition, the invoking country would need to demonstrate that it had reasons (information) to believe so at the time when the trade restrictions were imposed.

In terms of the meaning of ‘direct’ and ‘indirect’ supply, the text offers no further guidance. However, the internal deliberations of the US drafters indicate that addition of the terms ‘directly or indirectly’ to the initial wording of the security exceptions was prompted by the intention to align the multilateral agreement with the US domestic legislation, which at that time granted the US government ‘control over traffic in arms or other articles used to supply, directly or indirectly, a foreign military establishment, and in times of international crisis to permit control over any article the export of which would affect the security interests of the United States’.Footnote ¹⁰⁵ As a result, the representative of the US Department of War, Harold Hopkins Neff, drafted the following version of today's sub-paragraph (b)(ii): ‘Relating to the traffic in arms, ammunition and implements of war and to such traffic in other goods and materials as is carried on, directly or indirectly, for the purpose of supplying a military establishment.’Footnote ¹⁰⁶ This proposal prompted heated discussions within the US delegation as several officials disputed the added value of the terms ‘directly or indirectly’ and were concerned with the fact that ‘the security exceptions were not limited to controlling “what the US may do”, but instead would permit covered actions by any member state, thereby providing “means for unilateral action [that] w[ould] surely be abused by some countries’.Footnote ¹⁰⁷ As a result, Neff's proposal was voted down and the ‘July 4’ draft circulated among the national delegations in Geneva contained the following wording: ‘Relating to the traffic in arms, ammunition and implements of war and to such traffic in other goods and materials as is carried on for the purpose of supplying a military establishment.’Footnote ¹⁰⁸

Yet, during the negotiations of the predecessor of Article XXI GATT, the Australian government proposed to include a broader clause permitting export control measures ‘[r]elating to the conservation, by export prohibitions, of exhaustible natural resources, if such measures are made effective in conjunction with restrictions on domestic production or consumption or are considered by the Member to be necessary to its long term plans for defence or security’.Footnote ¹⁰⁹ In that regard, the US delegate pointed out: ‘the Australian proposal may be a little too broad, because it is very difficult to say what may be necessary to a Member's long-term plans for security. I think that perhaps you could restrict almost anything in the world on that ground and I wonder whether the exception already in the Charter does not meet their point.’ Footnote ¹¹⁰ Along the same lines, the Canadian representative added that ‘[t]he words “long-term plans” are extremely wide and we feel that they may allow the taking of action which is contrary to the general intent of the Charter under those broad terms. Long-term plans may include almost anything.’Footnote ¹¹¹ As a result of these deliberations, a compromise solution was achieved by adding the expression ‘directly or indirectly’ to the original phrase ‘for the purpose of supplying a military establishment’, which effectively reinstated Neff's proposal that was internally opposed by the majority of the US negotiators. The objection to the Australian amendment, as well as the internal deliberations of the US delegates, signify that the negotiators were concerned about and sought to preclude an unconstrained definition of military supply. It is clear from the negotiating history that GATT/WTO members recognized that the provision could be subject to a broader interpretation, but they were not prepared for the provision to cover ‘almost everything’. However, the issue remains: What precisely does it encompass?

A closer look at the deliberations during the GATT negotiations reveals that the resulting compromise was primarily reflected in two aspects. First, the term ‘military establishment’ should not be regarded as involving long-term plans for defense or security as had been initially put forward by Australia. Second, as a response to Australia's concern, the form of ‘indirect’ supply was included. Its meaning could be inferred from another reply of the US delegate to the Australian proposal. Reacting to the question of the Australian delegate as to whether a member could restrict the exportation of iron ore when it believed that it would be first used by ordinary smelters but ultimately supplied to the military of another country, the US delegate clarified that ‘it was always our interpretation of this clause that if a Member exporting commodities is satisfied that the purpose of the transaction was to supply a military establishment, immediately or ultimately, this language would cover it’.Footnote ¹¹² In accordance with this exchange of opinions, ‘indirect’ supply alludes to situations where products that might initially be used in a civilian setting could have an ultimate purpose of serving the military.

3.3 Regulation of Dual-Use Items within the Interpretation of Sub-Paragraph (b)(ii) of Article XXI GATT

To summarize the discussion in this section, we seek to draw the contours of legitimate restrictions on dual-use items under the auspices of sub-paragraph (b)(ii) of Article XXI GATT.

First, unlike sub-paragraph (b)(iii), sub-paragraph(b)(ii) can be invoked in the times of peace. Therefore, WTO Members may rely on this provision to restrict trade, even in the absence of war or comparable emergency in international relations.

Second, sub-paragraph (b)(ii) can only be used to justify trade restrictions on direct or indirect military supplies. Parts one and two of sub-paragraph (b)(ii) could potentially be applied to restrict transactions of dual-use items (goods, services, or technologies) when certain conditions are met. The military use of an item may place it in the category of ‘arms, ammunition and implements of war’, as stipulated in part one of sub-paragraph (b)(ii). This might include certain raw materials, technologies, and services that contribute to the manufacturing and maintenance of ‘arms, ammunition and implements of war’. Concurrently, dual-use items are expected to be categorized under the second part of sub-paragraph (b)(ii), which pertains to the trade of goods intended ‘directly or indirectly’ for the supply of military. However, the provision does not permit the restrictions to be imposed on the dual-use items identified on the basis of ethics- or value-based categories outside the civil–military dichotomy. Export controls imposed for the purpose of limiting the trade of goods and technologies that may be used for human rights violations are unlikely to meet the requirements of sub-paragraph (b)(ii) unless the aforementioned violations are being perpetrated by the military using the controlled items.

Third, sub-paragraph (b)(ii) offers a safety valve to curb disguised protectionism. In this regard, the ‘purpose test’ in this provision is vital. The invoking nation must provide evidence that the restricted items are being traded for the purpose of supplying a military establishment, rather than for long-term contingency plans. Furthermore, the respective WTO Member is required to demonstrate that the presence of a subjective consciousness of this reality at the moment when the trade-restrictive measures are imposed. This subjective awareness cannot be established retrospectively. The onus to provide evidence should lie with the country making the invocation. Consequently, nations implementing limitations on dual-use items must provide evidence of prior investigation or reliable information indicating that the restricted items are traded to supply the military, either directly or indirectly. The fulfillment of this test is crucial in preventing the misuse of sub-paragraph (b)(ii) of Article XXI GATT.

We posit, in relation to export controls, that the aforementioned ‘purpose test’ should be applied in two stages, which follow the structure of the modern export control systems. The first stage involves the compilation of the control lists. At this stage, the ‘purpose test’ under sub-paragraph (b)(ii) could be satisfied if the invoking WTO Member provides evidence of the possible military application of the listed items, which thus justify their placement on the control lists. The second stage of the ‘purpose test’ assessment concerns the granting of export license. At this stage, the competent authorities, as per the control list, need to determine whether the notified export of the dual-use items would be permitted or banned. At this latter stage, the invoking WTO Member would need to demonstrate its awareness that particular shipments of dual-use items are likely intended for supply of a military establishment, albeit directly or indirectly. However, upon closer examination of the situation, a thorny problem arises: due to the highly confidential nature of military supply, it cannot be assumed that the importing countries and the traders will readily disclose information about military-related transactions.

The determination of the ultimate purpose of products is always a matter of assessment based on the available evidence. Therefore, it is necessary to grant the imposing nation a certain leeway in identifying the intended use and final destination of a commodity. For example, in certain cases, if an exporter has a history of supplying goods to a military establishment, the invoking party may reasonably infer that specific items are likely to be utilized in a military facility. Nevertheless, this presents a difficulty for the reviewing panels in determining the boundaries for the invoking countries’ discretion. This challenge will ultimately translate to the issue of the standard of proof in adjudication. In addition, even if the invoking nation had accurate information, requiring them to reveal it to the panels may place their national security interests in peril. In the next section, these evidentiary difficulties will be fleshed out and examined in detail.

4.1 Czechoslovakia v US: Relaxed Standard of Proof in the ‘Purpose Test’

In 1949, Czechoslovakia (CSR) launched a complaint against the US in relation to the latter's export control measures.Footnote ¹¹³ While describing the wide coverage of the US export control licensing, the CSR delegate remarked: ‘Practically everything may be a possible element of war potential, but if we accept this meaning, it would mean rooting out important sections of vital peace-time industry, narrowing the field of important research and changing the face of modern civilization and make peaceful cooperation impossible.’Footnote ¹¹⁴ The US argued that its export controls are highly specific and only include items that have potential military applications while every license application is examined on a case-by-case basis considering factors such as type of product, the stated end use, and the named consignee.Footnote ¹¹⁵ For example, when examining the export applications for the ball bearings that were declared as destined for the manufacture of agricultural machinery, ‘experts who examined the specifications, however, were convinced that the size, type and degree of precision specified showed them to be destined for use in aircraft, or other military applications’.Footnote ¹¹⁶ The Czechoslovakia's complaint was ultimately rejected by the GATT Contracting Parties.Footnote ¹¹⁷ Again, this case was a testament to the importance of the ‘purpose test’, which mandates the invoking countries to prove a subjective knowledge of the ultimate use of the product concerned when the restrictions are imposed. However, as evidenced by the case, the determination of a country's compliance with the ‘purpose test’ could be onerous.

In the respective case, the US suggested that if a GATT Contracting Party is satisfied with the purpose of transaction, then it can act upon it, thus suggesting that the determination is to be made entirely by the invoking Party. The US voiced a similar view during the works of the Geneva session of the Preparatory Committee, suggesting that the original formulation, ‘such traffic in other goods and materials as is carried on for the purpose of supplying a military establishment’, permits ‘a country to restrict exports not only of exhaustible natural resources but of other things, such as scrap iron, to the extent that it felt that the export transactions were carried on for the purpose of supplying a military establishment abroad ’.Footnote ¹¹⁸ In line with its stance on the self-judging nature of security exceptions, the US would argue that the ‘purpose test’ in Article XXI GATT should also be subject to self-assessment, meaning that only the country imposing trade restrictions can determine if it has been met and adjudicators should not second guess member states’ determinations. This perception is at variance with the recent panel decisions, which have confirmed the objective nature of the elements in the sub-paragraphs of Article XXI GATT.

In accordance with the panels’ recent decisions related to sub-paragraph (b)(iii), the conditions set out in sub-paragraph (b)(ii) also must be proven objectively. However, as has been set forth in subsection 3.1 above, the ‘subjective’ element in the chapeau may be read as ‘softening’ the objective conditions in the subsequent paragraphs. In line with this approach, the fulfillment of the ‘purpose test’ will most likely be subject to a formalistic check. As explained, this could be satisfied if the invoking country demonstrates the fulfillment of due diligence in scrutinizing the purpose or end-user of the products, or that it may possess certain information indicating the ultimate use of the products.

In the latter scenario, however, the challenge is that the invoking members cannot be mandated to disclose confidential information which they consider relevant to their essential security interests. In parallel with the ‘self-judging’ debate, this issue has also been brought out in the dispute between Czechoslovakia and the US. In its reply to the CSR's complaint, the US delegate pointed out that the contracting party shall not be required to provide information that it considers contrary to its essential security interests and that ‘The United States does consider it contrary to its security interest – and to the security interests of other friendly countries – to reveal the names of the commodities that it considers to be most strategic.’Footnote ¹¹⁹ Such right is explicitly enshrined in Article XXI(a) GATT. The 1982 Decision Concerning Article XXI reiterated the right of the Members to refuse disclosure of information contrary to their security interests.Footnote ¹²⁰ Moreover, even absent an explicit permission in the WTO agreements, it could be readily argued that general principles of international law entitle states to refuse to provide evidence in international adjudication on the grounds that this would harm their security interests.Footnote ¹²¹

4.2 The Standard of Proof and Non-Disclosure of Security-Related Information

In a number of recent security-related disputes, the issues concerning the requisite standard of proof have also been tested. In Russia–Traffic in Transit case, Ukraine argued that Russia did not adequately identify the ‘emergency’ during which the contested measures were adopted.Footnote ¹²² Instead of describing the ‘emergency’ cited in the argument, Russia referred to a hypothetical scenario, and when being asked whether this hypothetical scenario corresponded to the actual situation on the ground, the Russian representative explained that hypothetical scenario was introduced ‘in order not to introduce again some information that Russia cannot disclose’.Footnote ¹²³ Hence, for political reasons, Russia did not further substantiate its reference to ‘the emergency in international relations that occurred in 2014 that presented threats to the Russian Federation's essential security interests’ and it ‘asserted that the circumstances that led to the imposition of the measures at issue were publicly available and known to Ukraine’ as they were described in Ukraine's 2016 Trade Policy Review Report.Footnote ¹²⁴ Further, the Russian delegation submitted inter alia that requiring it ‘to provide any information additional to that it has already disclosed in respect of the measures challenged in this dispute would be inconsistent with the provisions of Article XXI(a) of the GATT’.Footnote ¹²⁵ Ukraine contended that Russia cannot invoke Article XXI(a) of the GATT 1994 to evade its burden of proof under Article XXI(b)(iii).Footnote ¹²⁶ Other countries participating in the dispute as third parties also expressed their positions on this matter. For instance, Canada explained the relationship between the two paragraphs in the following manner: ‘requirement to substantiate under paragraph (b) and a requirement not to provide sensitive information under paragraph (a) demonstrates that the threshold for substantiation under paragraph (b) is low but not non-existent’ and suggested that the invoking Member should at least present information that is in the public domain.Footnote ¹²⁷ Along the same lines, the EU submitted that the invoking Member should at a minimum be required to explain why certain information cannot be shared with the panel under Article XXI(a).Footnote ¹²⁸ According to Japan, even if the disclosure of certain information may contradict to the Member's security interests, ‘the invoking Member could discharge its burden of proof to make its Article XXI(b) claim without compromising such information’.Footnote ¹²⁹ As evidenced by the above reactions, countries indeed consider that the right to refuse disclosure confidential information must take precedence, whereas the standard of proof in ascertaining Article XXI(b) should be relaxed, in one way or another.

Eventually, the panel in Russia–Traffic in Transit qualified that, ‘the less characteristic is the “emergency in international relations” invoked by the Member, i.e. the further it is removed from armed conflict, or a situation of breakdown of law and public order (whether in the invoking Member or in its immediate surroundings), the less obvious are the defence or military interests, or maintenance of law and public order interests, that can be generally expected to arise’.Footnote ¹³⁰ Since, in the case at hand, the Russia–Ukraine conflict was recognized as armed conflict by the United Nations General Assembly and a number of countries have imposed sanctions against Russia in connection with this situation, the panel found ‘Russia's articulation of its essential security interests is minimally satisfactory in these circumstances’.Footnote ¹³¹ This flexible approach to assessing security claims preserved the room for articulation of novel or unusual security threats, which would require a higher level of articulation.Footnote ¹³² However, as noted by scholars,Footnote ¹³³ in Russia–Traffic in Transit the panel managed to avoid the discussion on the standard of proof incumbent of the invoking Members, which left it to future panels to determine whether failure to produce evidence can be interpreted against the respective party.Footnote ¹³⁴

Similarly, the subject of the standard of proof under Article XXI GATT was also raised in Saudi Arabia–IP rights case, where the panel used the analytical framework for application of Article XXI(b)(iii) GATT in its assessment of whether Saudi Arabia had properly invoked Article 73(b)(iii) TRIPS. When ascertaining the existence of ‘emergency in international relations’ the panel considered several sources of information, including Saudi Arabia's decision to sever diplomatic and economic relations with Qatar, similar decisions taken by several other countries, and Saudi Arabia's accusations of Qatar's support for terrorism and extremism.Footnote ¹³⁵ All of this publicly available information allowed the panel to conclude that an ‘emergency in international relations’ exists in this case. The foregoing examples demonstrate that the panels, in line with the shared understanding of the WTO Members, attempted to conduct their assessments relying primarily on publicly available information insofar as it allowed ascertaining the existence of circumstances listed in the sub-paragraphs of Article XXI(b) GATT.

Turning back to sub-paragraph (b)(ii), if the same standard of proof is applied, WTO Members could invoke Article XXI(a) GATT to repudiate the disclosure of information. Without such information, it will be impossible for the panel to establish the two critical matters affirmatively when it comes to restricting dual-use items: first, whether a particular batch of dual-use items was destined for military application; second, whether the invoking Member was aware of this at the time when the restrictive measures were adopted. Moreover, it means that, in practice, the adjudicators may consider that the invoking country discharges its burden of proof insofar as it can demonstrate that particular products possess the potential for military use, rather than they have been, or will be used for military application de facto. The expert knowledge, that may be barred by the invoking state, could include information as to the dual-use nature of certain items, the determination that is normally reserved for military and defense officials.Footnote ¹³⁶

The aforementioned cases suggest that the standard of proof will likely to be relaxed in case of security-related restrictions on dual-use products. While the ‘purpose test’ is supposed to curb the misuse of Article XXI(b)(ii) GATT, the imposing countries retain a wide wiggle room in deciding how and to what extent the test should be satisfied. The question as to whether the invoking states adequately fulfill the ‘purpose test’, and discharge their burden of proof in so doing, can only be decided on a case-by-case basis. Furthermore, when the WTO Members refuse to provide certain information on the grounds of security, the panels will have to consider publicly available information. This provides significant room for maneuver to countries seeking to use, or abuse, security exceptions to justify trade restrictions on certain products.