The Unsettled Governance of the Dual-Use Items under Article XXI(b)(ii) GATT: A New Battleground for WTO Security Exceptions

Abstract

In an era marked by the ‘technologization’ of modern warfare and the privatization of military supply chains, numerous items possess dual-use potential, capable of serving both civilian and military ends. Concurrently, governments increasingly view the acquisition of specific goods, materials, services, and technologies by rival states as a threat to their security. As a result, economic restrictions imposed on dual-use items, including export controls, have proliferated in recent years. These measures have elicited concerns regarding disguised protectionism and potential non-compliance with trade agreements. Central to the debate is the difficulty to strike a balance between addressing legitimate security imperatives and preventing economic protectionism. This article delves into the intersection of trade and security in the regulation of dual-use goods. It offers a focused examination of Article XXI(b)(ii) of the General Agreement on Tariffs and Trade (GATT) concerning trade restrictions on products destined for military use. The paper first reveals limitations of this provision in governing the regulation of dual-use items. Furthermore, it introduces the concept of a ‘purpose test’ provided by the provision as a safeguard against abusive invocation. Lastly, it sheds light on the challenge posed by the standard of proof issue, which complicates the review and mitigation of bad-faith invocations of security exceptions.

1. Introduction

In contemporary discourse, the term ‘dual-use items’ pertains to a broad spectrum of products, services, and technologies capable of serving both civilian and military purposes. Despite historical precedence, trade restrictions on such items have traditionally occupied a peripheral status within international trade law.¹ However, recent geopolitical shifts, economic decoupling attempts,² and intensifying global competition for technological supremacy have propelled the issue of restrictions on dual-use items to the forefront of international trade governance.³

A discernible trend emerges wherein states increasingly equate security with their prowess in economic, technological, and ideological spheres.⁴ Consequently, regulatory frameworks aimed at controlling dual-use items have proliferated, manifesting primarily as export controls, foreign direct investment (FDI) screening mechanisms,⁵ and, most recently, outbound investment controls.⁶ The primary objective of these regulatory mechanisms is to prevent rival states and non-state entities from accessing goods, materials, services, and technologies perceived as posing threats to the regulating nations’ national security and global leadership. Simultaneously, the expansion of these measures raises concerns regarding protectionism and deviation from binding free trade commitments. This concern is underscored by the submission of several disputes to the World Trade Organization (WTO), challenging the legitimacy of various measures concerning dual-use items.⁷

In a recent trade dispute, China has contested the US' export control measures targeting semiconductor chips, supercomputer items, semiconductor manufacturing items, and other related commodities.⁸ The crux of the disagreement lies in the US' assertion that exporting these items would contribute to China's military modernization,⁹ a claim refuted by China, which contends that semiconductor items primarily serve civilian purposes and are traded by commercial entities. China argues that the US imposed export controls with the intention of undermining the scientific and technological development of other WTO Members and maintaining its own technological advantage.¹⁰ The focal point of contention, semiconductor products, epitomizes the intricate nature of dual-use items. In this case, the US seeks to justify its actions by invoking security exceptions for restricting trade of products used for military modernization.¹¹ As several recent WTO panels have examined the scope of sub-paragraph (b)(iii) of Article XXI and confined its invocation to the situations of ‘war or other emergency in international relations’¹², it is likely that the panel may have to consider the US measures under sub-paragraph (b)(ii) of Article XXI, even though the US may not invoke it explicitly, following its earlier practice of referring to Article XXI GATT as a whole.¹³

Despite its significance, sub-paragraph (b)(ii) of Article XXI GATT and its equivalents in the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) and the General Agreement on Trade in Services (GATS) currently occupy a marginal position in legal scholarship concerning the regulation of dual-use items.¹⁴ Existing research primarily revolves around two overarching narratives: the regulatory treatment of dual-use items in non-proliferation regimes and challenges in their implementation,¹⁵ as well as the dual-use items as targets of export controls and challenges in their enforcement.¹⁶ The former often assumes the legality of multilateral non-proliferation regimes without scrutinizing their compatibility with WTO rules, while the latter examines the legality of export controls under Article XXI(b)(iii) GATT as security-related measures imposed ‘in time of war or other emergency in international relations’¹⁷ or under Article XXI(c) GATT as non-proliferation efforts mandated by the United Nations Security Council to maintain international peace and security.¹⁸ In WTO dispute settlement mechanism, while the passing panels addressed the legality of export restrictions under Articles XI(2) and XX GATT,¹⁹ none of these cases concerned invocation of security exceptions. This paper argues that Article XXI(b)(ii) GATT, which is often overlooked in the literature, provides the proper framework for assessing peacetime export controls. However, its framework is clearly out of date and in absence of application practice, lacks an interpretation in light of the changing technological and geopolitical realities.

Our analysis sets forth the effects and coverage of Article XXI(b)(ii) GATT in relation to the regulation of dual-use items. Further, it demonstrates that sub-paragraph (b)(ii) provides a potential safety valve to curtail its abusive invocation: the invoking member is required to prove that the regulated items are traded for military use. This article coins this safety mechanism as a ‘purpose test’. Based on a thorough doctrinal analysis, the paper spells out how this test should be satisfied in practice. On this basis, the article further uncovers inherent limitations in scrutinizing invocations under Article XXI(b)(ii) GATT. It argues that the ‘self-judging’ elements of security exceptions, along with the right of non-disclosure of sensitive information under sub-paragraph (a) of Article XXI GATT, could potentially circumscribe the scope of legal review of this provision. Consequently, the relaxed standard of proof may enable opportunistic protectionism and undermine efforts to mitigate trade restrictions on dual-use items.

Our analysis proceeds in the following manner. In Section 2, we delineate the expanding domain of dual-use items, exploring the ongoing WTO disputes pertinent to restrictions on such commodities, and contextualizing these disputes within the broader geopolitical landscape. This section aims to underscore the heightened significance of trade and security considerations surrounding dual-use products in contemporary discourse. Section 3 delves into a comprehensive interpretation of Article XXI(b)(ii) GATT. Section 4 dissects the critical issue of the standard of proof, probing the challenges and implications associated with establishing the legitimacy of restrictions under sub-paragraph (b)(ii). Building upon this analysis, Section 5 scrutinizes the potential consequences of invoking Article XXI(b)(ii) GATT in cases pertaining to trade restrictions on dual-use items. It specifically underscores the substantial risk of abuse, whereby security exceptions may be wielded to curtail the trade of non-military goods, thereby compromising trade commitments. Finally, in Section 6, we draw together our findings and insights to offer a conclusive reflection on the implications of invoking Article XXI(b)(ii) GATT in the regulation of dual-use items.

2. Setting the Scene: The Blurring Line between Military and Civilian Applications in the Times of Global Insecurity

The concept of dual-use products is inherently a social construct, reflecting the ontological reality that certain goods possess utility for both civilian and military purposes. However, the connotation of this concept remains fluid, contingent upon the epistemological context and instrumental interpretation. Recent years have witnessed significant shifts in both dimensions, resulting in a notable expansion of the dual-use products. First, the phenomenon of modern warfare's ‘technologization’ and the commercialization of military applications have profoundly widened the spectrum of dual-use products. Concurrently, the evolving notion of security has led to the assimilation of items previously outside the purview of conventional security concerns. This section aims to present the expanding category of dual-use items and the emerging issues with regard to the regulation of such items. This section is divided into two parts. Subsection 2.1 introduces recent WTO disputes concerning restrictions on dual-use items and explores the underlying geopolitical dynamics shaping these disputes. Subsection 2.2 delves into the expanding array of dual-use products, driven by a compound of elements including the ‘technologization’ of modern warfare, the increasingly civilian origin of military applications, the eclipsed relevance of multilateral non-proliferation regimes, and the emergence of human rights considerations in export controls. We examine how these underlying currents have propelled trade restrictions on dual-use products from the periphery of trade law to the forefront of regulatory discourse.

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 measures²⁰ restricting the sales of advanced computing semiconductor chips, supercomputer items, semiconductor manufacturing items and related services, and technologies.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.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’.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’.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.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”’.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’.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’.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.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’.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.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.³² Other countries have followed suit. Similar export controls were introduced in Japan, another significant manufacturer of the lithography equipment.³³ In 2023, in spite of the domestic manufacturers’ reservations,³⁴ the Dutch government also imposed export controls on the advanced semiconductor manufacturing equipment.³⁵ Spain expanded its national export controls list by adding quantum computing, additive manufacturing and other emerging technologies.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.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.³⁸ The expansion of the export control lists by the individual Member States was criticized by the Chinese authorities,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’.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.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’.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.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.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.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.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.⁴⁷ In 2017, this approach was further articulated in the 13th Five-Year Special Plan for Science and Technology Military–Civil Fusion Development.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).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.⁵⁰ General Zaluzhnyi of Ukraine, whose book collection includes the Chinese President Xi Jinping's ‘The Governance of China’,⁵¹ asserted that the ‘central driver of this war is the development of unmanned weapons systems’.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.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’.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’.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.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’.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.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.⁵⁹ The EU, taking a step further, currently applies a broad ‘human security approach’⁶⁰ that goes beyond the traditional civilian–military dichotomy.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,⁶² ‘cyber-surveillance items’,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’.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.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.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’.⁶⁷ However, as the traditional security foundation at the basis of the Wassenaar Arrangement has transmogrified,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,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.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. Dual-Use Items within the Meaning of Article XXI(b)(ii) GATT

In principle, the WTO law does not inquire into the nature of the traded goods. This is to say, any measures concerning trade restrictions are subject to WTO law irrespective of the nature or purpose of the product. The WTO agreements do not distinguish between dual-use and single-use products. On this premise, any measures restricting international trade of dual-use items may violate a number of WTO commitments. For example, if the regulation of dual-use products takes the form of export controls, it could potentially contradict the prohibition on quantitative restrictions, non-discrimination standards, including most-favoured-nation treatment and national treatment, and frustrate commitments on trade-related investment measures, trade in services, and intellectual property rights.⁷¹

However, when trade-restrictive measures contravene these primary obligations, the recalcitrant WTO Member could exonerate itself from incurring liability by invoking exceptions contained in the WTO Agreements. In terms of dual-use products, the application of some of the exceptions has already been addressed in legal scholarship.⁷² This section examines the scope of possible restrictions that can be imposed on dual-use items under sub-paragraph (b)(ii) of Article XXI GATT.

Through a focused doctrinal analysis, the section produces the following findings. First, the sub-paragraph (b)(ii) can be applied in peacetime, unlike sub-paragraph (b)(iii) that can be only invoked in times of war or other emergencies in international relations. Therefore, countries might use this provision to limit commerce even without the incidence of war or other international emergencies. Second, the sub-paragraph (b)(ii) focuses solely on trade of military goods or other supplies to the military and thus does not accommodate the emerging concerns such as human rights. Third, the sub-paragraph (b)(ii) consists of two parts, both of which allow restricting dual-use goods, services, and technology transactions under specific scenarios. Raw materials, technology, and services used for manufacturing and maintenance of ‘arms, ammunition and implements of war’ may fall under part one of sub-paragraph (b)(ii). Part two of the sub-paragraph (b)(ii) could potentially cover a broad swipe of dual-use items whose trading is intended for the purpose of directly or indirectly provisioning military establishments. Nevertheless, in order to evoke this latter part, a ‘purpose test’ should be satisfied. The test mandates the invoking countries to show that prohibited items are traded to supply a military establishment. The country invoking part two of the sub-paragraph (b)(ii) should also demonstrate that it had a subjective awareness of the situation at the time when trade-restrictive measures were implemented. It means that, in practice, governments that wish to justify restrictions on dual-use items under part two of the sub-paragraph (b)(ii) must show that they have carried out a due diligence scrutiny or obtained proof from reliable sources which prompted them to believe that the restricted products ‘directly’ or ‘indirectly’ serve military ends. This criterion is crucial to prevent Article XXI(b)(ii) GATT from being misused.

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.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’.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’.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.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.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.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.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.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.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.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’⁸³ so that the panel is satisfied that the invocation is made in good faith and not to circumvent its obligations under WTO agreements.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.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.⁸⁶ In this regard, the panels held, with the notable exceptions of the Russia–Ukraine war⁸⁷ and Qatar blockade,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.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.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’.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’.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’.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.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’.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.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’.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’.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’.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.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,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’.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.’¹⁰³

Further, the connection to the word ‘traffic’ suggests that the intended purpose of such ‘traffic’ or ‘trade’ is to supply a ‘military establishment’.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’.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.’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’.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.’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’.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.’ 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.’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’.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. The Standard of Proof Issue

This section will focus on the challenge of providing sufficient evidence to satisfy the ‘purpose test’ in eventual adjudication. The challenge for effective review of the conditions specified in sub-paragraph (b)(ii) is posed by the WTO Members’ right to decline to provide specific confidential information. Consequently, the adjudicators may determine that the country making the claim has satisfied its burden of proof if it can show that certain products were placed on the control lists because they have the capability to be used by the military, regardless of whether they have actually been used or will be used by the military in practice. Ultimately, this challenge will limit the effectiveness of the ‘purpose test’ in preventing abusive invocation of the provision. This section will review a number of past cases to demonstrate the challenge.

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.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.’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.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’.¹¹⁶ The Czechoslovakia's complaint was ultimately rejected by the GATT Contracting Parties.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 ’.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.’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.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.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.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’.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.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’.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).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.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).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’.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’.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’.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.¹³² However, as noted by scholars,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.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.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.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.

5. The Provision in Action: Implications for Trade Governance

The previous section has examined the scope and limitations of Article XXI(b)(ii) GATT. Particularly, it highlighted the challenge of implementing a rigorous standard of proof in the ‘purpose test’ and the possible need for adjudicators to assess information that is publicly accessible. This section will pin down the likely implications of invoking this provision for global trading and the multilateral trade governance system. To exemplify the implications, the section draws on export control measures. Being the most prominent form of restrictions on dual-use goods, export control represents a pertinent illustration.

Along the lines of the Wassenaar Arrangement, Article XXI(b)(ii) GATT is based on the traditional conception of security. Hence, it exclusively concerns items that are intended for military deployment and appears obsolescent in reflecting the emerging security threats. To be clear, we are not suggesting that the provision should cover all the emerging categories of dual-use items. Controls of the dual-use items serving the objective of human rights protection could be addressed under general exceptions. However, Article XXI(b)(ii) GATT defies its bestowed regulatory objective, which is to permit trade restrictions that may threaten the essential security interests of the invoking states. It will be challenging for the multiple items concerning burgeoning security threats to fit squarely within this provision. Such limitations could be observed at the stage of inclusion of dual-use items in the control lists. In the US, for instance, the Bureau of Industry and Security issues export licenses for the items placed on the Commerce Control List (CCL).¹³⁷ Previously, the US followed the determination of the controlled items under the Wassenaar Arrangement pushing for amendments of the common export control lists and then amending its own export control legislation. Since 2018, however, the US has started to define the dual-use items unilaterally and impose its export controls extraterritorially, especially in relation to ‘emerging technologies’, without seeking international consensus. Recently, the US Congressman Michael T. McCaul has put forward legislation to restrict the export of AI systems.¹³⁸ This raises a number of questions in terms of the designation of such products as dual-use items under sub-paragraph (b)(ii) of the GATT security exceptions: their disruptive nature and uncertainty of military application.¹³⁹ For example, the additive manufacturing, commonly referred to as 3D printing, could potentially contribute to the arms race instability, although the experts continue to debate its likely effects on non-proliferation efforts.¹⁴⁰ As another example, drones that are increasingly becoming part of the modern warfare have a wide range of civilian applications, which makes the effective non-proliferation of the drone parts and drone-related technologies untenable.¹⁴¹ Technologies, such as AI, could also pose risks for cybersecurity, which falls outside Article XXI(b)(ii) GATT unless being used by a military.

Apart from this under-inclusiveness, another pitfall of Article XXI(b)(ii) GATT rests on its weakness in curbing bad faith invocation, pinned on the difficulty to implement the ‘purpose test’. As set forth in this paper, the ‘purpose test’ should be carried out in two stages. The authorities could place certain items on the control list and thus subject their exports to mandatory reporting and licensing, as long as they had prima facie evidence suggesting that the pertinent items could be traded for supplying military establishment. Only at the stage of licensing would the authorities need to verify whether the particular batch of restricted or prohibited items were ‘directly or indirectly’ destined for a military establishment. In practice, such knowledge could be acquired when the exporter applied for a license. However, the export controls would already permit punitive actions on companies as long as they deviated from notification requirements. This means that traders could incur prohibitions, fines, and other sanctions on the ground of primary administrative determination of dual-use items without any further interrogation into their actual transactions. Moreover, because of the relaxed standard of proof, the licensing authority would enjoy a wide margin of discretion in conducting such inquiries. By way of example, on 6 October 2023, the US Department of Commerce included 42 Chinese enterprises on the Entity List subjecting them to certain export restrictions. According to the US authorities, these Chinese entities were engaged in the supply of integrated circuits of US origin to Russian entities associated with the defense sector of Russia.¹⁴² Provided that the Chinese companies involved indeed engaged in supplying the Russian military, the US could legitimately invoke sub-paragraph (b)(ii) of Article XXI GATT to justify these measures. Hence, the crux of the issue here is whether the US indeed possesses information to support its determination of the supply to the military. As mentioned in the preceding section, because of the relaxed standard of proof, the verification of this knowledge will be difficult for the adjudicators.

As a result, the provision is susceptible to abuse. It could be invoked to restrict trade in products that have dual-use features but are not traded for supplying a military establishment. As demonstrated in this paper, the range of items that bear civilian and military functions is vast and expanding. Hence, sub-paragraph (b)(ii) could be invoked to vindicate restrictions on a broad array of raw materials, products, and technologies that could theoretically be used by a military establishment. The tendency of expanding export controls has been conspicuous over the past few years. According to the 2023 WTO report on export controls, Article XXI GATT was referred to as justification for 108 (or 15%) of the notified quantitative restrictions.¹⁴³ Moreover, as sub-paragraph (b)(ii) covers indirect supply, it permits the controlling of exports, re-exports, and transit and hence features high potential for extraterritorial application.

Therefore, it is likely that the WTO dispute settlement mechanism may encounter invocations of Article XXI(b)(ii) GATT as a justification for export controls in relation to dual-use items. The difficulties with assessment of security-related evidence in WTO DSM proceedings prompted the scholars to propose alternative venues for deliberation of security-related specific trade concerns within the existing WTO committees,¹⁴⁴ trade policy review mechanism, and notification of the trade-affecting security measures.¹⁴⁵ Some argued for ‘soft law’ best practices that would guide the states in their actions in relation to communication and implementation of the security-motivated trade restrictions, including transparency, proportionality, minimization of harm to other WTO Members, and limitations on duration of the security measures.¹⁴⁶ While the majority of these proposals turn to the forum for deliberations and procedural matters, it is also important to highlight and acknowledge the challenges of applying current security exceptions provision. Beyond the WTO, the pitfalls of sub-paragraph (b)(ii) present a common challenge for numerous free trade agreements that contain similar provisions.¹⁴⁷

6. Conclusion

The rapid ‘technologization’ of modern warfare and the privatization of military supply chains have significantly blurred the distinction between civilian and military products. Consequently, a wide array of tradable items and services inherently possess dual-use capabilities. Nations, recognizing the strategic significance of technological leadership, may leverage trade restrictions on dual-use items to impede the progress of their rivals, even in scenarios where these items are not explicitly earmarked for military establishment. Notably, major international players, such as China, the European Union, and the US, increasingly frame their security interests within broader global political and economic leadership aspirations. This geopolitical competition has catalyzed intensified limitations on trade and investments, aimed at preserving competitive advantages.

In light of these developments, trade restrictions on items with both civilian and military applications have assumed increased importance within the framework of multilateral trade governance. This article examined the authority of WTO Members to impose restrictions on trade in dual-use items, particularly within the ambit of the security exceptions delineated in sub-paragraph (b)(ii) of Article XXI GATT.

Our analysis, grounded in principles of treaty interpretation, concludes that sub-paragraph (b)(ii) of Article XXI GATT permits restrictions on trade of dual-use products under specific conditions. First, sub-paragraph (b)(ii) can be invoked both during the times of peace and armed conflict. Second, both parts of sub-paragraph (b)(ii) can be applied to restrict military-related transactions. Part one encompasses items categorized as ‘arms, ammunition, and implements of war’, including certain raw materials, technologies, and services contributing to their manufacturing and maintenance. Concurrently, part two encompasses goods intended ‘directly or indirectly’ for military supply. The dual-use items defined outside the civil–military dichotomy, including those suspected of ‘malevolent’ uses, are outside the scope of sub-paragraph (b)(ii) and the respective restrictions cannot be excused through its invocation.

Central to our analysis is the ‘purpose test’ inherent in this provision, which confines the meaning of dual-use items in the context of sub-paragraph (b)(ii). This test requires the invoking nations to substantiate that restricted items are specifically intended for military supply, rather than for general ‘long-term plans’. Moreover, the invoking nation must demonstrate a subjective awareness of this reality at the time of imposing trade-restrictive measures.

However, while the ‘purpose test’ aims to curb potential abuses, its effectiveness is constrained by the relaxed standard of proof typically associated with security-related measures. Consequently, countries retain substantial flexibility in determining how and to what extent the test should be satisfied. When WTO Members withhold evidence citing security concerns, panels will have to rely on publicly available data, providing significant leeway to countries seeking to exploit security exceptions to apply and justify trade restrictions on dual-use products.

The implications of this conclusion for the international trading system are potent. Security exceptions manifest a realist crack on the edifice of the current international trading system built upon the neo-liberal tenets. The incorporation of ‘self-judging’ elements in the respective provisions, as an embodiment of the consensual expectations of the negotiating nations, renders these provisions inherently prone to abuse. Nevertheless, it is also fundamental to acknowledge that over the 70-year lifespan of these provisions, the trading nations have demonstrated laudable self-restraint in their invocation, mostly driven by political rather than legal considerations.¹⁴⁸ The foundation for this political self-constraint, however, is no longer present. Countries are increasingly tapping on the potential of the security exceptions for safeguarding their expanding or pretextual security interests. It is discernable that security exceptions are increasingly used, or abused, for economic and protectionist endeavors, undermining global trading system. In this context, some of these attempts have culminated in disputes that have been submitted to the WTO, but most remain unchallenged. However, WTO panels adjudicating the invocation of security exceptions have made considerable effort to find a balance between the agreed-upon trade commitments and members’ sovereign right to protect their essential security interests. Yet, adjudicating security measures under trade law, as making a Hobson's choice, invariably takes toll on the legitimacy of the WTO regime. In this article, we presented the restrictions on dual-use items as the next challenge for the global trading system, and uncertainty looms large as the dysfunctional WTO dispute settlement mechanism is unable to provide for the final settlement of the emerging trade disputes.