1.       Executive Summary

Romania’s Law No. 163/2021, the constitutionality of which is currently being challenged in the Constitutional Court of Romania, offers an instructive case study of a broader and increasingly important question: how far may a State restrict access to its 5G and communications infrastructure in the name of national security while remaining within the boundaries of international trade rules? The Law No.163/2021 is particularly suitable for such an analysis because it bans certain technologies, equipment and software by reference primarily to the identity, ownership structure and corporate background of their manufacturer. Drawing on my experience in WTO adjudication and academic work in international economic law, this paper examines whether that form of manufacturer-based prohibition is consistent with Romania’s WTO obligations.

On that basis, this paper identifies several major distinct inconsistencies of Law No. 163/2021 with applicable WTO rules. The Law No.163/2021 treats like products differently depending on whether their manufacturers obtain authorization, thereby failing to provide imported products from unauthorized manufactures with equal competitive opportunities, as compared to like products from authorized manufacturers sourced domestically or from third countries. The Law is therefore inconsistent with the most-favoured-nation and national treatment obligations under the relevant WTO covered agreements, which are fundamental non-discrimination principles in the international trade order.  

Moreover, the blanket exclusion of all relevant products supplied by an unauthorized manufacturer under the Law No.163/2021 creates an unnecessary barrier to trade in violation of the WTO rules because more targeted and less restrictive measures are reasonably available. These could include product-specific security requirements, testing, certification, auditing and continuing monitoring, as well as differentiated treatment of core networks and network components of lower criticality.

The Law No.163/2021 also fails to use relevant international technical standards as the basis for its regulatory approach, as required by the WTO TBT Agreement to which Romania is a party.  Rather than assessing the security characteristics of particular technologies or products against objective technical benchmarks, it relies predominantly on broadly formulated criteria concerning their manufacturers. In addition, Romania did not notify the proposed measure through the applicable WTO transparency procedure or give other WTO Members an opportunity to comment before its adoption, which constitutes a violation of the applicable WTO transparency obligations.

Finally, the authorization system’s undefined criteria, the absence of reasons in the resulting decisions and the lack of transparency surrounding their factual basis also create a serious risk that the measure is not administered in the uniform, impartial and reasonable manner required by WTO law.

National security is unquestionably a legitimate and important objective. It does not, however, exempt a measure from WTO scrutiny merely because it is invoked as its purpose. The general exceptions under the GATT 1994 do not cover the security justification advanced for Law No. 163/2021. As for the GATT 1994 security exceptions, note that these exceptions do not confer unlimited discretion: their invocation remains subject to objective review and good-faith requirements. Law No. 163/2021 does not concern trade in arms or military supplies and was not adopted in time of an emergency in international relations. Also, the invocation of the national security exceptions to justify the Law is inconsistent with Romania’s obligation under international law to interpret and apply these exceptions in good faith.

The paper therefore concludes that the Law’s discriminatory, unnecessarily trade-restrictive, non-transparent and potentially arbitrary features amount to violations of Romania’s obligations under the relevant WTO covered agreements and cannot be justified under either the general or the security exceptions of the GATT 1994.

2.       Introduction

In July 2026, I made some preliminary observations on the WTO consistency of the European Commission proposal on the Cybersecurity Act 2.0 (CSA 2.0), which were posted on the IELP Blog and published as WTI Working Paper 4/2026.[2] While doing research on the draft CSA 2.0, I came across Law No. 163/2021[3], the 5G Security Law of Romania, which is currently being challenged before the Constitutional Court of Romania by a Chinese company.[4] I note that under Romanian constitutional law, international agreements, once ratified, are part of the domestic legal order and national courts directly apply these agreements.[5]  Therefore, one of the issues that could be addressed by the Court is the consistency of this Law with Romania’s obligations under the WTO Agreement,[6] because Romania is a party to this 1994 Agreement and, since 1995, a founding Member of the World Trade Organization.[7] In this article, I will first examine the relevant features of Law No. 163/2021, and subsequently assess the WTO consistency of this Law.

3.       Law No. 163/2021 on the Adoption of Measures Relating to Information and Communications Infrastructures of National Interest and Conditions for the Implementation of 5G Networks

The origins of Law No. 163/2021 must be traced back to a 2019 memorandum of understanding between Romania and the United States, and a 2020 European Union (EU) policy paper.  The Memorandum of Understanding signed in Washington, D.C., on 20 August 2019 by Romanian President Klaus Iohannis and US President Donald Trump laid down criteria for evaluating 5G vendors to ensure the security of 5G networks. [8] The criteria set out in the Memorandum were: first, whether the vendor is subject, without independent judicial review, to control by a foreign government; second, whether the vendor has a transparent ownership structure; and, third, whether the vendor has a history of ethical corporate behaviour and is subject to a legal regime that enforces transparent corporate practices. In the same vein, the EU policy paper of January 2020 on Cybersecurity of 5G networks: EU Toolbox of Risk Mitigating Measures recommended that: (a) EU Member States assess their national risk profiles of 5G network suppliers in light of the likelihood of interference by non-EU governments and the degree of control over their own supply chains; and (b) restrict or exclude ‘high-risk suppliers’ from participating in critical and sensitive assets, including core network functions, network management and orchestration functions, and access network functions. [9]

Romania’s 5G Security Law, Law No. 163/2021 on the Adoption of Measures Relating to Information and Communications Infrastructures of National Interest and Conditions for the Implementation of 5G Networks, was adopted by the Romanian Parliament on 7 June 2021 and published in the Official Gazette on 11 June 2021.[10] The Law entered into force on 14 June 2021.[11] Pursuant to Article 1, the Law aims ‘to prevent, counter and eliminate risks, threats and vulnerabilities to national security and defence of the country’. To this end, the Law adopts ‘measures concerning the authorization of manufacturers of technologies, equipment and software used in information and communications infrastructures of national interest, as well as in electronic communication networks through which 5G electronic communication services are provided, 5G networks’. The risks, threats and vulnerabilities referred to in Article 1 are defined in Article 2. Generally speaking, these risks, threats and vulnerabilities concern any action, fact, plan, event, situation, condition or deficiencies, which may impair Romania’s national security and defence.[12] The latter are understood to include national character, sovereignty, independence, unity and territorial integrity, functioning of state institutions, organization and communities of people, as well as life and bodily integrity of citizens.[13] The key feature of Law No. 163/2021 is that it allows only technologies, equipment and software of ‘authorized manufacturers’ to be used within: (i) information and communications infrastructures of national interest (hereinafter ‘communications infrastructures’); and (ii) 5G networks.[14] Manufacturers may be authorized by the Prime Minister of Romania when the Romanian Supreme Council for National Defence (‘CSAT’) gives its assent to such authorization.[15] However, if the CSAT identifies any risk, threat or vulnerability to Romania’s national security or defence, it will issue an adverse opinion and the application for authorization will be rejected by the Prime Minister.[16] In assessing whether there exists a risk, threat or vulnerability to Romania’s national security or defence, the Law sets out four criteria, namely: (1) whether the manufacturer is under the control of a foreign government, in the absence of an independent legal system; (2) whether the manufacturer has a transparent shareholding structure; (3) whether the manufacturer has a record of unethical corporate conduct; and (4) whether the manufacturer is subject to a legal system not requiring transparent corporate practices.[17] Note that the decision of the Prime Minister to grant or reject authorization, which is published in the Official Gazette, does not set out any reasons for the decision.[18] The opinion of the CSAT, on which the decision of the Prime Minister is based, is not made public and is not even shared with the manufacturer applying for authorization.[19] In addition to banning the use of technology, equipment or software of unauthorized manufacturers in the relevant infrastructures and networks, the Law requires that technology, equipment or software of these manufacturers which are already used in communications infrastructures and 5G networks are only to be used for 7 years (or in the case of core networks, 5 years).[20] Moreover, unauthorized manufacturers are excluded from any public procurement procedure for the supply of technologies, equipment or software for information and communications infrastructures of national interest.[21] Providers of electronic communications networks and services in Romania, which use (or, if they already use, continue to use longer than allowed) technologies, equipment or software of an unauthorized manufacturer, may be fined up to 1%-5% of their annual turnover.[22] While the Law does not explicitly single out the manufacturers of any particular country as ‘unauthorized’, Romania’s Minister stated in an interview with Free Europe at the time of the adoption of the Law that it is ‘very clear’ that technologies, equipment or software of Chinese manufacturers, and in particular of Huawei, are excluded from use in 5G networks in Romania.[23]

4.       WTO Consistency of Law No. 163/2021

According to public information, the consistency of Law No. 163/2021 with Romania’s Constitution, and in particular Article 11 thereof,[24] has been challenged before the Constitutional Court of Romania.[25] Relevant to this challenge could be claims that Law No. 163/2021 is in breach of Romania’s obligations under the WTO Agreement. The WTO Agreement is a short, 16-article agreement, with in the annex no less than 18 multilateral trade agreements, which are an integral part of the WTO Agreement and contain the specific substantive obligations imposed on all WTO Members, including Romania. Pursuant to Article XVI:4 of the WTO Agreement, a WTO Member must ‘ensure the conformity of its laws, regulations and administrative procedures with its obligations as provided in the Annexed Agreements’.

In examining the WTO consistency of Law No. 163/2021, the most relevant obligations under WTO law are those contained in the GATT 1994 and the more specific TBT Agreement. Of particular interest are Articles I:1 and III:4 of the GATT 1994 and Article 2.1 of the TBT Agreement, which prohibit discrimination; Article 2.2 of the TBT Agreement, which prohibits the creation of unnecessary barriers to trade; Article 2.4 of the TBT Agreement, which imposes an obligation to use international standards as a basis for technical regulations; Article X:3 of the GATT 1994, which requires the uniform, impartial and reasonable administration of trade measures; and Article 2.9 of the TBT Agreement, which imposes specific transparency obligations regarding technical regulations. As regards the inconsistency with any of the aforementioned GATT provisions, the question that could also be addressed is whether such inconsistency can be justified under Articles XX (General Exceptions) or XXI (National Security Exceptions) of the GATT 1994. Note that all the provisions of the GATT 1994 and the TBT Agreement referred to above have been the subject of interpretation and clarification by WTO dispute settlement panels, i.e., first instance, ad hoc arbitral tribunals, and by the WTO Appellate Body, i.e., the standing court of appeals of the WTO. Their interpretation and clarifications of WTO provisions should be given great weight.

As the panel in EC – Asbestos (2001) held that when a complainant challenges the WTO consistency of a measure under both the TBT Agreement, i.e., the WTO agreement specifically dealing with technical regulations, and the GATT 1994, i.e., the general agreement dealing with measures affecting trade in goods, an adjudicator must first examine whether the measure at issue is consistent with the TBT Agreement, since this agreement deals ‘specifically, and in detail’ with technical regulation.[26] However, should the adjudicator find a measure to be consistent with the TBT Agreement, it must still examine whether the measure is also consistent with the GATT 1994. There is no presumption of GATT-consistency when a technical regulation is found to be consistent with the TBT Agreement. In US – Tuna II (Mexico) (2012), the Appellate Body reversed the panel’s decision to exercise judicial economy regarding the complainant’s claim of inconsistency with Article III:4 of the GATT 1994 after it had found that the measure at issue was not inconsistent with Article 2.1 of the TBT Agreement.[27] Below, I therefore first address the consistency of Law No. 163/2021 with the TBT Agreement, before examining its consistency with the GATT 1994.

4.1.              Is Law No. 163/2021 in violation of the non-discrimination obligations under Article 2.1 of the TBT Agreement?

Article 2.1 of the TBT Agreement provides that:

Members shall ensure that in respect of technical regulations, products imported from the territory of any Member shall be accorded treatment no less favourable than that accorded to like products of national origin and to like products originating in any other country.

Pursuant to Article 2.1, technical regulations are thus subject to a national treatment obligation and a most-favoured-nation (MFN) treatment obligation. The Appellate Body first interpreted the national treatment obligation of Article 2.1 in US – Clove Cigarettes (2012) and further clarified this obligation in US – Tuna II (Mexico) (2012) and US – COOL (2012). In US – Tuna II (Mexico) (2012), the Appellate Body also addressed the MFN treatment obligation of Article 2.1. [28]

In US – Clove Cigarettes (2012), the Appellate Body ruled that Article 2.1 of the TBT Agreement sets out a three-tier test of consistency with the national treatment obligation. This test requires the examination of: (1) whether the measure at issue is a ‘technical regulation’ within the meaning of Annex 1.1; (2) whether the imported and domestic products at issue are ‘like products’; and (3) whether the imported products are accorded ‘treatment no less favourable’ than like domestic products.[29] As the Appellate Body found in US – Tuna II (Mexico) (2012), Article 2.1 sets out a largely similar test of consistency with the MFN treatment obligation. Under the third element of the test, however, instead of examining whether the imported products are accorded ‘treatment no less favourable’ than like domestic products, the adjudicator must examine whether products imported from one WTO Member are accorded ‘treatment no less favourable’ than like products originating in any other country.[30] Below, each element of the three-tier test of consistency with the national treatment and the MFN treatment obligations under Article 2.1 will be discussed in turn.

4.1.1.     Is Law No. 163/2021 a ‘technical regulation’?

In examining the consistency of Law No. 163/2021 with Article 2.1, the first question to be addressed is whether the Law is a ‘technical regulation’ within the meaning of Annex 1.1 of the TBT Agreement.

Annex 1.1 of the TBT Agreement defines a ‘technical regulation’ as a:

Document which lays down product characteristics or their related processes and production methods, including applicable administrative provisions, with which compliance is mandatory. This may also include or deal exclusively with terminology, symbols, packaging, marking or labelling requirements as they apply to a product, process or production method.

In EC – Sardines, the Appellate Body set out three criteria that a document must meet to constitute a ‘technical regulation’: (1) the document must apply to an identifiable product or group of products; (2) the document must lay down one or more characteristics of the product or their related processes and production methods (PPMs); and (3) compliance with the product characteristics (or PPMs) must be mandatory.[31] 

There is no doubt that Law No. 163/2021 is a ‘document’ that applies to ‘an identifiable product or group of products’. The identifiable group of products to which Law No. 163/2021 applies is technologies, equipment and software used in information and communications infrastructures of national interest and 5G networks.

Law No. 163/2021 also lays down one or more characteristics of the products to which the Law applies. It is clear from Annex 1.1, quoted above, that the concept of ‘characteristics’ of a product is a broad concept, which refers to more than the physical characteristics (i.e., form, weight, size, colour, composition, …) of the product, but also includes, inter alia, terminology, symbols, packaging, marking or labelling requirements. The Appellate Body defined the term ‘characteristics’ in EC – Asbestos (2001) as ‘any objectively definable “features”, “qualities”, “attributes”, or other “distinguishing mark” of a product’. [32] Such ‘characteristics’ include not only features and qualities that are intrinsic to the product itself, but also related characteristics.[33] According to the Oxford English Dictionary, a ‘characteristic’ is ‘a typical feature, quality, or distinguishing attribute that someone or something has’. The identity of the manufacturer of a product may be such a distinguishing attribute. For example, the fact that a handbag is manufactured by Louis Vuitton definitely distinguishes this product from handbags produced by Fila.[34] I do not argue that the identity of the manufacturer of a product is always a distinguishing attribute, i.e., a ‘product characteristic’. Whether it is, depends on the extent to which the manufacturer’s identity is relevant to the product and its other characteristics. I note that in EC – Seal Products (2014), the Appellate Body rejected the proposition that the identity of the hunter could be viewed as a product characteristic.[35] However, this finding is inapposite here. The identity of the hunter of the seals in that case was clearly less relevant to the seal products concerned than the identity of the manufacturer of the technologies, equipment and software concerned in the present case. Law No. 163/2021 explicitly distinguishes between the products concerned on the basis of the manufacturer of these products. The identity of the manufacturer is a distinguishing attribute of these products and therefore a ‘product characteristic’ within the meaning of Annex 1.1. Excluding the identity of the manufacturer from the concept of ‘product characteristic’ would open the door to easy and frequent circumvention of the obligations under the TBT Agreement.

In EC – Asbestos (2001), the Appellate Body further noted that product characteristics may be prescribed or imposed in either a positive or a negative form. For example, the regulation may provide, positively, that products must possess certain characteristics, or the regulation may require, negatively, that products must not possess ‘certain characteristics’.[36] Law No. 163/2021 provides that the products concerned may not possess a certain characteristic, namely, they may not be produced by an unauthorized manufacturer.

As noted above, a technical regulation also includes documents laying down the processes and production methods (‘PPMs’) of the products concerned, i.e., the product-related processes and production methods. Law No. 163/2021 distinguishes between authorized and unauthorized manufacturers based on factors such as the ownership of, government control over, and corporate conduct of the manufacturers. It could be argued that each of these factors is a product-related PPM, as they determine, or at least affect, the manner in which, or the conditions under which, the products concerned are manufactured. In EC – Seal Products (2014), the Appellate Body stated that the concept of ‘related PPMs’ ‘is yet to be clarified in WTO law’, but it did indicate that there would need to be ‘a sufficient nexus to the characteristics of a product in order to be considered related to those characteristics’.[37] In the present case, the factors referred to above, i.e., ownership, government control and corporate conduct of the manufacturers, constitute ‘a sufficient nexus’ with the technologies, equipment and software concerned and are thus product-related PPMs.

Law No. 163/2021 also meets the third criterion that a document must meet to constitute a ‘technical regulation’, namely that compliance with it is ‘mandatory’. As the Appellate Body stated in EC – Asbestos (2001), ‘mandatory’ indicates that a technical regulation regulates the subject matter ‘in a binding or compulsory fashion’.[38] The Law requires the use in the communications infrastructures and 5G networks in Romania of technologies, equipment and software of authorized manufacturers, and prohibits, or requires phasing out, the use of technologies, equipment and software of unauthorized manufacturers. The requirements on the use of these technologies, equipment and software are mandatory for the providers and operators of the communications infrastructures and 5G networks concerned. The mandatory compliance with these requirements is enforced through hefty fines.

As it meets the criteria for a ‘technical regulation’ laid down by the Appellate Body in EC – Sardines (2002), it follows that Law No. 163/2021 constitutes a technical regulation within the meaning of the TBT Agreement, although it does not set out typical technical specifications related to communications infrastructures and networks. Note that similar measures, i.e., measures aiming to exclude telecommunications products of ‘high-risk vendors’ from domestic markets, have been notified by WTO Members to other Members and the WTO Secretariat as technical regulations under Article 2.9.2 of the TBT Agreement. For example, on 28 September 2022, Belgium submitted a TBT notification (G/TBT/N/BEL/47) concerning a draft royal decree seeking to implement an authorization system on electronic communications and to establish restrictions on the use of active 5G network elements manufactured by high-risk manufacturers. [39]

4.1.2.     Are the products concerned ‘like products’?

The second element of the three-tier test of consistency with Article 2.1 of the TBT relates to the question of whether the imported and domestic products concerned (for the national treatment obligation) or the imported products originating in different countries (for the MFN treatment obligation) are ‘like products’. The non-discrimination obligations of Article 2.1 only apply to ‘like products’.

In US – Clove Cigarettes (2012), the Appellate Body considered that the determination of ‘likeness’ under Article 2.1 of the TBT Agreement is, as under the non-discrimination obligations of the GATT 1994:

[A] determination about the nature and extent of a competitive relationship between and among the products at issue. [40]

It is well established in WTO case law on the determination of ‘likeness’ (both in the context of the GATT 1994 and the TBT Agreement) that in assessing the nature and the extent of the competitive relationship, an adjudicator must examine on a case-by-case basis all relevant factors, including: (1) the products’ properties, nature, and quality (i.e., their physical characteristics); (2) the products’ end uses (i.e., the extent to which products are capable of performing the same, or similar, functions); (3) consumers’ tastes and habits, also referred to as consumers’ perceptions and behaviour, in respect of the products (i.e., the extent to which consumers are willing to use the products to perform these functions or the extent to which consumers perceive products to be substitutable); and (4) the products’ tariff classification.[41]

Determining whether products are ‘like’ is often a complicated and controversial exercise. Note, however, that in the case law on ‘likeness’ under the non-discrimination obligations under the GATT 1994, it was determined that one may ‘presume’ that the products concerned are ‘like’ when the measure at issue distinguishes between products solely on the basis of their origin.[42] To date, this origin-based presumption of ‘likeness’ has not been made yet in cases relating to Article 2.1 of the TBT Agreement, but there is no reason why such presumption cannot be made regarding a technical regulation which distinguishes between products based on their origin.

As discussed above, the products concerned in Law No. 163/2021 are technologies, equipment and software used in communications infrastructures and 5G networks. The questions which arise are thus:

·       regarding the MFN treatment obligation under Article 2.1 of the TBT Agreement, whether the technologies, equipment and software imported from some WTO Members are ‘like’ the technologies, equipment and software from other WTO Members

·       regarding  the national treatment obligation under Article 2.1 of the TBT Agreement, whether the technologies, equipment and software imported from other WTO Members are ‘like’ the domestic, i.e., Romanian technologies, equipment and software.

Law No. 163/2021 distinguishes between products from authorized and products from unauthorized manufacturers. Pursuant to the Law, authorization is granted to manufacturers mainly based on whether they are controlled by a foreign government and the legal system of the country in which they operate in, as well as their ownership structure and corporate behaviour. Hence, the predominant, if perhaps not the sole, criterion for distinguishing between the products concerned is, directly or indirectly, their origin. Their likeness may thus be presumed. However, if one would consider that the products concerned cannot be ‘presumed’ to be ‘like’, the application of the four-factor ‘likeness’ test, discussed above, swiftly leads to the conclusion that the products are ‘like’. There is no difference at all in the physical characteristics, end uses and tariff classification of the products concerned from authorized manufacturers and those from unauthorized manufacturers. As to the consumers’ tastes, preferences and perceptions, note that the consumers of the products concerned are not private citizens, but providers of electronic communications services. While the tastes, preferences and perceptions of private citizens may be influenced, if not determined, by their positive or negative political perception of the country of origin of the products, the tastes, preferences and perceptions of providers of electronic communications services are typically determined by the performance and quality of the products concerned. Their perception of the products concerned would therefore not differ, at least not on the basis of whether the products are products of authorized or unauthorized manufacturers. On the basis of the above-mentioned factors, it is clear that the products concerned are in a strong competitive relationship and therefore ‘like products’ within the meaning of Article 2.1 of the TBT Agreement.

4.1.3.     Does Law No. 163/2021 accord ‘treatment no less favourable’?

The third element of the three-tier test of consistency with Article 2.1 of the TBT relates to the question of whether Law No. 163/2021 accords ‘treatment no less favourable':

·       to the products concerned imported from some WTO Members than to products from other Members, in which case Law No. 163/2021 would be inconsistent with the MFN treatment obligation; and/or

·       to the products concerned imported from WTO Members than to Romanian products, in which case Law No. 163/2021 would be inconsistent with the national treatment obligation.

In US – Clove Cigarettes (2012), the Appellate Body ruled that, in line with the well-established case law on the concept of ‘treatment no less favourable’ of Article III:4 of the GATT 1994, that:

[A] panel examining a claim of violation under Article 2.1 should seek to ascertain whether the technical regulation at issue modifies the conditions of competition in the market of the regulating Member to the detriment of the group of imported products vis-à-vis the group of like domestic products.[43]

However, while a detrimental impact on the competitive conditions in the relevant market may be sufficient to establish a violation of the national treatment obligation under Article III:4 of the GATT 1994, in US – Clove Cigarettes (2012), the Appellate Body considered that the existence of such detrimental impact is, in case of de facto discrimination, not sufficient to establish a violation of Article 2.1 of the TBT Agreement. The Appellate Body ruled that:

where the technical regulation at issue does not de jure discriminate against imports, the existence of a detrimental impact on competitive opportunities for the group of imported vis-à-vis the group of domestic like products is not dispositive of less favourable treatment under Article 2.1.[44]

According to the Appellate Body, a panel must in such cases of de facto discrimination:

further analyze whether the detrimental impact on imports stems exclusively from a legitimate regulatory distinction rather than reflecting discrimination against the group of imported products.[45]

To determine whether the detrimental impact stems exclusively from a legitimate regulatory distinction rather than reflecting discrimination, a panel must, according to the Appellate Body in US – Clove Cigarettes (2012):

[C]arefully scrutinize the particular circumstances of the case, that is, the design, architecture, revealing structure, operation, and application of the technical regulation at issue, and, in particular, whether that technical regulation is even-handed, in order to determine whether it discriminates against the group of imported products.[46]

The Appellate Body reached this understanding on the basis of the context of Article 2.1 and the object and purpose of the TBT Agreement. The Appellate Body emphasised, in particular, that, while the TBT Agreement does not contain a ‘general exceptions’ clause similar to Article XX of the GATT 1994,[47] the WTO Members recognise in the sixth recital of the Preamble to the TBT Agreement that ‘no country should be prevented from taking measures necessary’ to pursue policy objectives such as the protection of public health, the protection of the environment and the protection of the consumer. As the sixth recital of the Preamble states, countries should not be prevented from taking such measures ‘subject to the requirement that [these measures] are not applied in a manner which would constitute a means of arbitrary or unjustifiable discrimination between countries where the same conditions prevail or a disguised restriction on international trade and are otherwise in accordance with the provisions of this Agreement’. Thus, the Appellate Body interpreted the ‘treatment no less favourable’ requirement of Article 2.1 as:

[P]rohibiting both de jure and de facto discrimination against imported products, while at the same time permitting detrimental impact on competitive opportunities for imports that stems exclusively from legitimate regulatory distinctions.[48]

The Appellate Body further refined and elaborated in later case law this ‘legitimate regulatory distinction’ test to determine whether a technical regulation accords ‘treatment no less favourable’.[49] In this regard, note that the Appellate Body in US – Tuna II (Mexico) (2012) introduced the ‘calibration’ analysis to assess whether the detrimental impact of a technical regulation stems exclusively from a legitimate regulatory distinction.[50]  This analysis examines whether there is a rational relationship between the regulatory distinction(s) and the objective of the measure to see if such distinction(s) is (are) ‘calibrated’ to the objective.

As discussed above, Law No. 163/2021 excludes the use in communications infrastructures and 5G networks in Romania of products of unauthorized manufacturers. The Law thus, quite obviously, modifies the conditions of competition of imported products of unauthorized manufacturers from certain WTO Members vis-à-vis products of authorized domestic manufacturers and authorized foreign manufacturers from other WTO Members. Law No. 163/2021 has a detrimental impact on the conditions of competition of some imported products. However, as Law No. 163/2021 does not explicitly identify the WTO Members of which the products may not be used in communications infrastructures and 5G networks and therefore does not constitute de jure discrimination, the Law can only be found to accord ‘treatment no less favourable’ when the detrimental impact on the conditions of competition of the products of unauthorized manufacturers exclusively stems from a legitimate regulatory distinction.

In assessing whether or not Law No. 163/2021 meets the ‘legitimate regulatory distinction’ test, an adjudicator should, in accordance with the case law discussed above, examine whether the Law is designed and implemented in a manner that constitutes arbitrary or unjustifiable discrimination and fails either the even-handedness requirement or the ‘calibration’ test. In this regard, note that the criteria for granting authorization are vague, subjective, and political, rather than technical, in nature. However, most importantly, the authorization is granted or denied without conducting any evidence-based and product-specific risk assessment and therefore is not ‘calibrated’ to the stated objective of protecting national security and defence.

This leads me to the conclusion that Law No. 163/2021 does not accord ‘treatment no less favourable’ and is inconsistent with both the MFN treatment obligation and the national treatment obligation of Article 2.1 of the TBT Agreement.

4.2.              Is Law No. 163/2021 an unnecessary obstacle to international trade in violation of Article 2.2 of the TBT Agreement?

Article 2.2, first sentence of the TBT Agreement, provides that:

Members shall ensure that technical regulations are not prepared, adopted or applied with a view to or with the effect of creating unnecessary obstacles to international trade.

This sentence is followed by a sentence which qualifies the terms of the first sentence and elaborates on the scope and the meaning of the obligation contained therein. The second sentence of Article 2.2 states:

For this purpose, technical regulations shall not be more trade-restrictive than necessary to fulfil a legitimate objective, taking account of the risks non-fulfilment would create.

The third sentence of Article 2.2 lists several of the ‘legitimate objectives’ to which the second sentence refers: national security; the prevention of deceptive practices; the protection of human health and safety, animal or plant life or health; and the protection of the environment. Note, however, that this list is not exhaustive as it is introduced by the word ‘inter alia’. The last sentence of Article 2.2 refers back to the final clause of the second sentence, namely, ‘taking account of the risks non-fulfilment would create’ and states that in assessing such risks, it is relevant to consider, inter alia: available scientific information; related processing technology; or intended end uses of products.

Article 2.2 sets out a four-tier test of consistency. This test requires the examination of: (1) whether the measure at issue is a ‘technical regulation’ within the meaning of Annex 1.1; (2) whether the measure at issue is ‘trade-restrictive’; (3) whether the measure at issue fulfils a legitimate objective; and (4) whether the measure at issue is ‘not more trade-restrictive than necessary’ to fulfil that legitimate objective. The first element of this four-tier test has already been addressed above. Law No. 163/2021 is a technical regulation within the meaning of Annex 1.1. The second, third and fourth elements of the test will be discussed in turn below.

4.2.1.     Is Law No. 163/2021 ‘trade restrictive’?

The second element of the test of consistency with Article 2.2 relates to the question of whether the measure at issue is ‘trade-restrictive’. In US – Tuna II (Mexico) (2012), the Appellate Body defined ‘trade-restrictive’ to mean ‘having a limiting effect on trade’.[51] The Appellate Body elaborated on the notion of trade restrictiveness in Australia – Plain Packaging (2020) as follows:

Where a measure modifies the conditions of competition in the market, a panel must be satisfied that such modification will have a limiting effect on trade in order to conclude that the measure is trade-restrictive.[52]

Law No. 163/2021 clearly has a limiting effect on trade and is therefore ‘trade-restrictive’ within the meaning of Article 2.2. The Law prohibits technologies, equipment and software from unauthorized manufacturers from being  used in communications infrastructures and 5G networks in Romania. As a result, imports of such products are less than they would have been in the absence of Law No. 163/2021.

4.2.2.     Does Law No. 163/2021 fulfil a legitimate objective?

The third element of the test of consistency with Article 2.2 relates to the question of whether the measure at issue fulfils a legitimate objective. As noted above, the third sentence of Article 2.2 lists specific examples of legitimate objectives, including ‘national security’. With regard to the term ‘fulfil’, the Appellate Body considered in US – Tuna II (Mexico) (2012) that, in the context of Article 2.2, ‘fulfil’ an objective does not mean the complete achievement of something, but is to be understood as referring to ‘the degree of contribution that the technical regulation makes toward the achievement of the legitimate objective’.[53] In US – Tuna II (Mexico) (2012), the Appellate Body also stated that in the context of Article 2.2, an adjudicator must ‘assess the contribution to the legitimate objective actually achieved by the measure at issue, not the contribution that is intended to be achieved’.[54]

Article 1 of Law No. 163/2021 states that the Law aims ‘to prevent, counter and eliminate risks, threats and vulnerabilities to national security and defence of the country’. While an adjudicator is not bound by a country’s characterisation of the objective(s) it pursues through the measure at issue and must make an independent and objective assessment of the objective(s) pursued, it may take a country’s characterization of the objective(s) as a starting point.[55] As for Law No. 163/2021, it is clear that the objective pursued by the Law is ‘national security’ and that ‘national security’ is a legitimate objective within the meaning of Article 2.2. However, does Law No. 163/2021 ‘fulfil’, i.e., make a contribution to the achievement of, that legitimate objective? As noted above, what is important is the contribution ‘actually achieved’, not the contribution intended to be achieved. To assess the actual contribution of Law No. 163/2021 to the achievement of ‘national security’, an adjudicator must, as the Panel in Australia – Tobacco Plain Packaging (2020) did, review a plethora of evidence, including scientific studies, expert accounts and empirical evidence. It is important to note that it is for the complainant to show that Law No. 163/2021 does not contribute to the achievement of ‘national security’. While it might be challenging for a complainant to show that the Law does not make any contribution to the stated objective, I note, as will be addressed below, that the complainant may be successful in raising doubts as to the degree of contribution Law No. 163/2021 makes to the achievement of ‘national security’.

4.2.3.     Is Law No. 163/2021 more trade restrictive than necessary to protect national security?

The fourth and last element of the test of consistency with Article 2.2 relates to the question of whether the measure at issue is ‘more trade-restrictive than necessary’ to fulfil a legitimate objective, taking account of the risks non-fulfilment would create. In US – Tuna II (Mexico) (2012), the Appellate Body found that in the assessment of whether a technical regulation is ‘more trade-restrictive than necessary’ within the meaning of Article 2.2 of the TBT Agreement an adjudicator should start by considering three factors, namely: (1) the degree of contribution made by the measure to the legitimate objective at issue; (2) the trade-restrictiveness of the measure; and (3) the nature of the risks at issue and the gravity of consequences that would arise from non-fulfilment of the objective(s) pursued by the Member through the measure.[56] The Appellate Body has referred to the consideration of these factors as the ‘relational analysis’. According to the Appellate Body, this relational analysis will, however, in most cases not allow to determine whether the trade-restrictiveness of the measure at issue is more than necessary. In most cases, a comparison of the challenged measure and possible alternative measures should be undertaken.[57] If there is an alternative measure that is: (1) less trade-restrictive; (2) makes an equivalent contribution to the relevant legitimate objective, taking account of the risks non-fulfilment would create; and (3) that is reasonably available, the trade restrictiveness of the measure at issue is more than necessary.[58] The Appellate Body referred to the comparison of the measure at issue and possible alternative measures as the ‘comparative analysis’. Note that with regard to the requirement that a proposed alternative measure is ‘reasonably available’, the Appellate Body stated, in US – COOL (Article 21.5 – Canada and Mexico) (2015), that, since the alternative measures is of a hypothetical nature (because it does not yet exist in the Member in question), a complainant cannot be expected to provide ‘complete and exhaustive’ description of the alternative measure it proposes.[59] The Appellate Body stated:

It would appear incongruous to expect a complainant, under Article 2.2 of the TBT Agreement, to provide detailed information on how a proposed alternative would be implemented by the respondent in practice, and precise and comprehensive estimates of the cost that such implementation would entail.[60]

Starting with the relational analysis of the necessity of the trade restrictiveness of Law No. 163/2021, it must be noted, regarding the first factor to be considered in this analysis, that the degree of contribution to the achievement of national security is rather low, if not zero. The Law prohibits the use in communications infrastructures and 5G networks of technologies, equipment and software from unauthorized manufacturers, i.e., manufacturers which: are under the control of a foreign government; do not have a transparent shareholding structure; have a record of unethical corporate conduct; and/or are subject to a legal system not requiring transparent corporate practices. However, in the absence of evidence that the technologies, equipment or software of these unauthorized manufacturers constitute a national security risk, it is difficult to sustain that Law No. 163/2021 makes a significant contribution to the achievement of national security. In this respect, focus placed on the identity and shareholding of a product manufacturer rather than the product itself bears little relevance to the security risks that the specific product may pose. Therefore, the contribution of Law No. 163/2021 is, if any, small and uncertain.

As to the trade restrictiveness of Law No. 163/2021, the second factor to be considered in the relational analysis, namely trade-restrictiveness of the measure, I note that there will be no trade in technologies, equipment and software of unauthorized manufacturers used in communications infrastructures or 5G networks. As Law No. 163/2021 bans these products, the Law is highly trade restrictive.

As to the nature of the risks at issue and the gravity of consequences that would arise from non-fulfilment of the objective of Law No. 163/2021, i.e., the third factor to be considered in the relational analysis, the possible risks to national security related to communications infrastructures and 5G networks are, in theory, significant and consequences arising from the non-fulfilment of the objective of national security are grave. However, in light of the first and second factors considered above, it is not possible to decide on the basis of the relational analysis whether Law No. 163/2021 is more trade restrictive than necessary.

Hence, I proceed to the comparative analysis. Is there a less trade-restrictive measure that is reasonably available and that makes an equivalent contribution to national security? I argue that there is. Rather than imposing a blanket ban on all technologies, equipment and software of unauthorized manufacturers to be used in communications infrastructures and networks, Romania could clearly distinguish between core networks and other networks of less criticality (for example, radio access network (RAN) and internetwork exchanges).[61] Moreover, Romania could demand that these technologies, equipment and software meet product-specific technical requirements which address possible national security risks that may arise with respect to these products. Together with imposing these technical requirements, it could introduce product-specific testing, auditing and monitoring to ensure compliance. For technologies, equipment and software used in 5G networks, for example, Romania can, in defining the relevant technical requirements, be inspired by the work of the International Telecommunication Union (ITU), which sets inter alia performance and design requirements for 5G networks. I discuss the relevant work of the ITU in more detail below in the context of Article 2.4 of the TBT Agreement. In defining alternative technical requirements to address possible national security risks, Romania could also fall back on EU research under the 5G Public Private Partnership (5G-PPP).[62] Under the 5G-PPP, a dedicated working group on security was set up. An alternative measure setting out product-specific technical requirements addressing national security risks is less trade restrictive (as it is not a blanket ban on all relevant products of unauthorized manufacturers, but is more selective), is reasonably available (as the implementation of this alternative measure would not give rise to excessive costs or major technical difficulties), and makes at least an equivalent contribution to national security. In fact, while Law No. 163/2021 may arguably make no or a small contribution to national security, this alternative measure would specifically focus on products posing actual national security risks and thus make a much more substantial contribution to national security.

This leads me to the conclusion that Law No. 163/2021 is more trade-restrictive than necessary to achieve Romania’s policy objective of national security and is therefore inconsistent with Article 2.2 of the TBT Agreement.

4.3.              Is Law No. 163/2021 in violation of the obligation to base technical regulations on international standards under Article 2.4 of the TBT Agreement?

Article 2.4, first sentence of the TBT Agreement provides:

Where technical regulations are required and relevant international standards exist or their completion is imminent, Members shall use them, or the relevant parts of them, as a basis for their technical regulations.

However, Article 2.4, second sentence of the TBT Agreement further states that Members do not have to use international standards as a basis when:

such international standards or relevant parts would be an ineffective or inappropriate means for the fulfilment of the legitimate objectives pursued, for instance because of fundamental climatic or geographical factors or fundamental technological problems.

There are three main questions, which must be answered to determine whether or not a technical regulation is consistent with Article 2.4, namely: (1) whether there exists a relevant international standard; (2) whether the relevant international standard is ‘used as a basis’ for the technical regulation at issue; and (3) whether the relevant international standard is an effective and appropriate means for the fulfilment of the legitimate objectives pursued. Below, each element of this three-tier test of consistency is discussed in turn.

4.3.1.     Does a relevant international standard exist?

The first element of the test of consistency with Article 2.4 relates to the question of whether a ‘relevant international standard’ exists or its completion is imminent. A standard is an ‘international standard’ within the meaning of Article 2.4 if it is approved by an international standardising body[63], and an ‘international standardizing body’ is a body which has ‘recognized activities in standardization’[64] and of which the membership is ‘open to relevant bodies of at least all WTO Members’.[65] An international standard is ‘relevant’ when it concerns the same product as the technical regulation and addresses the same issue(s).[66]

The International Telecommunication Union (ITU), a recognized international standardizing body, approved on 2 September 2022, IMT-2020, the international standard for 5G networks. IMT-2020 explicitly addresses cybersecurity through dedicated international standards and security frameworks. The ITU established detailed recommendations – such as Recommendation ITU-T X.1814 – to define security requirements, threat identification, and guidelines for IMT-2020 (5G) communication systems.[67] Just as Law No. 163/2021, IMT-2020 concerns 5G networks and addresses the issue of cybersecurity. It is therefore a ‘relevant international standard’ within the meaning of Article 2.4. According to public record, before Law No. 163/2021 was adopted on 7 June 2021, the 5G security guidelines had been negotiated since 2019 and had gone through a number of draft versions. These circumstances demonstrate that completion of the standard was ‘imminent’.[68] Note that Romania has an active presence in the ITU,[69] and is therefore familiar with the ITU standardization work. Therefore, IMT-2020 is the ‘relevant international standard’ for Law No. 163/2021, the existence of which was imminent at the time of adoption of the Law.

4.3.2.     Is IMT-2020 used as a basis for Law No. 163/2021?

The second element of the test of consistency with Article 2.4 relates to the question of whether the relevant international standard is ‘used as a basis’ for the technical regulation at issue. The panel in EC – Sardines (2002) stated that the requirement to use, as a ‘basis’ imposes the obligation to ‘employ or apply’ the international standard as ‘the principal constituent or fundamental principle for the purpose of enacting the technical regulation’.[70]

IMT-2020 has clearly not been used as a basis for Law No. 163/2021. The Law takes a very different approach to regulating the use of technologies, equipment and software in 5G networks and, in particular, to addressing security issues such products may raise.

4.3.3.     Is IMT-2020 an effective and appropriate means for the fulfilment of the legitimate objective pursued by Law No. 163/2021?

The third, and last, element of the test of consistency with Article 2.4 relates to the question of whether the relevant international standard is an ineffective or inappropriate means for the fulfilment of the legitimate objectives pursued by the technical regulation at issue. The Appellate Body in EC – Sardines (2002) stated that an international standard is ‘effective’ if it has the capacity to accomplish the objective(s) pursued, and it is ‘appropriate’ if it is suitable for the fulfilment thereof.[71]

As discussed above, the policy objective pursued by Law No. 163/2021 is national security. This is obviously a legitimate objective and an objective which all countries pursue.  IMT-2020, the international standard for 5G networks, addresses security issues. As noted above, Recommendation ITU-T X.1814 defines security requirements, threat identification, and guidelines for IMT-2020 (5G) communication systems. IMT-2020 is effective since it has the capacity to address adequately national security issues, and it is appropriate since it is suitable for the fulfilment of national security. The fact that IMT-2020 is of a recent date and thus reflects all recent developments regarding cybersecurity strongly supports the conclusion that IMT-2020 is both effective and suitable for the fulfilment of national security.

On the basis of the above, I conclude that Law No. 163/2021 is in violation of Article 2.4 of the TBT Agreement.

4.4.              Is Law No. 163/2021 in violation of the transparency obligations under Article 2.9 of the TBT Agreement?

Article 2.9 of the TBT Agreement imposes on WTO Members detailed transparency and notification requirements. These requirements apply to proposed technical regulations whenever a relevant international standard does not exist, or the proposed regulation is not in accordance with the relevant international standard, and the proposed regulation may have a significant impact on the trade of other Members.  Under Article 2.9.2, Members are required to notify other Members, through the WTO Secretariat, of the proposed technical regulation.[72] Such notification must be done at an early appropriate stage of the process, when comments made by other Members can still be taken into account and amendments to the proposed technical regulation made.[73] In the same vein, Article 2.9.4 requires Members to allow reasonable time for other Members to make comments in writing, discuss these comments upon request, and take these written comments and the results of these discussions into account.

As discussed above, Law No. 163/2021 is not in accordance with the relevant international standard and has a significant impact on trade of other Members. The transparency obligations under Article 2.9 therefore apply to Law No. 163/2021. Romania was under an obligation to notify the draft of Law No. 163/2021 to other Members at an early stage of the legislative process that led to the adoption of the Law. It failed to do so in violation of its obligation under Article 2.9.2.[74] Moreover, in violation of its obligation under Article 2.9.4, it also failed to allow reasonable time for other Members to make comments in writing, discuss these comments upon request, and take these written comments and the results of these discussions into account.

4.5.              Is Law No. 163/2021 in violation of the most-favoured treatment obligation under Article I:1 of the GATT 1994?

Article I:1 states in relevant part:

With respect to … all matters referred to in paragraph […] 4 of Article III, any advantage, favour, privilege or immunity granted by any [Member] to any product originating in or destined for any other country shall be accorded immediately and unconditionally to the like product originating in or destined for the territories of all other [Members].

 As stated by the Appellate Body in EC – Seals Products (2014), there are four questions which must be answered to determine whether or not a measure affecting trade in goods is consistent with the MFN treatment obligation of Article I:1, namely: (1) whether the measure at issue is a measure covered by Article I:1; (2) whether that measure grants an ‘advantage’; (3) whether the products concerned are ‘like products’; and (4) whether the advantage at issue is accorded ‘immediately and unconditionally’ to all like products concerned, irrespective of their origin or destination.[75] Below, I will discuss each element of this four-tier test of consistency in turn.

4.5.1.     Is Law No. 163/2021 a measure covered by Article I:1?

The first element of the test of consistency with the MFN treatment obligation of Article I:1 relates to the question of whether the measure is a measure covered by this provision. Article I:1 covers a very broad range of measures, including measures regarding all matters referred to in Article III:4 of the GATT 1994. The latter measures are ‘laws, regulations and requirements affecting [the] internal sale, … and use’ of products.

Law No. 163/2021 is unmistakably a law affecting the sale and use of products and, in particular, technologies, equipment and software. The Law is thus a measure covered by Article I:1.

4.5.2.     Does Law No. 163/2021 grant an advantage?

The second element of the test of consistency with the MFN treatment obligation of Article I:1 relates to the question of whether the measure at issue grants an ‘advantage’. The text of Article I:1 of the GATT 1994 refers to ‘any advantage, favour, privilege or immunity granted by any [Member]’. In light of the use of the word ‘any’, it is not surprising that the term ‘advantage’ has been given a broad meaning in the case law. The panel in EC – Bananas III (1997) considered that a measure granting an ‘advantage’ within the meaning of Article I:1 is a measure that creates ‘more favourable competitive opportunities’ or affects the commercial relationship between products of different origins.[76] In Canada – Autos (2000), the Appellate Body further clarified the meaning of the term ‘advantage’, and thus the scope of the MFN treatment obligation, by ruling:

The words of Article I:1 refer not to some advantages granted ‘with respect to’ the subjects that fall within the defined scope of the Article, but to ‘any advantage’; not to some products, but to ‘any product’; and not to like products from some other Members, but to like products originating in or destined for ‘all other’ Members.[77]

As discussed above, Law No. 163/2021 prohibits the use in communications infrastructures and 5G networks of technologies, equipment and software of unauthorized manufacturers, while allowing the use of the products of authorized manufacturers. The Law therefore grants an advantage to the products of authorized manufacturers.

4.5.3.     Are the products at issue ‘like products’?

The third element of the test of consistency with the MFN treatment obligation of Article I:1 relates to the question of whether the products at issue are ‘like products’. In Section 4.1.2 above, I already addressed the question whether the products at issue in Law No. 163/2021, i.e., technologies, equipment and software of unauthorized manufacturers and technologies, equipment and software of authorized manufacturers, are ‘like products’ within the meaning of Article 2.1 of the TBT Agreement. As I noted in Section 4.1.2, in the case law on ‘likeness’ under the non-discrimination obligations of the GATT 1994, including Article I:1 thereof, it was determined that one may ‘presume’ that the products concerned are ‘like’ when the measure at issue distinguishes between products solely on the basis of their origin.[78] On examining Law No. 163/2021, I established in Section 4.1.2 that Law No. 163/2021 distinguished between products on the basis of their origin. I therefore come to the conclusion that it may be presumed that the products concerned are ‘like’ within the meaning of Article I:1 of the GATT 1994.

In the alternative and pursuant to well-established case law, I argue that the determination of ‘likeness’ under Article I:1 of the GATT 1994 is (as under Article 2.1 of the TBT Agreement) a determination of the nature and the extent of the competitive relationship between and among the products at issue. In assessing this relationship, an adjudicator must examine the same factors (physical characteristics, end use, consumer tastes, preferences and perceptions, and customs classification) as when establishing ‘likeness’ under Article 2.1 of the TBT Agreement. In Section 4.1.2 above, I concluded that the products at issue in Law No. 163/2021 are ‘like products’ within the meaning of Article 2.1 of the TBT Agreement. Since the same test of determining ‘likeness’ applies under Article I:1 of the GATT, I conclude – for the same reasons as spelt out in Section 4.1.2 – that the products at issue in Law No. 163/2021 are also ‘like products’ within the meaning of Article I:1 of the GATT 1994.

4.5.4.     Does Law No. 163/2021 accord immediately and unconditionally the advantage at issue to all like products?

The fourth and final element of the test of consistency with the MFN treatment obligation of Article I:1 relates to the question of whether the advantage granted by the measure at issue is accorded ‘immediately and unconditionally’ to all like products irrespective of their origin or destination. Article I:1 of the GATT 1994 requires that any advantage granted by a WTO Member to imports from, or exports to, any country must be granted ‘immediately and unconditionally’ to imports from, or exports to, all other WTO Members. There is no debate on what ‘immediately’ means. ‘Immediately’ means ‘without delay, at once, instantly’. More problematic has been the meaning of the requirement to accord an advantage ‘unconditionally’. ‘Unconditionally’ does not mean that no conditions can be attached to the granting of an advantage. It means that these conditions may not have a detrimental impact on the competitive opportunities of the like products imported from or exported to any Member.[79] As discussed above, the ‘advantage’ granted by Law No. 163/2021 is the permissible use of technologies, equipment and software in communication infrastructures and 5G networks in Romania. This ‘advantage’ is only granted to the products of authorized manufacturers. It is therefore not granted ‘immediately’ and ‘unconditionally’ to all like products imported from any Member.

On the basis of the above, I conclude that Law No. 163/2021 is in violation of the MFN treatment obligation under Article I:1 of the GATT 1994.

4.6.              Is Law No. 163/2021 in violation of the national treatment obligation under III:4 of the GATT 1994?

Article III:4 of the GATT 1994 states in relevant part:

The products of the territory of any [Member] imported into the territory of any other [Member] shall be accorded treatment no less favourable than that accorded to like products of national origin in respect of all laws, regulations and requirements affecting their internal sale, offering for sale, purchase, transportation, distribution or use.

As stated by the Appellate Body in Korea – Various Measures on Beef (2001), there are three questions which must be answered to determine whether or not a measure affecting trade in goods is consistent with the national treatment obligation of Article III:4, namely: (1) whether the measure at issue is a ‘law, regulation, or requirement affecting their internal sale, offering for sale, purchase, transportation, distribution, or use’ of the products concerned; (2) whether the imported and domestic products at issue are ‘like products’; and (3) whether the imported products are accorded ‘less favourable’ treatment than that accorded to like domestic products.

The Law meets the above three elements. First, as already discussed above in the context of Article I:1 of the GATT 1994, Law No. 163/2021 is a law affecting the international sale and use of technologies, equipment and software and is therefore covered by Article III:4. Second, as the ‘likeness’ of products under Article III:4 is, pursuant to well-established case law, assessed in the same way as likeness under Article I:1 of the GATT 1994 and Article 2.1 of the TBT Agreement, the products at issue in Law No. 163/2021 are also ‘like products’ within the meaning of Article III:4 of the GATT 1994. Third, by excluding the use in communications infrastructures and 5G networks in Romania of products of unauthorized manufacturers, the Law quite obviously modifies the conditions of competition to the detriment of imported products of unauthorized manufacturers.[80] Law No. 163/2021 therefore accords ‘treatment less favourable’ to imported products within the meaning of Article III:4.

On the basis of the above, I conclude that Law No. 163/2021 is in violation of the national treatment obligation under Article III:4 of the GATT 1994.

4.7.              Is Law No. 163/2021 in violation of the obligation under Article X:3(a) of the GATT 1994 to administer trade measures in a uniform, impartial and reasonable manner?

Article X:3(a) of the GATT 1994 states:

Each [Member] shall administer in a uniform, impartial and reasonable manner all its laws, regulations, decisions and rulings of the kind described in paragraph 1 of this Article.

The panel in Thailand – Cigarettes (Philippines) (2011) ruled that to establish a violation of Article X:3(a), a complainant must show that: (1) the measure at issue is a legal instrument of the kind referred to in Article X:1 of the GATT 1994; and (2) this measure is administered in a manner that is non-uniform, partial and/or unreasonable.[81] Below, I will discuss each element of this two-tier test of consistency in turn.

4.7.1.     Is Law No. 163/2021 a legal instrument referred to in Article X:1 of the GATT 1994?

The first element of the test of consistency with the obligation of Article X:3(a) relates to the question of whether the measure at issue is a legal instrument of the kind referred to in Article X:1 of the GATT 1994. Article X:1 refers to, and applies to, ‘laws, regulations, judicial decisions and administrative rulings of general application’ affecting, inter alia, the sale or use’ of products. As already noted above in Sections 4.5.1 and 4.6.1, Law No. 163/2021 is unmistakenly a ‘law’ affecting the sale and use of products.

4.7.2.     Is Law No. 163/2021 administered in a manner that is non-uniform, partial and/or unreasonable?

The second element of the test of consistency with the obligation of Article X:3(a) relates to the question of whether the measure at issue is administered in a manner that is non-uniform, partial and/or unreasonable.  Note that the administration of a measure in any of these three manners constitutes a violation of Article X:3(a).[82] Also, as Article X:3(a) clearly indicates, the requirements of ‘uniformity, impartiality and reasonableness’ do not apply to the laws, regulations, decisions, and rulings themselves, but rather to the administration of those legal instruments.[83] However, as the Appellate Body clarified in EC – Selected Customs Matters (2006), it is possible to challenge under Article X:3(a) the content of a legal instrument to the extent that it regulates the application or implementation of that instrument.[84] As the Appellate Body stated in US – Shrimp (1998), underlying Article X:3 is the due process requirement. Article X:3 sets out ‘certain minimum standards for transparency and procedural fairness’ in the administration of trade measures.[85] The requirements of ‘uniformity’, ‘impartiality’ and ‘reasonableness’ have been clarified in WTO case law as follows.  ‘Uniformity’ requires that a trade measure be applied consistently and predictably and accords uniform treatment to persons similarly situated[86]; ‘impartiality’ requires that a trade measure is applied or implemented in a fair, unbiased and unprejudiced manner[87]; and ‘reasonableness’ requires that the administration of a trade measure is equitable, appropriate for the circumstances and based on rationality. The assessment of ‘reasonableness’ entails ‘a consideration of the factual circumstances specific to each case, for which purpose it is necessary to examine the features of the administrative act at issue, in the light of its objective, cause or the rationale behind it’.[88]  Notably, in China – Raw Materials (2012), the panel found that the lack of definition, guidelines or standards may pose a very real risk of criteria being applied non-uniformly and unreasonably.[89]

The administration of Law No. 163/2021 raises multiple concerns. First, the decision-making process on the granting or denial of authorization lacks transparency. Applicants for authorisation are merely informed of the final decision of the Prime Minister, but not of the reasons for that decision. Second, the standards for assessing whether to grant authorization are characterised by – to use the words of the panel in China – Raw Materials (2012) – a ‘lack of definition, guidelines or standards’. The criteria applied pursuant to Law No. 163/2021 to decide on granting or denying authorization, namely ‘control by a foreign government’, or ‘an independent legal system’, are not defined. Also, and more generally, it is unclear what constitutes a risk, threat or vulnerability in relation to national security and defence within the meaning of Law No. 163/2021. As decisions to deny authorisation do not give any reasons, the lack of clarity of Law No. 163/2021 on these issues will not be remedied through subsequent application of the Law. Law No. 163/2021 is, and will continue to be applied, in a non-uniform, partial and/or unreasonable manner. This constitutes a blatant violation of the principle of due process which underlies Article X:3(a).

On the basis of the above, I conclude that Law No. 163/2021 is in violation of Article X:3(a) of the GATT 1994.

4.8.              Can the otherwise GATT-inconsistent Law No. 163/2021 be justified under the general exceptions of Article XX of the GATT 1994?

Above, I have argued and concluded that Law No. 163/2021 is inconsistent with Articles I:1, III:4 and X:3(a) of the GATT 1994. The question which now arises is whether Romania can justify these GATT inconsistencies under the general exceptions of Article XX of the GATT 1994, and/or under Article XXI thereof. In this Section, I will address whether justification under Article XX is possible.

Article XX of the GATT 1994, entitled ‘General Exceptions’, states:

Subject to the requirement that such measures are not applied in a manner which would constitute a means of arbitrary or unjustifiable discrimination between countries where the same conditions prevail, or a disguised restriction on international trade, nothing in this Agreement shall be construed to prevent the adoption or enforcement by any contracting party of measures:

(a)         necessary to protect public morals;

(b)         necessary to protect human, animal or plant life or health;

(c)         relating to the importations or exportations of gold or silver;

(d)         necessary to secure compliance with laws or regulations which are not inconsistent with the provisions of this Agreement, including those relating to customs enforcement, the enforcement of monopolies operated under paragraph 4 of Article II and Article XVII, the protection of patents, trade marks and copyrights, and the prevention of deceptive practices;

(e)         relating to the products of prison labour;

(f)          imposed for the protection of national treasures of artistic, historic or archaeological value;

(g)         relating to the conservation of exhaustible natural resources if such measures are made effective in conjunction with restrictions on domestic production or consumption;

(h)         undertaken in pursuance of obligations under any intergovernmental commodity agreement …;

(i)           involving restrictions on exports of domestic materials necessary to ensure essential quantities of such materials to a domestic processing industry …;

(j)           essential to the acquisition or distribution of products in general or local short supply …

In its very first case, US – Gasoline (1996), the Appellate Body stated:

In order that the justifying protection of Article XX may be extended to it, the measure at issue must not only come under one or another of the particular exceptions – paragraphs (a) to (j) – listed under Article XX; it must also satisfy the requirements imposed by the opening clauses of Article XX.[90]

Article XX thus sets out a two-tier test for determining whether a measure, which is otherwise inconsistent with GATT obligations, can be justified.  As regards the first element of this test, namely whether the measure at issue comes under one of the specific exceptions exhaustively listed in paragraphs (a)–(j). These exceptions relate to the protection of societal values and interests such as public morals, human, animal, or plant life or health, exhaustible natural resources, and national treasures of artistic, historic, or archaeological value. In EC – Seal Products (2014), the Appellate Body explained that provisional justification under one of the paragraphs of Article XX requires that a challenged measure ‘address the particular interest specified in that paragraph’, and that ‘there be a sufficient nexus between the measure and the interest protected’.[91] The nexus required between the measure and the interest protected differs between paragraphs, but a common nexus requirement is that the measure is ‘necessary’ to protect the interest concerned.[92] When a measure is found to be provisionally justified under one of the paragraphs of Article XX, an adjudicator must, subsequently, assess whether the application of that measure meets the requirements of the opening clauses, commonly referred to as the chapeau, of Article XX. The chapeau requires that a provisionally justified measure is:

not applied in a manner which would constitute a means of arbitrary or unjustifiable discrimination between countries where the same conditions prevail, or a disguised restriction on international trade.

As discussed above, Law No. 163/2021 explicitly states that it aims ‘to prevent, counter and eliminate risks, threats and vulnerabilities to national security and defence of the country’. The policy objective pursued by Law No. 163/2021 is thus, as also already discussed above, national security. The exceptions exhaustively listed in Article XX do not include ‘national security’, and none of the exceptions that are included can, even if interpreted broadly, be understood to cover ‘national security’. This is not surprising since ‘national security’, as a possible exception, is specifically dealt with in Article XXI of the GATT 1994, discussed below.  As noted above, the provisional justification under one of the paragraphs of Article XX requires that the measure at issue addresses the particular interest, i.e., policy objective, specified in that paragraph. Law No. 163/2021 does not address an interest specified in any of the paragraphs of Article XX and can therefore not be justified under Article XX. Any further exploration of the conditions for the invocation of Article XX to justify Law No. 163/2021 is therefore pointless.

4.9.              Can the otherwise GATT-inconsistent Law No. 163/2021 be justified under the national security exceptions of Article XXI(b) of the GATT 1994?

Apart from Article XX of the GATT 1994, WTO Members can, and increasingly do, invoke Article XXI thereof to justify otherwise GATT-inconsistent measures. As with Article XX, Article XXI may be invoked to justify an inconsistency with any obligation under the GATT 1994, but it is not available to justify inconsistencies with obligations under the TBT Agreement. Article XXI, entitled ‘Security Exceptions’, states:

               Nothing in this Agreement shall be construed

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

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

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

(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;

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

                 (c)         to prevent any contracting party from taking any action in pursuance of its obligations under the United Nations Charter for the maintenance of international peace and security.

4.9.1.     Justiciability and legal standard of Article XXI(b) of the GATT 1994

Of possible relevance in the present case is Article XXI(b), and in particular subparagraphs (ii) and (iii) thereof.  A WTO Member may invoke Article XXI(b) (ii) to justify otherwise GATT-inconsistent measures relating to trade in arms or in other materials, directly or indirectly, for military use; and Article XXI(b)(iii) to justify measures taken in time of war or other emergency in international relations. In recent years, the invocation of, most notably, Article XXI(b)(iii) has escalated, and with it disputes relating to such invocation. Some WTO Members, and in particular, the United States, continue to argue that the question whether a Member correctly invokes Article XXI(b) is ‘non-justiciable’, i.e., cannot be decided by a WTO panel.[93] However, it is now well established in WTO case law that a WTO panel can examine, and decide on, whether a Member correctly invokes Article XXI(b) to justify an otherwise GATT-inconsistent measure.[94] The phrase in the chapeau of Article XXI(b), ‘which it considers necessary for the protection of its essential security interests’ (the ‘it’ referring to the Member invoking Article XXI(b)), does not give a Member unlimited discretion. As the panel in Russia – Traffic in Transit (2019) ruled, the words ‘which it considers’ referred to the ‘essential security interests’ and the necessity of the measure at issue to protect these interests. [95] The words ‘which it considers’ do not refer to the objective facts listed in the subparagraphs of Article XXI(b).[96] Whether the otherwise GATT-inconsistent measure is a measure ‘relating to the traffic in arms, …’; or a measure ‘taken in time of war or other emergency in international relations’, can, and must, be objectively established by a panel.

In old GATT and current WTO case law, there is very little guidance on which measures may be considered to be measures ‘relating to traffic in arms, …’ within the meaning of Article XXI(b)(ii).  While it is clear what traffic in ‘arms, ammunition and implements of war’ is, what is traffic in ‘other goods and materials as is carried on directly or indirectly for the purpose of supplying a military establishment’? This is, potentially, traffic in a very broad category of goods and materials, but, as Article XXI(b)(ii) explicitly states, it is traffic to supply the military. Therefore, measures relating to traffic for supplying non-military entities are not measures ‘relating to traffic in arms …’ within the meaning of Article XXI(b)(ii).

In contrast to Article XXI(b)(ii), there is plenty of recent WTO case law on Article XXI(b)(iii) and the question what measures may be measures ‘taken in time of war or other emergency in international relations’. While the concept of ‘taken in time of war’ needs no explanation, the concept of ‘other emergency in international regulation’ has been more problematic. The panel in Russia – Traffic in Transit (2019) found that an ‘emergency in international relations’ refers to:

a situation of armed conflict, or of latent armed conflict, or of heightened tension or crisis, or of general instability engulfing or surrounding a state.[97]

The panel also noted in this regard that ‘political or economic differences between Members are not sufficient, of themselves, to constitute an emergency in international relations’, unless they affect defence and military interests, or the maintenance of law and public order interests.[98] Note that the panel in Saudi Arabia – IPRs found that ‘Saudi Arabia's severance of all diplomatic, consular and economic ties with Qatar … falls into the category of cases in which such action can be characterized in terms of an exceptional and serious crisis in the relations between two or more States’, i.e., an emergency in international relations.^[99] Most recently, the panel in US – Steel and Aluminium (China) held that :

an “emergency in international relations” … must be, if not equally grave or severe, at least comparable in its gravity or severity to a "war" in terms of its impact on international relations.[100]

If a WTO Member invoking Article XXI(b) can successfully argue that an otherwise GATT-inconsistent measure is a measure ‘relating to traffic in arms, …’ or a measure ‘taken in time of war or other emergency in international relations’, that Member can justify this measure provided that ‘it considers’ the measure ‘necessary for the protection of its essential security interests’. It is now well-established caselaw that, while the words ‘it considers’ leave a Member wide discretion, this discretion – as already noted above – is not unlimited. The panel in Russia – Traffic in Transit (2019) noted, that ‘essential security interests’ may be understood to refer to:

those interests relating to the quintessential functions of the state, namely, the protection of its territory and its population from external threats, and the maintenance of law and public order internally.[101]

According to the panel, what specific interests are relevant to the protection of a state from external or internal threats ‘will depend on the particular situation and perceptions of the state in question and can be expected to vary with changing circumstances’.[102] The panel therefore concluded that:

[I]t is left, in general, to every Member to define what it considers to be its essential security interests.[103]

However, this does not mean that a Member may consider any concern to be an ‘essential security interest’. As the panel in Russia – Traffic in Transit (2019) stated:

[T]he discretion of a Member to designate particular concerns as ‘essential security interests’ is limited by its obligation to interpret and apply Article XXI(b)(iii) of the GATT 1994 in good faith.[104]

The panel noted that ‘the obligation of good faith is a general principle of law and a principle of general international law, which underlies all international treaties and binds all parties to such treaties. This obligation of good faith is codified in Article 31(1) of the Vienna Convention on the Law of Treaties regarding the interpretation of treaties, and in Article 26 thereof regarding the application of treaties. As the panel in Russia – Traffic in Transit (2019) observed, a ‘glaring example’ of a violation of the obligation to interpret and apply Article XXI(b) in ‘good faith’ is the relabelling of trade interests as ‘essential security interests’.[105] To avoid such and other violations of the obligation to interpret and apply Article XXI(b) in good faith, the panel in Russia – Traffic in Transit (2019) ruled that a Member must:

articulate the essential security interests said to arise from the emergency in international relations sufficiently enough to demonstrate their veracity.[106]

The obligation to interpret and apply Article XXI(b) in good faith applies not only to the designation of ‘essential security interests’, but also, and most importantly to the ‘connection’ between these interests and the measure at issue.[107] As noted above, the Member invoking Article XXI(b) must ‘consider’ this measure ‘necessary for the protection of its national security interests’. The panel in Russia – Traffic in Transit (2019) ruled that measures can only be regarded as ‘considered necessary’ by the Member invoking Article XXI(b) if these measures:

meet a minimum requirement of plausibility in relation to the proffered essential security interests, i.e. that they are not implausible as measures protective of these interests.[108]

A measure that would, for example, be very remote from, or unrelated to the essential security interests invoked, would not meet this minimum requirement of plausibility. If a Member would with regard to a measure that does not meet the minimum requirement of plausibility argue that it considers this measure necessary, this Member would not interpret and apply Article XXI(b) in good faith, and an adjudicator must reject the invocation of this provision.

4.9.2.     Can Law 163/2021 be justified under Article XXI(b) of the GATT 1994?

As repeatedly noted above, Law No. 163/2021 aims ‘to prevent, counter and eliminate risks, threats and vulnerabilities to national security and defence of the country’. However, Romania cannot successfully invoke Article XXI(b) to justify the otherwise GATT-inconsistent Law No. 163/2021.

Regarding Article XXI(b)(ii), I note that Law No. 163/2021 does not relate to the kind of trade covered by this subparagraph. The Law does not relate to ‘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’. The technologies, equipment and software used in communications infrastructures and 5G networks are not ‘arms’, ‘ammunition’ or ‘implements of war’. Also, the technologies, equipment and software are traded for use in communication infrastructures and 5G networks. They are not ‘other goods and materials’ traded ‘for the purpose of supplying a military establishment’. The possibility that the products concerned may occasionally be supplied to a military establishment does not alter the nature of Law No. 163/2021 as a measure relating to trade in the products concerned for use in communications infrastructures and 5G networks.

Regarding Article XXI(b)(iii), I note that Law No. 163/2021 is not a measure ‘taken in time of war or other emergency in international relations’. Romania is happily not at war with any country in which the products of unauthorized manufacturers originate. Law No. 163/2021 is therefore not a measure taken in time of war. The Law is also not a measure taken in time of emergency in international relations with any of the countries in which the products of unauthorized manufacturers originate. While Romania may have political and economic differences with these countries, such differences are not an emergency in international relations, defined by the panel in Russia – Traffic in Transit (2019) as ‘a situation of armed conflict, or of latent armed conflict, or of heightened tension or crisis, or of general instability engulfing or surrounding a state’[109], or defined by the panel in US – Steel and Aluminium (China) as a situation which ‘must be, if not equally grave or severe, at least comparable in its gravity or severity to a "war" in terms of its impact on international relations’.[110] I also note that Romania has clearly not severed its diplomatic, consular and economic ties with the countries in which the products of unauthorized manufacturers originate.

I conclude that Law No. 163/2021 is not a measure that covered by either subparagraph (ii) or subparagraph (iii) of Article XXI(b), and can therefore not be justified under Article XXI(b). However, if Law No. 163/2021 would be covered by either of the subparagraphs (quod non), Romania could not successfully invoke Article XXI(b) to justify the otherwise GATT-inconsistent Law, because such invocation would constitute a violation of Romania’s obligation under international law to interpret and apply Article XXI(b) in good faith. This is so because Law No. 163/2021 only refers in the most general terms to the protection of ‘national security and defence of the country’ and fails thus to articulate (sufficiently to demonstrate their veracity) the ‘essential security interests’ at issue. Also, and even more importantly, Law No. 163/2021 is not a measure that Romania can – in good faith – consider to be ‘necessary’ because the Law fails to meet the minimum requirement of plausibility. The publicly available material does not establish that the technologies, equipment and software of unauthorized manufacturers constitute a risk to Romania’s ‘essential security interests’. In the absence of any evidence of a risk to Romania’s ‘essential security interests’, Law No. 163/2021 is so remote from, or unrelated to, the essential security interests invoked that Romania cannot – in good faith – consider the Law to be ‘necessary’.

On the basis of all the above, I conclude that the inconsistency of Law No. 163/2021 with Articles I:1, III:4 and X:3(a) of the GATT 1994 cannot be justified under the national security exceptions of Article XXI(b) thereof.

5.       Conclusion

Pursuant to Article 11 of Romania’s Constitution, international agreements concluded and duly ratified by Romania, such as the WTO Agreement, are part of its domestic legal order and can be directly applied by its domestic courts. On the basis of the factual information that is publicly available and the legal analysis undertaken above, I conclude that Law No. 163/2021 is inconsistent with:

·       the MFN treatment and national treatment obligations of Article 2.1 of the TBT Agreement;

·       the prohibition of creating unnecessary barriers to trade of Article 2.2 of the TBT Agreement;

·        the obligation to use international standards as a basis for technical regulations of Article 2.4 of the TBT Agreement;

·       the transparency obligations regarding technical regulations of Article 2.9 of the TBT Agreement;

·       the MFN treatment obligation of Article I:1 of the GATT 1994;

·       the national treatment obligation of Article III:4 of the GATT 1994; and

·       the obligation to administer trade measures in a uniform, impartial and reasonable manner of Article X:3(a) of the GATT 1994.

I also conclude that the inconsistency of Law No. 163/2021 with Articles I:1, III:4 and X:3(a) of the GATT 1994 cannot be justified under the ‘general exceptions’ of Article XX of the GATT 1994, or the ‘security exceptions’ of Article XXI(b) thereof.

Romania’s Law No. 163/2021 offers a case study of the broader questions arising at the intersection of 5G security regulation and WTO obligations. Drawing on my experience as a former Member and Chair of the WTO Appellate Body and my academic work in international economic law, I seek in this article to contribute to a balanced analysis of those questions, recognizing both the importance of national security objectives and the legal disciplines established by international trade agreements.

[1] Former Member and Chair of the WTO Appellate Body; Professor Emeritus of International Economic Law, World Trade Institute, University of Bern; and external legal advisor, King & Wood.

[2] See https://www.wti.org/research/publications/1499/preliminary-thoughts-on-the-european-commission-proposal-for-the-cybersecurity-act-20/.

[3] See https://legislatie.just.ro/public/DetaliiDocument/243213

[4] See e.g.  https://informat.ro/en/current-affairs/huawei-contests-the-exclusion-from-5g-networks-in-romania-at-the-constitutional-court-116111

[5] Constitution of Romania 1991 (rev. 2003), Article 11(2), see https://www.constituteproject.org/constitution/Romania_2003. 

[6] Marrakesh Agreement on the Establishment of the World Trade Organization (‘WTO Agreement’) signed on 15 April 1994 and entered into force on 1 January 1995, at https://www.wto.org/english/docs_e/legal_e/marag_e.htm.

[7] See https://www.wto.org/english/thewto_e/whatis_e/tif_e/org6_e.htm. When Romania joined the European Union in 2007, it remained a full-fledged Member of the WTO with all the rights and obligations which Members have under the WTO Agreement. Note that the European Union as well as its 27 Member States are all WTO Members, albeit that within the WTO the European Commission will act on behalf of the European Union and its Member States.

[8] On file with the author.

[9] NIS Cooperation Group, CG Publication 01/2020, see https://digital-strategy.ec.europa.eu/en/library/cybersecurity-5g-networks-eu-toolbox-risk-mitigating-measures.

[10] Published in the Official Gazette, Part I, No 590 of 11 June 2021.  See: https://legislatie.just.ro/public/DetaliiDocument/243213

[11] Note that Article 9 and Articles 12 to 16 entered into force 30 days after the publication in the Official Gazette.

[12] See Article 2, paras. a, g, and i of Law 163/2021.

[13] Ibid.

[14] Note Article 2, para. d of Law 163/2021, which defines ‘information and communications infrastructures of national interest’ broadly.

[15] See Articles 3, 4 and 5 of Law 163/2021.

[16] See Article 6 of Law 163/2021. Note that a manufacturer can challenge such rejection within 30 days of its publication in the Official Gazette. See Article 8 of Law 163/2021.

[17] Article 4(3) and Article 5(2) of Law 163/2021. These criteria are virtually identical to the criteria set out in the 2019 Memorandum of Understanding between Romania and the United States referred to above.

[18] See: https://legislatie.just.ro/public/DetaliiDocument/279515

[19] Ibid.

[20] Article 10 of Law 163/2021. The same applies to products of manufacturers of which the authorization was withdrawn. See Article 7 of Law 163/2021.

[21] Article 11(2) of Law 163/2021.

[22] Article 14 of Law 163/2021. If the provider is newly established, a fine ranging from RON 10,000,000 to RON 50,000,000 may be imposed. See ibid. 

[23] See: https://romania.europalibera.org/a/interviu-100-de-minute-cu-premierul-alegeri-pensii-fmi-5g-fara-china/30923127.html

[24] See: https://www.profit.ro/povesti-cu-profit/it-c/ultima-ora-gigantul-huawei-reclama-in-romania-la-ccr-legea-in-baza-careia-i-a-fost-interzis-accesul-la-5g-ul-romanesc-precedent-recent-in-justitia-uniunii-europene-22410424

[25] See: https://dosare.ccr.ro/#/CautareDosare

[26] Panel Report, EC – Asbestos (2001), para. 8.16.

[27] Appellate Body Report, US – Tuna II (Mexico) (2012), para. 405.

[28] Most recently, the panels in EU and Certain Member States – Palm Oil (Malaysia) (2024) and EU – Palm Oil (Indonesia) (2025) interpreted and applied the non-discrimination obligations of Article 2.1.

[29] Appellate Body Report, US – Clove Cigarettes (2012), para. 87.

[30] Appellate Body Report, US – Tuna II (Mexico) (2012), para. 202.

[31] Appellate Body Report, EC – Sardines (2002), para. 176.

[32] Appellate Body Report, EC – Asbestos, para. 67.

[33] Ibid.

[34] For Louis Vuitton handbags, see https://de.louisvuitton.com/deu-de/homepage. For Fila handbags, see, https://www.fila.de/en/Women/Accessories/Backpacks-Waist-Bags/.

[35] Appellate Body Report, EC – Seal Products (2014), paras. 5.45 and 5.69.

[36] Appellate Body Report, EC - Asbestos, para. 69.

[37] Appellate Body Report, EC – Seal Products (2014), para. 5.12.

[38] Appellate Body Report, EC – Asbestos, para. 68.

[39] See https://www.epingalert.org/en/Search/Index?countryIds=C056&distributionDateFrom=2022-08-01&distributionDateTo=2022-09-30&viewData=G%2FTBT%2FN%2FBEL%2F47

[40] Appellate Body Report, US – Clove Cigarettes (2012), para. 120. See also, ibid, para. 156.

[41] For ‘likeness’ under the non-discrimination obligations of the GATT 1994, see Appellate Body Report, EC – Asbestos (2001), para. 101. The Appellate Body in US – Clove Cigarettes (2012) explicitly referred to its approach to ‘likeness’ its Report in EC – Asbestos (2001) and to determine ‘likeness’ under Article 2.1 of the TBT Agreement assessed the same factors. See Appellate Body Report, US – Clove Cigarettes (2012), para. 119. See also ibid., paras. 120 and 156.

[42] See e.g. Panel Report, Colombia – Port of Entry (2009), paras. 7.182 and 7.357.

[43] Appellate Body Report, US – Clove Cigarettes (2012), para. 180.

[44] Ibid., para. 182.

[45] Appellate Body Report, US – Clove Cigarettes (2012), para. 182.

[46] Ibid.

[47] Ibid., para. 101.

[48] Ibid., para. 175.

[49] See e.g., US – Tuna II (Mexico) and US – COOL.

^[50] Appellate Body Report, US – Tuna II (Mexico) (2012), para 297.

[51] Appellate Body Report, US – Tuna (Mexico) (2012), para. 319.

[52] Appellate Body Report, Australia – Plain Packaging (2020), para. 6.406.

[53] Appellate Body Report, US – Tuna II (Mexico) (2012), para. 315.

[54] Ibid., para. 317.

[55] Ibid., para. 314.

[56] Appellate Body Report, US – Tuna II (Mexico) (2012), para. 322.

[57] Ibid.

[58] Ibid.

[59] Appellate Body Report, US – COOL (Article 21.5 – Canada and Mexico) (2015), para. 5.334.

[60] Appellate Body Report, US – COOL (Article 21.5 – Canada and Mexico) (2015), para. 5.338. According to the Appellate Body, it would be more appropriate, for the respondent, once the complainant made a prima facie case, to establish that a proposed alternative measure is not reasonably available. See ibid., para. 5.339.

[61] Note that according to the EU 5G Toolbox, core networks functions are generally considered as critical, while radio access networks, transport and transmission functions, and internetwork exchanges are rated with lower criticality. See: https://digital-strategy.ec.europa.eu/en/library/cybersecurity-5g-networks-eu-toolbox-risk-mitigating-measures (pp. 39-40).

[62] Available at: https://digital-strategy.ec.europa.eu/en/policies/5g-research-standards.

[63] Appellate Body Report, US – Tuna II (Mexico) (2012), para. 356.

[64] Ibid., para. 363.

[65] See TBT Committee Decision on Principles for the Development of International Standards, Guides and Recommendations with Relation to Articles 2, 5, and Annex 3 to the Agreement, in WTO document G/TBT/1/ Rev.10, dated 9 June 2011, pp. 46–8, para. 6.

[66] Panel Report, EC – Sardines (2002), paras. 7.69 – 7.70, as upheld by the Appellate Body in Appellate Body Report, EC – Sardines (2002), para. 233.

[67] See https://www.itu.int/ITU-T/recommendations/rec.aspx?rec=14992&lang=en

[68] See https://www.itu.int/ITU-T/workprog/wp_item.aspx?isn=15006

[69] See https://www.ancom.ro/en/about-us/media-en/press-releases/romania-chairs-the-council-of-the-international-telecommunication-union/#:~:text=Romania%20became%20a%20member%20of%20the%20organization%20in%201866,%20just%20one%20year%20after%20the%20establishment%20of%20the%20Union

[70] Panel Report, EC – Sardines (2002), para. 7.110.

[71] Appellate Body Report, EC – Sardines (2002), para. 288.

[72] Article 2.9 of the TBT Agreement. The WTO Secretariat maintains the Technical Barriers to Trade Information Management System (TBT IMS), a publicly available database of all information provided by WTO Members in relation to technical regulations and other measures covered by the TBT Agreement standards. See https://eping.wto.org. Members submit over 3,000 to 4,000 new TBT notifications every year.

[73] Ibid.

[74] There is no entry in the TBT IMS relating Romania’s proposed Law 163/2021.

[75] Appellate Body Reports, EC – Seal Products (2014), para. 5.86.

[76] Panel Report, EC – Bananas III (Guatemala and Honduras) (1997), para. 7.239. See also Panel Report, Colombia – Ports of Entry (2009), para. 7.341; and Panel Report, US – Poultry (China) (2010), para. 7.415.

[77] Appellate Body Report, Canada – Autos (2000), para. 79. See also Appellate Body Reports, EC – Seal Products (2014), para. 5.86.

[78] Panel Report, Colombia – Port of Entry (2009), paras. 7.182 and 7.357; and Panel Report, US – Poultry (China) (2010), paras. 7.431–7.432.

[79] Appellate Body Reports, EC – Seal Products (2014), para. 5.88.

[80] Appellate Body Report, Korea – Various Measures on Beef (2001), para. 137. ‘Treatment no less favourable’ is assessed by examining whether a measure modifies the conditions of competition in the relevant market to the detriment of imported products.

[81] Panel Report, Thailand – Cigarettes (Philippines) (2011), para. 7.866.

[82] Ibid., para. 7.867.

[83] Appellate Body Report, EC – Bananas III (1997), para. 200.

[84] Appellate Body Report, EC – Selected Customs Matters (2006), para. 200.

[85] Appellate Body Report, US – Shrimp (1998), paras. 182-183; see also, Panel Report, US – COOL (2012), para. 7.861.

[86] Panel Reports, US – COOL (2012), para. 7.876.

[87] Panel Report, Colombia – Textiles (Article 21.5 – Colombia) / Colombia – Textiles (Article 21.5 – Panama) (2016), para. 7.361 (referring to Panel Report, Thailand – Cigarettes (Philippines) (2011), para. 7.899).

[88] Panel Reports, Colombia – Textiles (Article 21.5 – Colombia) / Colombia – Textiles (Article 21.5 – Panama) (2016), para. 7.362 (referring to Panel Reports, China – Raw Materials (2012), para. 7.696; Panel Report, US – COOL (2012), para. 7.851).

[89] Panel Report, China – Raw Materials (2012), paras. 7.751-7.752; 7.745-7.746.

[90] Appellate Body Report, US – Gasoline (1996), p. 22.

[91] Appellate Body Reports, EC – Seal Products (2014), para. 5.169

[92] This is the nexus required under paragraphs (a), (b) and (d) of Article XX of the GATT 1994.

[93] Panel Report, US – Steel and Aluminium Products (China), para. 7.128.

[94] Ibid. In this report, circulated in 2022 but subsequently appealed into the void by the United States, the panel sets out in detail the interpretative analysis leading to its conclusion that Article XXI(b) is not ‘self-judging’ or ‘non-justiciable’. See ibid, paras. 7.104 – 7.128.

[95] Panel Report, Russia – Traffic in Transit (2019), para. 7.82.

[96] Ibid.

[97] Ibid., paras. 7.76 and 7.111.

[98] Ibid., para. 7.75.

[99] Panel Report, Saudi Arabia – IPRs, para. 7.262.

[100] Panel Report, US – Steel and Aluminium Products (China), para. 7.139.

[101] Panel Report, Russia – Traffic in Transit (2019), para. 7.130.

[102] Ibid., para. 7.131.

[103] Ibid., paras. 7.119 and 7.131.

[104] Ibid., para. 7.132.

[105] Panel Report, Russia – Traffic in Transit (2019), para. 7.133.

[106] Ibid., para. 7.134.

[107] Ibid., para. 7.138.

[108] Ibid.

[109] Panel Report, Russia – Traffic in Transit (2019), paras. 7.76 and 7.111. Also, the political and economic differences, which may well be present, do presently not affect directly the defence and military interests of Romania, or the maintenance of law and public order in Romania.

[110] Panel Report, US – Steel and Aluminium Products (China), para. 7.139.

At the Ministerial Conference held in March this year, 67 WTO Members participating in the negotiations for the Agreement on Electronic Commerce (ECA) issued a “Declaration on Interim Arrangements for the Agreement on Electronic Commerce”.[2] In a nutshell, those Members announced their intention to have the ECA provisionally enter into force outside of the WTO legal framework, while continuing to push for a consensus-based decision for the incorporation of the ECA into Annex IV to the WTO Agreement as a “plurilateral agreement”.[3] 

The key elements of the interim arrangements are as follows: (i) the Interim Arrangements Annex becomes an integral part of the ECA and prevails over the latter in the event of conflicts; (ii) the Interim Arrangements Annex will automatically lapse if and when the ECA is incorporated into Annex IV to the WTO Agreement; (iii) the provisions of the DSU are incorporated by reference into the Interim Arrangements Annex, mutatis mutandis, and appeal arbitration is established as the default mechanism for appeals. 

In a recent news article, Chris Horseman reported that the European Union might be interested in pushing for a similar solution for the other plurilateral agreement that is currently in a state of limbo due to lack of consensus at the General Council: the Agreement on Investment Facilitation for Development (IFDA).[4] 

WTO Members are actively debating the matter. At the latest General Council meeting held in July this year, Members discussed a number of issues concerning the ECA interim arrangements, including the involvement of the Director-General as the depositary of the agreement and the extent to which the Committee on Trade-Related Aspects of Electronic Commerce (“ECA Committee”) to be established pursuant to the Interim Arrangements will receive support from the Secretariat once the agreement enters into force.[5] 

This post focuses on one aspect that has not caught much attention: the role of the ECA Committee in the administration of disputes under the Interim Arrangements. 

As the ECA Interim Arrangements will not formally be a WTO Agreement, the WTO DSU will not apply to it. The drafters of the ECA Interim Arrangements opted to incorporate by reference the rules of the DSU, mutatis mutandis, thereby creating a parallel set of provisions on dispute settlement, identical in content to, but formally distinct from, those of the DSU. This approach certainly has a merit: the drafters did not have to conduct extensive negotiations over dispute settlement matters, and could instead refer to a set of rules everyone was already accustomed to. 

The WTO DSU entrusts the Dispute Settlement Body (DSB) with a number of key functions that are crucial for the operations of the dispute settlement mechanism. Since recourse to the DSB was not possible for a non-WTO agreement, the drafters of the ECA Interim Arrangements assigned those functions to the ECA Committee. 

Paragraph 3.3 of the ECA Interim Arrangements reads as follows: 

3.3 For the purposes of this Annex:

(…)

(b) references to the "Dispute Settlement Body" or to the "DSB" in the DSU shall be understood as references to the Committee (…).

In practical terms, this means that for disputes under the ECA Interim Arrangements, the ECA Committee “shall have the authority to establish panels, adopt panel (…) reports, maintain surveillance of implementation of rulings and recommendations, and authorize suspension of concessions of obligations under the covered agreements”.[6] 

The implications for the ECA Committee are not negligible. The ECA Committee, in fact, will have to perform the same functions as other committees established under the WTO covered agreements. In addition to those, it will also administer dispute settlement operations, which is a task assigned to the DSB for WTO covered agreements. This is likely to increase the workload of the ECA Committee, in a situation where certain non-participants have already raised concerns regarding the extent to which the WTO Secretariat will support the activities of the committee. 

One way to minimize the burden on the ECA Committee would be for parties in disputes under the ECA (or the IFDA, if participants will also proceed with interim arrangements) to resort to arbitration pursuant to Article 25 of the DSU, as incorporated by reference in the interim arrangements annex, also for first instance disputes. Replacing panel proceedings with arbitrations means that the ECA Committee would not have to perform certain functions normally assigned to the DSB under the WTO DSU. Arbitration proceedings, in fact, are premised on an agreement between the parties and arbitration bodies need not be established by the DSB. Furthermore, the parties to arbitration proceedings agree to abide by the arbitration awards, such that the DSB is not called to adopt arbitration awards. 

Pursuant to Article 25.2 of the DSU, in the event that the parties agree to resort to arbitration, they shall also “agree on the procedures to be followed”. This can be done on the basis of ad hoc arrangements or by means of a template set of procedures to be followed in every dispute, similar to the MPIA. The procedures need not be unnecessarily complex or different from panel proceedings under the DSU. In this respect, inspiration could be drawn from the EU-US arrangements in the Steel and Aluminium and Additional Duties disputes, where the parties agreed to replace panel proceedings with arbitration proceedings. They did so to “freeze” their dispute while they sought a negotiated solution, anticipating that negotiations could take longer than the 12-month limit envisaged in Article 12.12 of the DSU for the suspension of disputes. The parties thus terminated their panel proceedings and initiated arbitration proceedings, appointing the original panelists as arbitrators. In terms of the procedures, the parties agreed to reproduce the “rules, procedures and practices on panel proceedings that would have applied to the panel proceedings”, with the exception of the twelve-month limit set forth Article 12.12 of the DSU.[7] At the time of writing this post, both arbitration proceedings are still formally suspended. 

Replacing panel proceedings with arbitrations also for first instance disputes would thus be coherent with economical approach of the ECA Interim Arrangements. When negotiating interim arrangements for the IFDA, drafters could consider the solutions suggested in this post as a way to ensure that the committee to be established under that agreement will devote all its energies to its core business, without having to also perform the functions originally assigned to the DSB under the WTO DSU.

[1] International trade lawyer, ccantore@pm.me.

[2] WTO, Declaration on Interim Arrangements for the Agreement on Electronic Commerce, WT/MIN(26)/42 (27 April 2026).

[3] Plurilateral agreements are binding only for those WTO Members that have accepted them and do not create rights and obligations for other WTO Members (Article II:3 of the WTO Agreement). Pursuant to Article X:9 of the WTO Agreement, the Ministerial Conference (or the General Council in the intervals between meetings of the Ministerial Conference) may decide “exclusively by consensus” to add a new plurilateral agreement to Annex IV to the WTO Agreement.

[4] C. Horseman, “EU to push for interim fix on stalled WTO investment pact” (available at: https://borderlex.net/2026/09/02/eu-to-push-for-interim-fix-on-stalled-wto-investment-pact/, last accessed, 15 September 2026).

[5] P. Ungphakorn, “Text: WTO Head on the E-Commerce Plurilateral Deal and Secretariat Support” (available at: https://tradebetablog.wordpress.com/text-dg-reply-secretariat-and-e-commerce-plurilateral/, last accessed, 15 September 2026).

[6] Article 2.1 of the DSU.

[7] WTO, US – Steel and Aluminium Products (EU) – Recourse to Article 25 of the DSU, WT/DS548/19, paras. 5 and 6; EU – Additional Duties (US) – Recourse to Article 25 of the DSU, WT/DS559/7, paras. 5 and 6.

First, he describes the industrial policy threat coming from China:

[China has constructed] a vast, decentralized system of state-backed innovation finance, channeling subsidized capital into priority high-tech sectors and encouraging firms to compete ruthlessly for scale. The result is a market systematically tilted by state intervention, one in which Chinese firms can overbuild capacity and flood markets before American and allied competitors can scale up production.

Then he explains what he thinks a U.S. version of industrial policy that responds to the Chinese threat should look like:

But Washington should not aim to copy Beijing’s state-backed playbook wholesale. After all, the United States has plenty of advantages over China, including its entrepreneurs, universities, and capital markets, which are the deepest in the world. But those markets optimize for efficiency, not resilience or security. The challenge, then, is to turn capital into capability, to channel investment toward the industries that anchor modern national strength and economic vitality: advanced computing, biotechnology, robotics, critical minerals, pharmaceutical precursors, advanced manufacturing, and the electric-energy backbone. And the solution is an ambitious national strategic investment enterprise, headlined by a federal Strategic Investment Fund.

...

Throughout its history, the United States built institutions that channeled capital toward a national strategy. Resurrecting that tradition today would mean establishing a new kind of public investor: a U.S. Strategic Investment Fund (SIF), a federally chartered, market-facing public investor with its own balance sheet, designed to invest alongside private capital in strategically critical industries, both mature and emerging, where markets alone have proven insufficient.

He offers a lot of detail in support of his plan, a key element of which is to try to institutionalize industrial policy rather than have it be carried out in an ad hoc way (as has been the case).

From what I can tell, Sullivan's views on industrial policy reflect the conventional wisdom among many establishment Democrats these days. However, after the experience with these kinds of policies during the Biden administration, it may be time for a rethink. I have several specific criticisms of Sullivan's piece that could help move things in this direction (for any Democrat who is interested!): (1) he gets some of the history of U.S. economic policy wrong; (2) I'm not sure he's learning the right lessons from China's successes; (3) his proposals could lead to an international subsidy race, and that hasn't been a great outcome generally for the economy; and (4) the Biden administration tried a more limited version of his proposal and it didn't work out very well politically for the Democrats. Let's go through each one. (I'll have more to say about all this at some point, but it's going to have to wait.)

Learning the right lessons from U.S. economic history

First, Sullivan tells a story of U.S. economic history in which most of our industrial success is based on state capitalism. However, the history he presents is a bit dubious at times.

For example, he says "U.S. shipbuilding ... collapsed after the 1980s, when government subsidies were withdrawn, ... ." But as shown here, the U.S. shipbuilding situation is more complicated than that. From 1974 to 1978, the U.S. built 14 to 24 ships per year. It's true that after subsidies were cut in 1981, production declined to between 5 and 12 from 1981 to 1986, and continued to fall after that, so maybe this is what he has in mind by "collapse." To put these figures in perspective, however, in 1978 Japan delivered 471 ships. And for additional perspective, note that in 1981 the U.S. ranked 15th worldwide in shipbuilding, while last year the U.S. ranked 19th. So basically, without the subsidies, U.S. shipbuilding went from small to tiny, which doesn't feel like a "collapse." (A 1980 NY Times headline makes clear that serious problems existed in the industry even before the subsidy cuts: "U.S. Shipyards in the Doldrums; Orders Down Sharply")

Learning from history is always a good idea, but we need to make sure we have an accurate understanding of what happened. 

China's subsidies are a problem but let's not idealize them too much

Second, a key element of the policy vision Sullivan sets out is that it is a response to China: China provides massive subsidies for its industries, he says, and therefore the U.S. needs to adopt its own version of these subsidies in order to compete.

China does provide a lot of subsidies. Something missing here, though, is an evaluation of why exactly China has been successful in certain industries, and how much of a role the subsidies played in the success. In my view, it's an open question as to whether subsidies have been the driving force, as there are a wide range of factors other than subsidies and state interventions that help explain China's economic success. Having a 1.4 billion person internal free trade area as a market certainly seems like a plus!

Furthermore, evaluations of China's industrial policy are not always very glowing, and there are sectors where China gave lots of subsidies and still isn't competitive.

Having said that, by all means, the U.S. and other countries should challenge China on its subsidies. But just like with our own history, we should make sure to learn the right lessons from China's successes and failures.

(And speaking of history, keep in mind that back in the 1980s, people were talking about Japan overtaking the U.S. economically, but by the late 1990s it was clear that their predictions had missed the mark. Along the same lines, the situation with China could look very different in 10 years.)

Does anyone really win a subsidy race?

Third, it's important to think through the likely international response to a further expansion of industrial subsidies. China and the U.S. aren't the only players here. What is likely to follow from a ratcheted up U.S. industrial policy is that other countries will increase subsidies to their industries as well. At the end of all the ratcheting, the countries that can afford to participate in this subsidy race would be back on somewhat equal ground in terms of competitiveness (while less developed countries would fare badly), with taxpayers having funded the race for little if any advantage. That outcome does not seem ideal. In past decades, we saw the impact of similar developments on industries such as steel, and it's not great.

A better approach would be agreements not to subsidize, or at least to limit subsidies or to regulate the economic harm they may cause. As it happens, we have something along these lines in place at the WTO, but for some reason people haven't been enforcing it. I think I have a sense of why the Trump administration didn't enforce the WTO's subsidy rules. I'm not sure I understand why the Biden administration didn't do so though. If Sullivan is going to make the case for industrial policy, I think it's worth pushing him to explain how he thought about this issue while he was in the Biden administration. Why didn't he argue for challenging China's industrial policies? And how, in his view, are the U.S. and the world better off with China, the U.S., and many other countries competing with subsidies rather than competing in a more market-oriented system? Is the future he envisions for many industries something like the endless Airbus-Boeing subsidy battle (which was on a brief hiatus but now could be back), and is that really where we want to go with all of the industries that have been declared strategic? If so, it's going to cost taxpayers a lot of money, and it's not clear that any country will improve its position in the end. (And with the current state of the budget deficit/debt, it seems like a bad time for the U.S. to go down this road.)

Didn't we already try all this?

Speaking of things that happened during the Biden administration, my final point here is that, as I read it, Sullivan's proposal is basically to take the Biden administration's approach to economic policy and go a step or two further, turning industrial policy from an ad hoc effort to an institutionalized one. Before we go that route, however, I think we should ask whether this approach worked the first time, either in terms of policy outcomes or politics. Given the results of the 2024 election, there's a strong argument that it did not work out politically for those who pushed it.

Of course, Sullivan might argue that things didn't work out because of Biden personally, and that it would all go much better with a different Democratic president (i.e., Bidenomics would work better without Biden). Or perhaps he would say the efforts needed to have been undertaken more vigorously and comprehensively, and the institutional aspects of his "Strategic Investment Fund" proposal would make everything go more smoothly the next time around.

It's difficult to evaluate hypothetical arguments like these. However, if I were a Democrat thinking about a post-Trump economic policy, rather than just press forward with more of what they did last time, I would do a thorough post-mortem on the Biden administration's economic policies and think about the results of those earlier efforts to revive industrial policy. Figuring out what went wrong is essential for not ending up in the same place next time.

Then some Canadian wildfire smoke blew down to the U.S., and in a social media post on July 17 Trump put forward a vague tariff threat in response. But there are plenty of tariff threats these days, and it didn't necessarily seem like this one was going anywhere.

A few days later, though, on July 20, the Trump administration announced the imposition of tariffs under Section 338 based on accusations that Canada was discriminating against U.S. products in various ways. As the first ever use of Section 338 to impose tariffs, I think it's fair to say that this came as a bit of a surprise. Sure, people had mentioned Section 338 as a possible avenue for tariffs all the way back in 2016, but it hadn't been used to date, so why would it be now? Did the somewhat offhand wildfire tariff threats pave the way for a new tariff action that otherwise would not have happened? Did trade officials feel like that threat needed to have some sort of follow through? Who knows! Maybe the insider details will come out at some point.

Regardless of exactly how we got there, the Section 338 tariffs led to Canada-U.S. relations spiraling downwards. Canada announced its retaliation on August 25, and these tariffs went into effect today. In terms of possible U.S. escalation, all we have as of this writing is a threat by Trump to ban the sale of Bombardier jets. We'll see if anything else emerges in the coming days. [UPDATE: Action to remove Canadian-origin products from GSA procurement was just announced by Trump.][UPDATE II: Here comes more retaliation: Excluding Certain Canadian Products from Importation into the United States in Response to Continued Discrimination Against the Commerce of the United States with Respect to Dairy; Excluding Certain Canadian Products from Importation into the United States in Response to Continued Discrimination Against the Commerce of the United States with Respect to Motor Vehicles; Excluding Certain Canadian Alcoholic Beverages from Importation into the United States in Response to Continued Discrimination Against the Commerce of the United States with Respect to Alcoholic Beverages; there was also this: Modifying the Scope of Products of Canada Subject to the Additional Duties Imposed to Offset Canadian Discrimination Against the United States with Respect to Motor Vehicles; Modifying the Scope of Products of Canada Subject to the Additional Duties Imposed to Offset Canadian Discrimination Against the Commerce of the United States with Respect to Alcoholic Beverages]

I'm not sure what the Trump administration expected in response to the Section 338 tariffs. Administration officials often make reference to the fact that most U.S. allies eventually agreed to deals in the face of U.S. tariffs, whereas only China and Canada retaliated. It's possible these officials see trade leverage as the key here. Thus, in their view, the use of trade leverage pushed other allies into signing deals, and Canada should/would follow the same path with enough of this trade leverage in place. Section 338 would ratchet things up and get the deal done.

In my view, though, it's a mistake to see trade leverage as the key here. Arguably, it was security leverage that made the other deals happen. Everyone's situation is a little different, but, to take a couple examples, many EU nations feel dependent on the U.S. in the face of security threats from Russia, as do Japan and South Korea based on threats from China and North Korea. It was this security dependency that made these deals possible.

Importantly, Canada is in a very different position from most U.S. allies in relation to security, as it doesn't face the same security threats. I occasionally hear vague suggestions that Canada is under threat from China or Russia, but in a world where China or Russia are bombing/attacking/invading Canada, so much else has gone wrong that this is the least of anyone's worries. Canada simply does not face security threats in the way that other U.S. allies do. And the practical impact of this difference is that the U.S. doesn't have the same leverage over Canada as it does over other allies.

And then on top of that, while nobody's domestic politics are positive in relation to being publicly pushed around by the U.S., Canadian politics are particularly sensitive to this, especially in light of the rhetoric coming from President Trump and others in the administration.

So what happens now? My sense is that when they were at the negotiating table, Canada and the U.S. were not actually that far apart on the substance. A few tweaks here and there could be enough to close the gaps. However, the politics of all this could get in the way of a deal. The rhetoric has gotten heated and it can be hard to walk things back.

In the short term, this situation probably helps the Liberals in Canada politically, which makes a deal less attractive to the leadership there. In the U.S., it could hurt the Republicans slightly in the upcoming midterms, but this may not register enough for the administration to care (and Trump likes tariffs and trade wars, so he is not likely to back down in response to a small amount of political uncertainty).

Summing all this up: It's quite a mess we are in, and I'm not sure whether and how the key actors can get us out of it. While there's always a way for each side to make small concessions, declare victory, and get out, I don't know when people will feel the economic or political incentive to do so.

The ANZSIL International Economic Law Interest Group (IELIG) was established in 2011 to provide a forum for those interested in international economic law, especially international regulation of trade, investment, intellectual property, and monetary law. The work of the Interest Group covers all areas of international economic law and its implementation in domestic law, embracing both theoretical and practical aspects. The IELIG aims to maintain relevance for government, practitioners and academics alike and encourages discussion and exchange of ideas among all those involved in this field.

2026 Workshop (online)

The 2026 ANZSIL IELIG Workshop will be held via Zoom on Friday 27 November 2026. Its aim is to provide an informal setting for discussion of topical issues and works in progress.

We are honoured to have Professor Jarrod Hepburn, University of Melbourne, as this year’s keynote speaker. He will speak on ‘The Flexible Sources of Investment Law’. Jarrod received the ANZSIL Article Prize this year for ‘The Legal Justification for the Doctrine of Legitimate Expectations in International Investment Law’ (2025) 36 European Journal of International Law 43. As part of the workshop, we celebrate this achievement with panel on Jarrod’s work.

We invite paper proposals from ANZSIL members, practitioners in the field as well as staff from MFAT, DFAT, and the Attorney-General’s Department. We encourage PhD students, early career practitioners and academics to submit a proposal. Paper proposals may be on any topic within the work of the IELIG.

All speakers are invited to submit their paper following the workshop to the Australian Year Book of International Law or the New Zealand Yearbook of International Law for possible publication. We also invite ANZSIL members and MFAT/DFAT/AGD staff to attend the workshop even if not presenting a paper.

Submission of proposals

If you wish to submit a paper proposal, please submit an abstract of 300 words by email to christian.riffel@canterbury.ac.nz by 9 November 2026. Please include the heading on your email message ‘ANZSIL IELIG Workshop Proposal: [Your Name]’.

Prof Chris Riffel, University of Canterbury | Te Whare Wānanga o Waitaha

Dr Jose-Miguel Bello Villarino, University of Sydney ANZSIL IELIG Co-Chairs

A monthslong crackdown within the Trump administration to soften its trade rules all started with a flap over pasta.

In the aftermath, Commerce Secretary Howard Lutnick and other top Commerce Department officials are leaning on agency staff not to drive up tariffs on grocery items and other goods ahead of the midterm elections.

...

Agency leaders first stepped in late last year, after a Commerce Department investigation determined in September that leading Italian pasta makers had been underpricing the products they sold in the U.S., in violation of trade rules. The proposed punishment — sky-high tariffs on the pantry staple — drew headlines at a time when the administration was trying to convince Americans its trade agenda was not driving up prices.

...

In particular, he and other top Commerce officials have discouraged staff from pursuing the most aggressive possible tariffs in the pasta case and other similar investigations into trade violations, according to four people familiar with the matter, granted anonymity to share unreported details about the internal discussions.

In some cases, senior officials instructed staff to recalculate rates that were deemed too high, two of the people said. The result has been much lower duties on violators in recent months compared to the past.

...

Lutnick signaled to staff across meetings and private correspondence in the fall of 2025 that the administration would no longer prioritize the most aggressive forms of enforcement, according to the four people. In a meeting in March, senior Commerce officials told private-sector lawyers about the new approach, stressing that the administration would no longer back the most aggressive form of enforcement, two of the people said.

...

Since the fall, Commerce officials overseeing antidumping and countervailing investigations have selected the lowest tariff option under consideration in the overwhelming majority of cases, two of the people familiar said. One of the people said that career staff initially chose the highest rate 90 percent of the time, but senior leadership would “ask to redraw the numbers” if all the options seemed too high.

I'm not totally sure how this story made its way into the press, although I can imagine that domestic industry lawyers were getting annoyed at the Commerce Department and perhaps looked for ways to make their concerns public.

Thinking about the big picture here, one possible takeaway from recent U.S. trade policy developments is that the unrestrained use of tariffs other than AD/CVDs has brought a bit of restraint to AD/CVDs. But I'd want to hear from practitioners about this before reaching any firm conclusions on the connection.

The last time I remember mainstream media reporting on the inner workings of AD/CVD calculations was a 2003 Washington Post article, which was about Commerce pushing duties in the direction people would generally expect:

... Lawyers who defend foreign companies in dumping cases complain that [Joseph] Spetrini stands out for his tendency to use his discretion in ways that hurt their clients, the result often being higher "dumping margins" set by the department on imported goods, with concomitantly higher duties.

"Bending the rules in favor of petitioners – that's what he views as his job, frankly," said one trade lawyer, speaking of the U.S. parties in dumping cases.

...

... [Spetrini] has considerable latitude in interpreting the rules and deftly skews the outcome toward high margins, according to a number of former employees who relate similar accounts of how he operates. A term several used to describe Spetrini's approach is "margin shopping," or looking at different ways to calculate an importer's costs or prices and choosing the one that helps fatten the prospective margin to a desirably high level.

First up, Colombia explains how the TRIPS Agreement uses intellectual property rights to reward, and therefore encourage, innovation, focusing on the example of patents. A key point of the submission is that there are two separate elements (which it refers to as "prerogatives") – exclusivity and remuneration – of the patent reward system:

1. Colombia notes that the TRIPS Agreement establishes two main mechanisms to reward innovation through the protection of intellectual property rights (IPRs): i) exclusivity and ii) profit potential. Indeed, in the example of patents, Article 28 of the TRIPS Agreement states:

Article 28 – Rights Conferred

1. A patent shall confer on its owner the following exclusive rights:

(a) where the subject matter of a patent is a product, to prevent third parties not having the owner's consent from the acts of: making, using, offering for sale, selling, or importing for these purposes that product;

(b) where the subject matter of a patent is a process, to prevent third parties not having the owner's consent from the act of using the process, and from the acts of: using, offering for sale, selling, or importing for these purposes at least the product obtained directly by that process.

2. Patent owners shall also have the right to assign, or transfer by succession, the patent and to conclude licensing contracts.

2. According to the aforementioned article, a patent holder has two distinct prerogatives. The first is a prerogative of exclusivity, which prevents third parties from performing certain acts with respect to the patented product or process without their consent. The second is a prerogative of remuneration, under which a given use generates an entitlement to compensation. In the ordinary patent bargain, the two are bundled. However, as discussed below, the multilateral IP framework already recognizes configurations in which a right to remuneration is preserved independently of any power to prohibit use. Because the two prerogatives differ in scope and serve different functions, they do have distinct effects on innovation, access, market competition, technological diffusion, industrial learning and development.

(footnote omitted)

Colombia then notes the downsides to the exclusivity element:

4. Regarding the possibility of excluding third parties from using a technology, this prerogative can generate tensions with technology users, limit cumulative innovation, hinder interoperability, restrict competition, or increase barriers to entry, both for companies in industrialized and developing countries. Among countries, this exclusivity often serves as an industry protection, as it creates a "moat" in a specific market, with potential significant effects on other countries and its technological upscaling and development strategies.

On the other hand, Colombia sees the remuneration element as more positive and talks about the possibility of applying it independently of exclusivity:

5. By contrast, the remuneration prerogative – the possibility of receiving remuneration for a technological contribution – is in many cases sufficient on its own as a legitimate and sufficient incentive for innovation, without the negative impacts. The differentiated effects of exercising or privileging one of these two prerogatives over the other are relevant to the design of an optimal IPR reward system. In this context, the question is not whether innovators should be rewarded, but whether the reward should necessarily depend on the possibility of exclusion.

With regard to alternative approaches to remuneration that do not involve exclusivity, Colombia notes the following:

10. Other existing incentive mechanisms can inform this discussion. These include statutory licencing, related rights, pools, innovation awards, innovation inducement prices, reward funds, conditional public funding, open knowledge models, etc., under which the remuneration prerogative is kept, but the exclusivity one is not.

11. Taken together, these experiences demonstrate that the promotion of innovation could be better anchored focusing on the monetary rewards, without the potential adverse effects of granting temporary monopolies of exclusion.

(footnote omitted)

Colombia later elaborates on the point about separating out the two elements:

12. The international community could achieve significant progress on several IP discussions by independently analysing the two described prerogatives. Discussions on technology transfer, IP rights in pandemics, IP rights for environmental objectives, the role of the objectives, principles, and flexibilities of the TRIPS Agreement, VMAT for tech transfer (Voluntary and Mutually agreed terms), industrial protection and competition, etc., have proven to be almost intractable in different international fora. Exploring different models with autonomous prerogatives – for example, using different periods of protection for the two prerogatives, or providing market reward incentives by augmenting the remuneration mechanisms when a commitment to limit the right to exclude is made – could provide venues for better collective solutions. Likewise, the developmental dimension related to technology access and technology learning and technology upgrading, could be better served by treating the two prerogatives independently, and create a better scenario for trust and rule making at the WTO, even in the context of the reform of the organization.

Going forward, Colombia suggests the following discussion:

13. Consequently, Colombia, invites an exploratory thematic discussion on alternative innovation reward mechanisms currently in use, in which exclusivity has been modulated without eliminating the remuneration reward, and its effects on different types of industries and different types of countries.

In the English version of the document, the word "prizes" does not appear, but when I look at the Spanish version I think I see (using Google Translate) "innovation inducement prizes" (or maybe "awards") rather than "innovation inducement prices" (as stated in the English document), and Colombia clearly has something along these lines in mind. Prizes are arguably a better way to encourage innovation than the current patent system, so I'm glad Colombia raised this.

Colombia's communication was discussed at a meeting of the TRIPS Council on July 22-23. So far, all we have is the brief summary of the meeting put out as a WTO news item, which explains the reaction to the communication as follows:

Members also discussed a submission by Colombia entitled "Autonomous examination of TRIPS prerogatives, and existing examples" (IP/C/W/742). The submission invited members to consider thematic discussions on alternative innovation reward mechanisms, with a proposal to assess the impact of IP exclusivity as separate from a right to remuneration, and to discuss their effects across industries and regions. Some members expressed interest in continued dialogue on this topic, while others questioned the premise of the submission and emphasized the adequacy of the existing IP framework.

It's not too difficult to imagine which Members fell on each side of the debate. More details should be available soon when the minutes of the meeting are released, and I'll come back to it then if anything interesting was said.

Neguse: So going back to you, Mr. Steil, socialism [is] owning the means of production, right?

Steil: That's a piece of it, yeah.

Neguse: Okay. What would you call the Trump administration acquiring a 10% equity stake in Intel, a 15% stake in rare earth producer MP Materials, and a 10% stake in Lithium Americas and Trilogy Metals?

Steil: It's an America first policy to make sure we can outcompete China.

Neguse: Do you believe that the government should own grocery stores?

Steil: No.

Neguse: That would be socialist to you?

Steil: I think that would be a type of overreach in the federal government, absolutely.

Neguse: Would that be socialist to you?

Steil: Yeah, that'd be a form of socialism.

Neguse: Okay. And if Donald Trump announces tomorrow that he's going to take a 20% equity stake in Kroger, would that be socialism to you?

Steil: I'd want to see the details.

Neguse: Oh, you want to see the details?

Steil: No, but understand the difference. ...

...

Neguse: ... You said if government owns grocery stores, that would be socialist. But if, of course, Donald Trump were to announce an equity stake in a grocery store, well, I'd have to really see the details ...

...

Steil: I think there's a meaningful and substantive role of the federal government as it relates to foreign policy and national defense. I think that's why we look at the National Defense Production Act. ...

... [Rare earth materials, oil] have national policy and national security interests at stake, which is why we have the Defense Production Act, which gives the federal government certain roles that are non-traditional in the capitalist system. ... And then the question is, where on that dial? And so the question then becomes: Some folks on your side of the aisle would like to have government-run grocery stores. I think that's nuts.

...

Neguse: ... I asked you a very simple question. You answered it. The government owning grocery stores would be socialist in your view. I then posed the question as to whether or not the President announcing that the United States government would take a 10% or 20% equity stake in one of the largest national retailers in our country, whether that would be socialist, you said, I'd have to know the details. ... So, you then answered and took us down the road of the justifications with respect to national security and foreign mining and all the rest, so we are back to square one. As you said, you said very clearly, there are people in my party who believe the government should own grocery stores, and I'm trying to get a better sense as to whether or not you agree. And I sense that if Donald Trump were to announce that he believes it's the right thing to do, that you would agree. You have an opportunity to tell me no, that if the President announced a 10 or 20% equity stake in a national grocer, that you would not support that, you believe that's socialist.

Steil: I think there would need to be a national security rationale for the action. I don't see any ... in your hypothetical, I don't see a national security purpose.

...

Neguse: Okay. So, a test for you as to whether or not the U.S. government acquiring a stake in any American company or any American industry, as to whether or not that is socialist, is whether or not there's any kind of national security implication. And essentially, the United States government can, so long as they can put together this cogent rationale, Donald Trump can take equity stakes in whatever companies he deems necessary – the U.S. government, I should say, taking equity stakes, he's initiating – as long as that national security prerogative is established ... .

Steil: No, I think you're wildly overstating.

In reaction, let me first note that I mentioned similar issues last year, when Commerce Secretary Howard Lutnick didn't seem to deny that what the Trump administration was doing could be characterized as state capitalism, suggesting that it was justified as a response to China. In the Neguse-Steil exchange, the characterization has shifted from state capitalism to socialism, which I think is a fair way to describe some of the things that have been happening recently.

It seems to me that Steil's position as set out above is that, contrary to the title of the House Resolution, socialism should not, in fact, be condemned or denounced in all its forms, but rather should be embraced in particular circumstances. I understand that he only wants government ownership of the means of production when "national security" is at stake, but the concept of national security is very broad in theory, and in practice recently has been much broader than I ever imagined, with few if any guardrails. As a result, a limitation on government ownership of industry to circumstances involving national security does not provide much of a restriction on what I always thought was one of the main concerns people had with socialism, and instead appears to be a somewhat open-ended endorsement of such ownership.

Bringing this topic closer to the main focus of this blog, what does the current Trump administration/Republican party position on these issues mean for U.S. international economic policy? If Republicans are going to support U.S. government ownership of the means of production for national security reasons, that could, in theory, lead to some significant changes in international economic governance and relations. Perhaps the rules need to be refined and clarified so that all actors in the trading system are aware of the specific circumstances when government ownership of the means of production is permitted, so governments can have more certainty in knowing when and how they can do it?

That, of course, was a rhetorical question, and I'm pretty sure I know the answer: The Republican politicians who support U.S. government ownership of the means of production for security reasons would not apply these principles to foreign governments, and do not think international rules need to be adjusted to provide more guidance or leeway. Rather, the Trump administration and these Republican politicians are still likely to object to foreign government policies and practices of this sort when they have an impact on U.S. companies. And if foreign governments invoked national security in the same way the Trump administration has been doing, I don't think that would make a difference for the Trump administration's response. To make that point more concrete, if you think about possible foreign government interventions of this sort in the tech sector, I can imagine the Trump administration's response would be highly aggressive.

Having said all that, I'd still like to see some questions from trading partners on these issues, to clarify the scope of what the Trump administration thinks governments – both the U.S. and others – may do here. What exactly are its views on government ownership of industry based on a national security rationale? Given how common the practice has become, this seems worth fleshing out. Ideally, these conversations would happen at the WTO, so there would be a published account of the discussion, but even bilateral conversations could be useful.

And while we are on the subject of questions for the Trump administration, Congress could ask some about the nature and purpose of the equity stakes the administration has been taking, as well as how these actions would be considered under our international obligations. That might be more useful than voting on non-binding resolutions such as the one that triggered this discussion.

Finally, a big issue here is whether, or to what extent, this view on the part of some Republican politicians will outlast Trump's presidency. In 2029, will they return to the more free market views they held for several decades before Trump? I have been skeptical that there is a coherent big government version of economic policy that works in terms of politics on the right, in the absence of someone with Trump's personality dominating the scene. We'll find out in a couple years whether I'm correct about this.

In August 2023, China suspended imports of all Japanese aquatic products following the release of ALPS-treated water from the Fukushima Daiichi Nuclear Power Plant. The measure was widely discussed in Japan as an example of economic coercion. I previously examined its consistency with the WTO SPS Agreement in an EJIL: Talk! post.

More than three years have now passed, but Japanese seafood exports to China remain negligible. We can now look beyond the immediate impact of the measure and ask a different question: did this economic coercion work?

This post looks at how Japanese seafood exports changed after the restrictions, with particular attention to whether exports that could no longer go to China were redirected to other markets. 

What happened to Japanese seafood exports?

For this analysis, I identified 213 seafood product categories using Japan’s nine-digit statistical classification, based on the seafood categories in Japan’s trade statistics. I then traced their monthly export values from January 2022 through July 2026. The accompanying spreadsheet file contains the underlying product-level data.

Of these 213 products, 106 had been exported to China before the restrictions. Their average monthly export value to China fell by 99.9 per cent after the measure was introduced. In practical terms, direct exports to China almost disappeared.

At the same time, the average monthly export value of these products to all other markets — the rest of the world (ROW) — increased by about 30 per cent. At first sight, this looks like trade diversion: exports that could no longer go to China may have been redirected to third-country markets.

But there is one important problem. The other 107 products had not been exported to China before the restrictions, yet their average monthly export value to the ROW also increased by about 26 per cent. It is therefore possible that Japanese seafood exports were simply increasing more generally during this period. 

Was it trade diversion?

The next question is whether the increase in ROW exports really reflected trade diversion following the loss of the Chinese market. I focused on how dependent each product had been on China before the restrictions.

The idea was simple. If trade diversion occurred, products that had relied more heavily on China should show larger increases in exports to other markets after the restrictions. I therefore tested whether this pattern could be seen in the trade data.

The result supports this view. A 10 percentage point higher pre-restriction share of exports going to China was associated with an estimated 16.9 per cent higher monthly export value to the ROW after the restrictions. This makes it difficult to explain the increase in ROW exports only by general economic conditions or a broader increase in Japanese seafood exports. The result is consistent with trade diversion following the loss of the Chinese market. The accompanying Python replication code allows readers to reproduce this estimate. 

What does this tell us about economic coercion?

China’s restrictions almost eliminated direct exports of the affected products to the Chinese market. At the same time, however, products that had relied more heavily on China showed larger increases in exports to third-country markets. The restrictions therefore caused a major disruption in bilateral trade, but the trade data also point to reallocation toward alternative markets.

Previous research has shown that exporters targeted by economic coercion can reduce its economic impact by redirecting trade elsewhere. Ferguson, Waldron and Lim (2023), for example, identify reallocation to alternative markets as an important response to China’s trade restrictions against Australia. The Japanese seafood case shows a similar pattern.

What is Pix?

The Central Bank of Brazil explains Pix as follows (using Google Translate): "Within the Brazilian instant payment (IP) ecosystem, Banco Central do Brasil (BCB) created Pix, the Brazilian IP scheme that enables its users — people, companies and governmental entities — to send or receive payment transfers in a few seconds at any time, including non-business days." In "The Political Economy of Brazil’s Pix Payment System," Jeff Alvares, senior counsel at the Central Bank of Brazil, provides more background as follows:

Brazil’s payments landscape before Pix launched in 2020 exemplified market frictions. A few major banks dominated both infrastructure and customer-facing applications. Wire transfers cost $1.50-3.00 USD and took hours or days to clear. Credit card fees reached 2.2% on average, compared to 1.7% in the U.S., 1.5% in Canada, and 0.3% in the European Union. Around 45 million Brazilians, about 29% of the population, remained unbanked and excluded from digital commerce. Existing players had little motivation to incur the fixed costs of infrastructure that could cannibalize their card fees and transfer charges. Incumbent banks and card networks benefited from fragmented, costly rails.

The Central Bank conceived Pix to overcome this structural inertia. It now operates the Instant Payment System (SPI) infrastructure to provide real-time settlement around the clock. Use of this rail is mandatory for banks and major PSPs. Pix itself, the payment scheme running atop the SPI, is also Central Bank-controlled, with mandatory participation, and zero pricing for services to individuals and small businesses. (In common parlance, Pix refers to both the payment scheme and the vertically-integrated payment system including SPI).

The cornerstone of Pix’s integrated design is a combination of legal and economic barriers to potential competing payment schemes.

...

Pix’s fast, affordable, and near-universal model has achieved transformative results. It now reaches 177 million users (83% of population) and accounts for 51% of all payment methods, displacing payment cards and bank instruments alike (see chart). It processes seven billion monthly transactions worth $550 billion. For millions, it became the gateway to the digital economy.

...

Pix delivers transformative social benefits, but it does so through foreclosure rather than through competition among payment schemes. This tradeoff raises profound questions for antitrust policy and international trade law. ...

See also this follow-up piece by Alvares: "What Brazil’s Pix Reveals About WTO Rules for the Platform Economy."

USTR investigates Pix

USTR's July 2025 notice of initiation of a Section 301 investigation of Brazilian trade practices raises a general concern about "government-developed electronic payment services" without mentioning Pix by name:

Brazil also appears to engage in a number of unfair practices with respect to electronic payment services, including but not limited to advantaging its government-developed electronic payment services.

In the public comments it received, USTR heard from groups on both sides of the issue, setting out the opposing views. Below are some excerpts.

Pix critics

Critics of Pix included the US Chamber of Commerce and the Information Technology Industry Council (ITI).

The US Chamber argued the following:

PIX has been successful in expanding financial inclusion and digital access in Brazil since its launch by the Central Bank of Brazil (BCB) in November 2020. American firms partner with PIX, laud its success, and consider the continued success of PIX to be a priority. However, we maintain concerns around the fact that BCB both regulates and competes with payment arrangements (PAs). While it is not uncommon for central banks to operate one or more payment systems and supervise the private sector, the BCB has failed to establish governance procedures that avoid conflicts of interest and crowding out the private sector.

The BCB is the sole systemic regulator of Brazil's financial sector, setting market entry conditions, operational standards, and pricing frameworks for all PAs. At the same time, it operates PIX, a Central Bank-run retail payment network, which is itself a PA and a direct competitor in the marketplace. As such, U.S. electronic payment system companies must compete against their own regulator. This contradicts international best practices and guidance from the OECD, World Bank, CPMI, BIS, and WTO, all of which call for a clear separation between oversight functions and commercial activities.

Further, PIX is not subject to regulation and supervision from a third-party regulator and overseer or price or budget controls in a way that would be materially equivalent to the kind of regulation and supervision that private players face. ...

...

Brazil should commit to implementing effective measures that address the regulatory conflict of interest and anticompetitive practices described above. PIX should comply with the same regulatory, cybersecurity, and operational standards imposed on private-sector platforms and be subject to independent third-party regulation and supervision. Brazil should also ensure a clear separation between PIX's operational and regulatory/supervisory functions within the BCB to safeguard regulatory neutrality, mitigate conflicts of interest, and align with international best practices.

And ITI said:

Dual Role and Governance Concerns: The BCB is the sole systemic regulator of Brazil’s financial sector, setting market entry conditions, operational standards, and pricing frameworks for all PAs. At the same time, it operates PIX, a Central Bank-run retail payment network, which is itself a PA and a direct competitor in the marketplace. In practice, U.S. EPS are subject to a serious unlevel playing field as they must compete against their own regulator. This contradicts international best practices and guidance from multilaterals, which call for a clear separation between oversight functions and commercial activities. Furthermore, PIX is not subject to regulation and supervision from a third-party regulator and overseer, or price or budget controls, in a way that would be materially equivalent to the kind of regulation and supervision that private players are subject to.

Anti-Competitive Distortions: Private players are required to create ecosystems where participation is voluntary, whereas PIX receives certain benefits. Examples of anti-competitive distortions include:

• Access to competitive information: The BCB has access to confidential and sensitive information from its private competitors (including pricing, product development, and commercialization plans), and also manages the development and operation of PIX. Without effective governance safeguards, this structure enables the BCB to shape both market dynamics and regulatory standards in ways that advantage its own platform.

• Mandated bank investments in PIX improvements, new features, and products – and guidance that banks should prioritize PIX – without equivalent obligations for private platforms.

• Regulation dictating priority placement of the PIX icon within bank apps, ensuring prominent exposure and enhanced user experience.

• Mandatory network integration with PIX that channels retailer and consumer behavior toward the government platform, constraining growth opportunities for U.S. companies in Brazil’s payments ecosystem.

• Not equal standards applicable to PIX as competitor: Private-sector payment providers are expected to deliver top level security at all times. This means being under constant oversight from the Central Bank and investing heavily to keep systems safe and resilient. PAs are also subject to technical standards, tax collection requirements, and supervisory costs that do not apply to PIX.

• Unequal treatment of card networks to initiate transactions on PIX: U.S. card networks are not authorized to use their credentials to initiate payments via PIX. As a measure of good faith, the BCB should promptly expand the concept of payment initiation in order to permit all card networks and digital wallets to initiate payments on PIX, and to do so in a manner that would not require ongoing access to or retention of customer and transaction-related data. Such a measure would complement—not replace—existing PIX offerings and support broader adoption.

Pix defenders

On the other side, groups defending Pix were Public Citizen/Data Privacy Brasil Research and – not surprisingly! – the Brazilian government.

Public Citizen and Data Privacy Brasil Research said:

Pix is the result of a collaboration between the Central Bank, which regulates and operates it, and other private sector stakeholders, which began in 2018 with the establishment of a working group on instant payments. The private sector has been particularly involved since the design phase of Pix, and currently, there are over 900 payment and financial institutions participating in the system. The creation and continuous development of Pix is supported by the Pix Forum, created in 2019, which is composed of various market players with the goal of dialoguing with and supporting the Central Bank in defining the operating rules for instant payment ecosystems. Thus, key regulations around Pix have been developed through open and consultative mechanisms involving private banks, payment system operators, fintech companies, etc. There are therefore numerous mechanisms for private sector operators to liaise with and suggest improvements to the regulatory ecosystem around Pix.

As with other Digital Public Infrastructure (DPI) systems, Pix attempts to “open up” the payments ecosystem, thereby enabling greater competition and innovation in the fintech sector. A number of smaller companies and startups are said to be developing new services and applications around the Pix ecosystem, creating an ecosystem of diversified product offerings. As noted by an International Monetary Fund (IMF) study, the use of Pix has enabled greater competition in the financial sector as it has “led to the growth of several payment services institutions, which have established banking subsidiaries, increasing competition for deposits with big banks. The open sharing of transaction information on Pix users has helped to strengthen competition in the sales of banking products and services, including for better and cheaper payment services, among various institutions.”

...

It is important to note that Pix is not a substitute for traditional digital payment mechanisms such as credit cards. Pix has not replaced credit services; instead, it has been widely used as a real-time payment tool. If anything, Pix has driven a move away from the use of cash, rather than traditional digital payment methods. Notably, despite the huge uptake in use of Pix, studies indicate that the use of credit card systems has also increased over the past 5 years. Data show that the credit card market expanded in Brazil (with an 11.6% increase in the number of transactions in credit in the first half of 2024 compared to 2023), reaching R$4.1 trillion in total transaction volume.

And the Brazilian government said:

In establishing itself as the entity responsible for defining Pix’s rules, the BCB—recognizing the need for neutrality—chose to develop Pix as a Digital Public Infrastructure. As such, it is not exclusionary by design. This open-access system is available to all eligible institutions/entities/persons as a means to encourage innovation and the development of new business models by market participants, thereby promoting financial inclusion. Foreign entities also have the opportunity to integrate it into their systems through regulated financial institutions. To this end, BCB also became the operator and manager of Pix’s technological infrastructure, consisting of the Instant Payment System (“SPI”)—the central settlement platform; and the Transactional Account Identifier Directory (“DICT”)—the centralized database of transactional accounts linked to Pix keys.

...

The development of Pix, therefore, is at the forefront of a global trend to which the United States’ own Federal Reserve is actively contributing. The fact that the instant payment infrastructure provided by the BCB makes available to the Brazilian public an additional service, alongside other electronic payment options offered by different providers, is remarkably analogous to the development of FedNow in the United States and of similar infrastructure in other jurisdictions.

Brazil’s policies and measures do not restrict the operations or undermine the competitiveness of U.S. companies engaged in electronic payment services. Brazil does not apply differential treatment to foreign payment providers, nor does it impose licensing or operational barriers specifically on U.S. providers.

There is no prohibition on digital platforms—such as WhatsApp, Facebook, or Instagram—offering their own digital payment services. However, like any other domestic or foreign digital payment service providers, they must obtain authorization from the BCB; comply with the LGPD; and adhere to relevant, non-discriminatory regulatory requirements.

...

To reiterate, there is no discrimination against U.S. digital payment service providers—whether in the form of additional or distinct requirements, or regulatory bias vis-à-vis domestic or third-country providers. All providers, regardless of origin, must meet the same criteria for the provision of digital payment services.

There are no specific restrictions, for example, preventing U.S. digital wallets from operating in Brazil, whether for consumer or merchant accounts. They are not required to use Pix and are free to process transactions through their own systems. They must, however, like all other digital wallets (domestic or foreign), integrate with BCB-authorized institutions and comply with Brazilian regulations. No additional or differential regulatory requirements are imposed on U.S. digital wallet providers compared to Brazilian or other foreign counterparts.

There is likewise no prohibition on private payment platforms such as the U.S.-based Zelle and Venmo operating in Brazil, provided they—like all other domestic or foreign private payment platforms—secure BCB authorization, establish a local commercial presence (or partnership), and comply with Brazilian regulations.

The evidence presented herein makes it abundantly clear that Pix does not discriminate against, and does not unfairly disadvantage, U.S. companies engaged in digital trade or electronic payment services. By incorporating millions of users in the market for digital payments—an achievement of financial inclusion that has been widely recognized internationally, including by U.S. private companies—Pix has leveraged the potential of the Brazilian market for all providers, including U.S.-based companies.

Questions from the U.S. government

Pix came up at the public hearing as the Section 301 Committee asked questions of the witnesses. Nicholas Voltaggio of the Department of the Treasury asked Neil Herrington of the US Chamber the following:

The first question is for Mr. Herrington. Your testimony summary and your written comments, as well as your testimony here today, address conflicts that arise from the Brazilian Central Bank's operation and regulation of the government run electronic payment system, or PIX.

Could you please expand on these concerns and how specifically they may affect U.S. firms?

Herrington replied:

Sure, I'll say – I appreciate the question. I'll say, first of all, I want to – I'll submit a written rebuttal response just to be very, very thorough.

But on the surface, I think, certainly, as I said in my testimony, we agree, and I think our members agree, that PIX has been beneficial for issues like financial inclusion and digital inclusion.

The real – the question becomes, what – when you work as both regulator and operator. And we believe that there's a lack of transparency and separation in the governance of – between the – being a regulator and being an operator in that sense.

And obviously, there are other Central Banks around the world that operate in that sphere, but they have, frankly, much more robust governance procedures. So again, it's the division between – we'd like to see the division between – ensure competitive playing field for private providers, ensuring that the peak system is divided between – that the responsibilities between operator and regulator are clearly defined and ensure transparency and fair competition.

Sarah Bonner of the U.S. Small Business Administration then asked Sean Murphy of ITI the following:

In your written comments, you state that private players in the digital payments landscape are required to create ecosystems where participation is voluntary.

Whereas, PIX receives benefits such as access to competitive information, mandated bank investments in PIX, and regulation dictating priority placement of the PIX icon.

Could you please elaborate on how, if at all, electronic payment service providers from the U.S. or elsewhere may be affected by these alleged benefits received by PIX?

Murphy replied: 

... Let me begin by saying that Information Technology Industry Council and our members have no issue with government managed payment systems when they compete fairly with private sector payment systems.

And as my colleague from the U.S. Chamber said a moment ago, in the case of Brazil, this is not the situation.

The Central Bank of Brazil both regulates the financial sector, including many of the measures you just referred to, which are in our long form comments, but also operates a state championed entity that is a competitor to the private sector companies.

... on digital devices where you would normally access and conduct financial transactions, PIX is given prominent places.

For example, if you were – an equivalent would be if you were to walk into a bricks and mortar store and a competing product that is supported by the government has prime placement in a store and you have to then go search to find other alternatives that are operated by commercial service providers, it is itself, a de facto discrimination, and an impediment to fair access. ...

Some post-hearing comments are here: US Chamber; ITI; Brazil.

The USTR Determination

USTR took all this in, and in its determination came out on the side of the critics, offering the following conclusions on the issue:

Brazil has unfairly disadvantaged U.S. companies engaged in competing electronic payment services, including by policies that favor its national champion Pix. The Brazilian central bank established the instant payment system Pix in November 2020. Pix connects financial and payment institutions (‘‘participating institutions’’) with individuals, firms, and government entities to provide instant or scheduled payments, cash withdrawals, payment invoices, and short-term borrowing, among other services. The Brazilian central bank’s dual role as regulator and owner/ operator of Pix creates a conflict of interest, in the absence of adequate procedural safeguards. The bank has acted as a regulator to disadvantage U.S. electronic payment services providers and preference Pix. For example, the central bank mandates the use of Pix by financial institutions with more than 500,000 accounts and requires that Pix be displayed on participating institutions’ main application screen with no less prominence than any other payment or transfer functionality. In addition, the central bank encourages use of Pix over other services by mandating that participating institutions (including institutions that it requires to participate in Pix) offer Pix for free to individuals and by capping the fee those institutions may charge businesses for Pix transactions.

The acts, policies, and practices of Brazil related to its preferential treatment of Pix are unfair and discriminatory. It is unfair to require competitors to provide advantages to Pix, such as availability, visibility, and fee caps, and Brazil discriminates against U.S. electronic payment services suppliers by providing those advantages only to Brazil’s national champion. The acts, policies, and practices of Brazil related to its preferential treatment of Pix are a burden or restriction on U.S. commerce by imposing costs on U.S. services providers and by forcing U.S. providers to promote their Brazilian competitor, without compensation.

Some other views

In a recent article, The Economist says not so fast on USTR's conclusions:

The Trump administration also complains that Brazil’s central bank both operates Pix and regulates it. The arrangement does raise legitimate questions about giving so much control over a payments system and the financial data it generates to a single institution. But those are concerns about concentration of power, not about discrimination against foreign firms. Governments build, own and regulate essential infrastructure routinely. There is nothing inherently discriminatory about applying the same model to payments, points out Monica de Bolle of the Peterson Institute for International Economics, a think-tank in Washington.

The second assumption—that Pix has harmed American payment companies—is also weak. It rests on a misunderstanding over why Pix was created, says Daniel Santos Kosinski, a professor of economics at the State University of Rio de Janeiro. Before Pix, existing services, including those offered by foreign firms, charged fees for electronic payments that poor Brazilians could not afford. Pix was built to change that. The central bank estimates that at least 70m people have entered the formal financial system since its launch.

Far from cannibalising other electronic payment methods, Pix has expanded the market. It has done so at the expense of cash and cheques, the use of which has plummeted. The number of cash withdrawals made every quarter has fallen by 46% since Pix was introduced ...

That does not mean incumbents face no pressure. Pix has changed the economics of payments. Bernardo Guimarães of Getulio Vargas Foundation, a university in Rio de Janeiro, says this may eventually squeeze the profits of Visa and Mastercard—but through lower fees, not fewer transactions. Businesses in Brazil typically pay around 2% of credit-card sales to a payment processor. A Pix payment costs next to nothing. The result is greater pressure on all payment processors, big and small, to justify the fees they charge.

The Trump administration’s deeper concern may be that Pix becomes a model for the rest of Latin America, eroding American card networks’ profits and their influence across the region. But that, too, is misguided. Pix was the product of Brazil’s particular circumstances: a powerful and trusted central bank that could require big banks to join the system, and a huge domestic market in which millions of people still relied on cash. It also took years to design and test. Few if any countries in Latin America have the same conditions.

The Pix trade dispute and sovereignty

It's clear that USTR has a good deal of discretion under Section 301 to find that foreign acts, policies, and practices are unfair trade practices on the basis of their impact on U.S. economic interests, and my sense is this discretion is broad enough to allow USTR to have reached its Pix determination. But should it have done so? How hard should the U.S. push in areas that U.S. trading partners will consider sensitive ones that implicate sovereignty?

In the pre-Trump trading system, some U.S. critics of that system – including those who later worked on trade policy under Donald Trump – worried about its infringements on sovereignty. However, when Trump was in charge, it sometimes seemed like these critics were looking to defend U.S. sovereignty but not particularly interested in foreign sovereignty. In the case of Pix, we are dealing with a policy/program of a central bank, which I would think involves a high degree of domestic political sensitivity. If sovereignty is a concern for people, payment systems set up by a central bank seem like an area where this principle should play a role in the decision-making about whether an unfair trade practice exists.

I can see the argument that the way Pix has been set up has some degree of negative economic impact on U.S. companies such as Visa and Mastercard. But if negative economic impact is the standard, a wide range of what most people would consider to be non-discriminatory laws, regulations, and policies could be the subject of trade disputes. The further we go beyond non-discrimination as the core principle underlying the trading system, the more we encroach on sovereignty and the more trade conflict we are likely to see.

Of course, if you believe that as the largest economic power, you can push others on their sensitive domestic policies while avoiding anyone pushing on yours, this is not necessarily a problem for you. But if, on the other hand, you are looking for principles that can sustain a durable trading system, this may not be the right approach.

During an interview with the CBC in the afternoon, host Rosemary Barton asked U.S. Trade Rep. Jamieson Greer the following question:

It's a better deal ... than yesterday, but it's not a better deal than when CUSMA itself was being respected by the United States, which is not the case now, Ambassador, as you know. And one of the things that Prime Minister [Carney] also raised was this issue of reliability. He points to the fact that other countries have made deals with your country. He said it looked like they were signed on pencil because they've changed overnight. So why would Canada sign a deal without even any guarantees that you would continue to respect the deal, given that you haven't respected CUSMA?

Greer started his reply by focusing on the Trump administration's recent bilateral deals:

So which deal is that that we haven't respected that we've signed over the past year? Like, I'm just not familiar with that. I mean, we have made a dozen deals ...

Barton jumped in to make it clear she had the USMCA in mind:

You've added tariffs. I'm talking about Canada. You've added tariffs, right, above and beyond the free trade agreement, and that sort of contradicts what a free trade agreement is, of course.

In response, Greer talked about the USMCA review, and then emphasized the agreement's security exception as the reason for any new tariffs that would otherwise violate the deal:

... USMCA includes specific exclusions for national security measures, and so steel, aluminum, autos – these are all national security measures for us that fall within the exclusions within USMCA. ...

This exchange jumped out at me (the whole interview is worth a watch), because earlier in the day I had been doing a webinar on the USMCA where I was asked a very similar question. Karthik Sankaran of the Quincy Institute asked me the following:

[Juan Carlos Baker Pineda] already mentioned you have all these different tariffs ... floating around. And you have this process where some things are exempt, but clearly other things are a violation of U.S. commitments under USMCA, and one of the lines in Mark Carney's kind of ... walkout mic drop, so to speak, was these agreements were signed with a pencil. What does that tell you about U.S. intentions? ...

My response was:

My sense of the Trump administration is it sees all of these statutes that Juan Carlos referred to – 301, 232, 338 – ... as taking priority over the USMCA and any other international trade obligations. So they will observe those agreements unless they decide not to and decide to impose tariffs under one of those statutes for whatever reason. And when they do decide to impose tariffs, they will often have a justification under the agreement. So they'll say, "Well, this is for national security. We're invoking the national security exception," which they've done many times now. ... So ... , in their view, or their stated view, [they] are technically complying with the agreement.

Or in the case of these Section 301 tariffs that are being imposed in response to concerns about forced labor, maybe they would say, "Well, here, these are justified under the public morals exception."

So, regardless of what their explanation is, and they will generally have one if pressed, they don't feel too constrained by these agreements. So, I think that's the written in pencil part right there.

So, maybe what we're seeing here is sort of the famous saying of "we're observing it in the breach." ...

Am I right about that? Given the scope of the Section 232 tariffs in Trump's second term, it certainly feels that way. So many products are being covered at this point that the connection to national security, which was weak to begin with, seems totally severed. And if that's the case, how reliable are any of these agreements? I feel like Greer's point at the outset may have been that, putting aside the older trade agreements, the Trump administration will comply with the tariff commitments in the new bilateral agreements it has signed during the second term. But how can trading partners rely on that? What is special about these agreements that distinguishes them from, say, the USMCA or the revised KORUS FTA, which were negotiated by Trump during his first term? What is stopping the Trump administration from imposing tariffs for security reasons in ways that conflict with the tariff commitments made in the recent bilateral agreements? These are questions that U.S. trading partners are probably wondering about, and I'm not sure what the answers are.

Americans pay far more for brand-name drugs than Europeans do. For years that was treated as a domestic health problem. The Trump administration has started treating it as a trade problem, and in June 2026 it opened a Section 301 investigation into Germany's drug pricing.

This post asks whether a German-style 301 investigation would work against France (another European country often cited as a problem by U.S. government officials and pharmaceutical companies). It proceeds in three parts. First, it explains the Section 301 investigation against Germany, specifically, what USTR is alleging and why. Second, it explains how France sets drug prices, since similar action against France would have to address the system in place there. Third, it looks at the debate inside France over how to respond to American pressure, which turns out to be less united than it first appears.

The German Section 301 case gets things rolling

Section 301 of the Trade Act of 1974 provides a set of procedures under which USTR can investigate a foreign country's trade practices. Among other things, actionable conduct under Section 301 includes acts, policies, and practices of a foreign country that are “unreasonable or discriminatory” and “burden or restrict U.S. commerce.” If USTR finds that such conduct exists, the U.S. government can take action, usually with tariffs.

The key evidence behind USTR’s investigation in the German pharmaceutical case is that U.S. consumers pay roughly 3.9 times what German consumers pay for brand-name drugs. USTR traces that gap to two specific German practices. The first is a rule that ties price confidentiality to a discount, meaning a manufacturer can keep its negotiated price secret only by accepting a discount. The second is a draft law that would add a mandatory rebate on patented medicines, which the industry expects to grow over time. USTR’s initiation notice suggests that these practices – one in effect now, and the other possibly coming into force soon – together push German prices below fair market value and leave Americans to fund the difference (through higher prices in America).

The investigation against Germany grew out of the May 2025 “Most-Favored-Nation” executive order, which set out to ensure that U.S. consumers should pay no more for a drug than the lowest price paid by any comparable developed country. The Trump administration is trying to achieve this goal by raising foreign prices through trade pressure (there have also been moves by the administration to lower U.S. prices, although it remains to be seen how successful they will be). The United Kingdom already agreed to pay more for new U.S. medicines in exchange for tariff relief. USTR urged Germany to follow suit, but after months of talks, Germany instead advanced legislation to further cut spending on innovative drugs, and the investigation followed.

Will France be next?     

Beyond Germany, France is another EU nation that has often been cited as a particular problem in relation to its drug pricing. In France, a single national insurer, the Sécurité sociale, covers nearly everyone. Because one buyer pays for almost every prescription (an arrangement that is used in many countries around the world), that buyer negotiates prices that apply to every purchase, and there is no free market price for prescription drugs.

Every individual drug proceeds through four steps before it has a price and enters the market:

• Authorization (AMM): The drug is approved for safety and effectiveness, usually through the European Medicines Agency. Price is not discussed at this stage.
• Grading (HAS): A public agency, the Haute Autorité de Santé, evaluates the drug through its Commission de la Transparence. It issues two ratings. The SMR decides whether the drug is useful enough to be reimbursed and sets the reimbursement rate; and the ASMR measures how much better the drug is than existing treatments, on a scale from I (major advance) to V (no improvement), and is the main driver of price.
• Price (CEPS): A government committee, the Comité économique des produits de santé, negotiates the actual price with the company. This is the step that matters most for what’s at issue with the current U.S. 301 investigation, for three reasons given below.
• Reimbursement rate (UNCAM): A separate body sets what percentage of the price the state pays back, and the health minister formally lists the drug.

The CEPS stage is worth focusing on, because it is where the price is negotiated between the state and the manufacturer, which is the same kind of government price-setting that the investigation against Germany targets. Three features of CEPS matter here:

• It is not independent. CEPS sits under the authority of the health, social security, and economy ministries. The state appoints its president and holds the majority of its seats.
• The company cannot simply walk away. If no agreement is reached, CEPS sets the price by its own decision. The relevant ministers can override that decision, jointly setting the price themselves within fifteen days, but the initial decision by CEPS does not require approval.
• The published price is not the real price. On top of the official price, the manufacturer pays confidential rebates (remises) back to the state, negotiated drug by drug. So the list price overstates what France actually pays.

The CEPS stage is likely to be the focus of a Section 301 investigation. The first two are clinical judgments, while the third is a budget decision. And the confidential rebates in step three are the same hidden discount USTR is attacking in Germany. This raises the question: does France show the same price gap that started the German case?

To check, we can turn to RAND Corporation's 2022 price comparison, which looks at U.S. prices compared to 33 OECD countries. Its figure for Germany, 3.87 times, lines up almost exactly with USTR's 3.9× claim, which makes it a reliable benchmark for measuring France in the same way. The RAND Corporation finds that U.S. prices for brand-name drugs are 445% of French prices and 387% of German prices. In other words, the U.S. pays about 4.45 times French prices versus 3.87 times German prices. RAND states plainly that France and Japan generally have the lowest prices for brand-name drugs among the 33 OECD countries in the analysis.

Source: RAND Corporation, International Prescription Drug Price Comparisons: Estimates Using 2022 Data (RRA788-3), Figure 3.2, p. 17.

The same pattern appears in the HHS/ASPE analysis of the same data: U.S. brand-name prices were 422% of the average across 33 OECD countries, whereas U.S. generic prices were actually lower, about 67% of other countries' prices. The price gap is therefore a brand-name phenomenon. On generics, which make up 90% of U.S. prescription volume, the U.S. pays less than its peers.

There are two caveats to keep in mind. All of these figures are gross, list-based prices. RAND could adjust the U.S. side down for rebates (bringing the brand-name figure from 422% to 308%), but it could not do this adjustment for countries where there are confidential rebates. RAND notes specifically that “German sickness funds receive statutory rebates not reflected in manufacturer sales.” France's remises work the same way. So the true gap that France has is actually smaller than 4.45 times by an unknown amount. However, the main claim holds: on brand-name drugs, the category USTR is targeting, France is a bigger outlier than Germany.

The French response

France is not united on the issue, with a strong internal divide emerging. In 2026, a rare coalition formed on a proposal to make the real, net prices of drugs public, instead of keeping the rebates secret. The left (La France insoumise, the Socialists, the Communists) and the National Rally, parties that agree on almost nothing, all supported it. In parliament, opposition came from the presidential majority, which voted against. The executive and the relevant ministries also warned that transparency would strip France of its main bargaining tool, prompting some criticism, such as from Médecins du Monde.

Other groups have entered the debate too:

• Industry (LEEM): The drugmakers' association, LEEM, wants the government to freeze price cuts and abolish the safeguard clause (a yearly cap on total drug spending: if national drug sales exceed the limit, manufacturers collectively pay back the overage). It warns that new tariffs on medicines would threaten patient access and drug supply. LEEM was also questioned directly by the Assemblée nationale on pricing mechanisms in March 2025.
• Critical press (Veltis/Le Figaro): The argument here is that if the U.S. forces lower prices at home, companies will raise prices in Europe to compensate, and France, with some of the lowest prices in Europe, is the most exposed. Veltis, a consultancy quoted by Le Figaro, put it bluntly: to protect their main market, companies “will have to raise the prices negotiated in Europe… or give up selling their new drugs there.”
• Reformist (Bizard): The economist Frédéric Bizard argues that France should treat the American pressure as a chance to rebuild a pricing model he considers broken, rather than a threat simply to resist. He also confirms the mechanism that connects the American policy to France directly: because the U.S. now targets the lowest price in Europe, and that price is often France's, companies such as Pfizer threaten not to launch new drugs in France at all.

Some tentative predictions

So, would a Section 301 investigation along the lines of the one underway against Germany work against France? Just looking at the numbers, if you apply USTR's analysis of the German policies to the French situation, the case is there. The price gap that justified the German investigation is not only present for France, it is larger. And the French system concentrates price-setting in a state committee, which provides evidence of a government practice of the type Section 301 is designed to address. At the same time, whether any such practices – in Germany, France, or elsewhere – are “unreasonable or discriminatory” and “burden or restrict U.S. commerce” is a question still open for debate (although the Trump administration's views on this are probably not in doubt at this point). 

A complicating factor here is the same one running through the whole story: France's low prices come partly through confidential rebates, the same secrecy USTR is attacking in Germany. That makes the gap real but hard to measure exactly. Whether the administration turns to France next may come down not to the strength of the case, but to the particular priorities of the Trump administration, along with various non-trade aspects of the U.S.-French relationship.

Despite intensive negotiations from August 19-21, 2026, Canada and the United States failed to reach an agreement that would have suspended the 50% tariffs the U.S. threatened a month ago on $20 billion worth of Canadian exports to the U.S. (The list includes hockey equipment, cement, alcoholic beverages, paper pulp, and many others.) Those tariffs, covering roughly 5% of total Canadian exports to the U.S., are significant in part because they are the first large group of USMCA-compliant goods that have been subject to prohibitively high (50%) U.S. tariffs, reflecting another flagrant U.S. violation of the USMCA. (New US Tariffs) (The previous U.S. tariffs on USMCA-compliant goods were on goods in specific sectors, pursuant to investigations under Section 232.)

It emerged Saturday morning, August 22, that several major new demands by the United States had contributed to Prime Minister Carney's decision to suspend the negotiations, even though it meant imposition by the U.S. of the 50% tariffs. According to Carney, "the U.S. introduced at the last hours efforts to restrict our ability to have other trade deals."  (Carney August 21 Speech) Canada in recent years has concluded various trade deals along with economic and security partnerships, including major free trade agreements with the EU, the UK, India, South Korea, and members of the Comprehensive and Progressive Agreement for Transpacific Partnership. A major element of Carney's economic policy since he took office in 2025 has been to seek to double Canada's non-U.S. trade from 20% to 40% of the total over the next decade, to be achieved in significant part by concluding new or expanded trading arrangements. (Trade Diversification) Given that Ambassador Greer, Commerce Secretary Lutnick, and other U.S. officials knew that trade diversification through an expanded network of trade agreements is a hallmark of Carney's economic policy, it is reasonable to ask whether this apparently last-minute demand was a deliberate (and successful) effort to scuttle the negotiations.

 Other ‘red lines” for Canada reportedly included a refusal to reduce tariffs on medium and heavy trucks, which was considered essential to preserve the Canadian automotive industry, and pressure to curtail the use of French in streaming services and French content requirements in Quebec and elsewhere. (Red lines) The U.S. reportedly also sought “exclusive access” to Canada’s critical minerals. (Critical mineral access) (Bloomberg reported that U.S. negotiators had been willing to eliminate a lumber tariff, a decades-long irritant in bilateral relations, but no details were provided.)

While I have no inside information, reliable public reports last week (before the alleged last-minute U.S. demands) suggested that Prime Minister Carney was willing to remove retaliatory tariffs on Canadian imports of U.S. automotive products, steel and aluminum; improve to at least a limited degree U.S. access to Canada's dairy market; and pressure the premiers of eight Canadian provinces to discontinue the boycott of American wine and distilled spirits in provincial liquor stores. (These were three of the major U.S. complaints that generated the July threat of tariffs.) (Settlement Proposals) Ultimately, what Carney was reluctant to accept, presumably for the longer-term, were lower but still prohibitively high tariffs on three mainstays of the Canadian industrial economy, steel and aluminum (50% to 25%) and autos (25% to 15% on the non-US content). According to some published reports the 25% steel tariff (but not the lower aluminum tariff) would only have applied to the first four million tons, with additional exports reverting to 50%. (Steel quota) Derivative steel and aluminum products were to be subject to varying rates.

It also seems evident Canada had concluded, based on their own and other countries' experience with the constantly varying levels of Trump's tariffs over the past 18 months, that any deal concluded at this time could never be regarded as final, as it might be changed in the future based on Trump's whims. As Carney suggested in his August 22 speech to the nation, U.S. trade agreements are “signed in pencil.” (Written in Pencil)

Implications

The failure to conclude the deal is a defeat for many stakeholders throughout North America, with the ongoing uncertainties continuing to retard both foreign and domestic investment and job creation now and for the foreseeable future. Some of the smaller Canadian producers of the goods subject to the new 50% tariffs, and their U.S. customers, may not survive. Canadian exporters and U.S. buyers of steel, aluminum and autos will continue to face higher prices. American wine producers will continue the emasculation of their largest export market. (US wine exports)

The tariffs on aluminum seem particularly unwise since the U.S. is not a leading producer due to prohibitively high electricity costs and local opposition to new based on adverse environmental impacts. (Oklahoma Smelter Opposition) Nor is it in the U.S. national interest to become more dependent on leading foreign producers Russia and China--where "national security" concerns are not bogus– for U.S. aluminum needs. (Aluminum shipments from the UAE may be welcome at 50% tariffs but have been disrupted by the Iran war.) Petroleum interests in Alberta, which for the first time in a decade saw the possibility of a revival of the XL pipeline project (an objective shared by Trump), are no doubt disappointed, as are refineries throughout the U.S.

The U.S. attempt to restrict Canada’s agreements with third countries, the latest escalation of the Trump/Greer trade war with Canada and potentially all other U.S. trading partners, is deeply troubling for reasons that go well beyond additional fracturing of what until January 2025 was the United States' premier political and military as well as economic relationship. Significantly, the United States is attacking trade agreements that are consistent with Article XXIV of the GATT, which has authorized limited discrimination under certain conditions in favor of FTA partners since 1947 (as with the USMCA among the U.S., Canada, and Mexico, which Trump championed). It is one affront to provide in USMCA Article 32.10 that Canada and Mexico may not conclude new trade agreements with non-market economies (e.g., China). These broader restrictions on third country trade agreements are a challenge to sovereignty that neither Canada nor most other U.S. trading partners are prepared to accept, even when many have already tentatively concluded one-sided trade "deals" with the United States.

This latest iteration of the trade “war” (Carney’s term) is far from over. Carney, with strong backing from l provincial leaders and many Canadian businesses, has promised “dollar for dollar” retaliation, which will almost certainly result in counterretaliation by the United States, as Ambassador Greer has promised. (U.S. Counterretaliation) Whether and when bilateral negotiations will resume is uncertain. One likely result of the impasse is that even more Canadian citizens will continue to boycott U.S. wine, distilled spirits and other products and avoid vacation travel to the United States.

Impact on Mexico

If anyone sees a silver lining to this cloud it could be Mexican President Sheinbaum. Had the deal with Canada gone forward, her cordial relationship with Trump compared to Carney's frostier one could have attracted broad criticism in Mexico, despite recent friction over Trump’s demand that the U.S. be permitted to send troops into Mexico to fight drug cartels. (US Troops to Mexico). Formal USMCA negotiations between the U.S. and Mexico that began months ago are scheduled to continue in September. (Next US-Mexico Talks)  At present, Mexico has little to show from them other than Trump and Greer's kind words.

 Moreover, Mexico is apparently still facing US demands for 50% U.S. content to qualify for (unspecified) reduced auto tariff access to the U.S. It also seems likely that the U.S. will be reluctant to offer Mexico a better deal than the 25%-25%-15% (steel, aluminum, autos) reduced tariffs offered to (and rejected by) Canada. Unlike Canada, Mexico is benefitting from greatly increased AI computer server exports to the U.S. which for the most part enter duty-free. (Mexico recently passed Taiwan to become the number one U.S. source.) (Mexican Server Exports) Unfortunately, with a Mexican value added estimated at only 5%-7%, the server boom helps the export numbers (with the increasing trade surplus a risk), but it will do little to stem unemployment in the automotive sector.

The new U.S. demand relating to Canada’s third-country trade agreements is also a wake-up call for Mexico, which currently has free trade agreements with the UK, the EU, the European Free Trade Association, Japan, the Transpacific Partnership members, the members of Latin America's Pacific Alliance, and dozens of other countries. (Mexico's free trade agreements) Those agreements are designed to be consistent with GATT Article XXIV or the 1979 GATT Enabling Clause (which permits FTAs among developing countries under more flexible rules than GATT article XXIV). (Enabling Clause) If Canada is being challenged because of its third-country trade agreements, Mexico has no assurance that it will not be next on the list.

David A. Gantz

Will Clayton Fellow for Trade and Int’l Economics,

Baker Institute for Public Policy

Peter offered some partial support for my view, in particular on my skepticism of the political importance of tariffs as revenue generators. Heather, whose original comments on this issue were the trigger for my post, pushed back a bit. In doing so, she elaborated on her points about the incremental value of tariff revenue in making the U.S. fiscal situation a bit more sane, as well as the constituencies that develop around specific tariffs.

Let me first note that, as their LinkedIn profiles make clear, both Heather and Peter have spent a good deal of time working in the executive branch, and I am quite sure their insider experience gives them insights that I don't have. As a result, I take seriously anything they say on these matters.

At the same time, I think that working on the inside can put you in a bubble to some degree, and sometimes the thoughts of a naive, idealistic outsider can be useful. I'm happy to play that role here!

Putting that theory into practice, as an outsider, it seems to me that there can sometimes be so much caution and care in political decision-making that we end up with something close to paralysis. If people agonize too much about how every interest group and faction will view a policy decision, they may end up with a bad decision or no decision at all. Sometimes it's better to trust your instincts and just go for it.

The practical risk here is that the agonizing could lead to an overly cautious approach to changing various Trump policies. There are a wide range of these policies that a future Democratic president would likely object to, including foreign policy interventions, insufficient protections for workers and minorities, immigration crackdowns, income taxes, spending choices, and, of course, tariffs. Our hypothetical Democratic president – it could be a Republican taking over, of course, but at this point let's just focus on how a Democrat would handle things because it is simpler in some ways – will have to look at each one, and make a decision on how quickly and thoroughly to reverse course. As part of this, a key question will be whether to shut things down immediately and develop a replacement later, or to keep things in place while a new policy is worked out.

I understand the desire for caution, but I can also imagine that if you don't do certain things right away, they may not get done at all. Four years can seem like a long time as you are entering office, but with everything you will have to deal with, it may feel like it goes by more quickly than you expected.

Getting back to the policy at issue here, in the case of the Trump tariffs a key point is the link between tariff revenue and budget deficits. It seems to me that a future president could make the case for revisiting both of the main components of that deficit (income tax and spending levels), and in that context explain that the tariffs haven't made much of a dent in the budget deficit (and in the meantime have led to problems in the domestic economy and in international relations).

Will there be certain constituencies that object to the removal of these tariffs? Sure, but as part of the broader policy review that would be taking place, I suspect that the people staffing this new administration can think of ways to keep these constituencies happy.

Anyway, let's have the midterms first, then come back to this issue next year!

During the past several years, the composition of Mexican exports to the United States has significantly changed. Auto and auto parts exports have decreased because of high Trump administration tariffs designed to shift production for the U.S. market from Mexico (and Canada) to the United States. (Trump tariff policy) Simultaneously, Mexican exports of the servers required to satisfy the apparently insatiable demand for AI data centers has skyrocketed. An estimated $650 billion investment in U.S. data centers is expected in 2026. (data center investment) Mexico now provides an estimated 40% of U.S. server imports, with sales reaching $46.9 billion for the first seven months, second to Taiwan, with sales of $53.5 billion. (Vietnam is third, with those three country suppliers largely replacing U.S. imports from China.) (US AI Server imports)

Servers and related hardware made up almost one-fifth of the $317 billion of goods Mexico exported to the United States between January and May 2026. This reflected a more than 100% increase compared to the same period a year earlier. (Mexican AI server exports)

Although specific import/export data is not available, it appears that at least some companies producing AI data servers in Mexico are exporting them to Canada as well as to the United States. Nvidia is reported as importing servers for its Canadian operations from Mexico under the USMCA. (Canadian server imports) Although volumes are not available, Nvidia recently announced that it had partnered with Ciara Technologies to assemble some AI servers in Montreal. (Nvidia Canadian server production) AI data center operators in Canada including Nvidia have greater sourcing flexibility since Imports of servers from third countries such as Taiwan, unlike into the United States, are generally duty free under the WTO’s Information Technology Agreement. (Zero Server Tariffs)

In the short and medium-term at least, server exports have thus helped to keep the Mexican economy from tanking. (Mexico's economic outlook) However, the server production phenomenon in Mexico raises major questions. Can the mostly Taiwanese new investment be sustained despite such concerns as Mexico's poor investment climate and the uncertainties that are a product of Mr. Trump's tariff policies, including the uncertain future of the United States-Mexico-Canada Agreement (USMCA)? Can server production help to replace lost automotive jobs? Will a significant percentage of the materials and components required for server production ever be sourced in Mexico, or will most of the value added continued to be imported from Taiwan, China, and the United States? Finally, what happens to the industry in the unlikely event that U.S. demand for new data centers diminishes because of electric power shortages or public opposition or, worse, if the AI craze turns out to be a "bubble" that bursts sooner rather than later? I seek to provide at least partial answers below.

Mexican Server Production and Exports

Development of Mexico as a major source of server production for U.S. AI data centers is a recent phenomenon. (Mexico: cornerstone of American AI boom) The servers, specialized computers used to store, manage, and process data inside AI data centers, are the major building blocks for data centers. “An AI data center is purpose-built … to handle AI training and inference.” AI workloads are estimated to require ten times more computational power than traditional applications. Three thousand AI centers are estimated to be under construction or planned globally by 2030, representing an estimated $7 trillion in total investment, with $3 trillion in the United States. For example, Texas currently has more than four hundred projects in the pipeline. (AI data centers)

Several considerations appear to have encouraged nearshoring of AI data server production to Mexico. Taiwanese enterprises, the principal source of Mexican server investment, were not surprisingly attracted by lower tariffs on exports to the United States, close physical proximity to the rapidly expanding U.S. market for data centers (and other high-tech electronic components), the desirability of diversification away from Taiwan, and lower labor costs for competent workers. (Taiwan nearshoring to Mexico) In particular, for most AI server exports to the United States, tariff treatment is favorable, with no specific regional value content requirement under USMCA rules or origin. (Server Tariff treatment) For Mexico, the disadvantage of this treatment is that there is less incentive for foreign producers to seek local components when imported components can be used without affecting duty-free entry.

Significantly, Taiwanese enterprises such as Foxconn decided to locate manufacturing facilities in Mexico in part because American tech giants including Amazon, Google, Microsoft, and Nvidia strongly encouraged Taiwanese enterprises to establish server production there. (Encouraging Server production in Mexico)

Among the logistical advantages, supplying necessary third-country parts and components to factories in Ciudad Juárez and other border cities is relatively straightforward: a container unloaded at Los Angeles or Long Beach can reach Mexico in less than two days. Shipping the heavy servers to data centers in most of the United States by truck is also manageable. Like many other Asian enterprises that have established manufacturing operations in Mexico, Taiwanese businesses do not seem overly worried with rule of law issues in Mexico although some concerns have been expressed. The advantages of co-production with Texas are also considered important. (Texas Co-production)  Other challenges, such as electric power and water shortages, could become more of a problem for new Taiwanese and other investments in the future.

Many of these same considerations have encouraged substantial direct foreign investment in Mexico for many years, despite concerns over crime, corruption, poor infrastructure, electricity shortages and a weak judiciary and regulatory uncertainty. (Mexico's Investment Climate)

Developing Mexican Supply Chains

A major challenge for Taiwanese server enterprises and for the Mexican economy itself is developing component production in Mexico so imports can be reduced and more local jobs created. The North American automotive industry, because of integration of Mexico with the United States and Canada, has done remarkably in producing parts and components by more than twelve thousand enterprises both large and small. (North American auto parts producers) However, historically, the electronics industry in Mexico has not done nearly as well. Consequently, flat screen TV producers (e.g., Visio in Baja California) and laptop producers (e.g., Lenovo in Nuevo León) must rely on a limited if slowly growing local supply chain. (electronics supply chain) Currently, as Christine Smith and Alan Murray for the Financial Times have noted,

Ciudad Juárez, across the Rio Grande from El Paso in Texas, is at the heart of the [AI server] industry, building off the border town’s existing Taiwanese factories that have been making simpler electronics for decades. The new production lines assemble parts shipped from Asia into AI servers, which are stacked in refrigerator-sized racks that can weigh more than two tonnes each and are then trucked across the border. (server assembly)

Taiwanese producer Pegatron is reported to have five assembly plants in Ciudad Juárez. It is arguably in everyone's interest to encourage relocation of Taiwanese and other Asian parts production for servers (and other electronic products) to Mexico, to lower costs and shorten supply chains, among other considerations.

Impact on Employment in Mexico

The Mexican automotive industry lost an estimated 320,000 jobs (out of almost two million), 2024-2025, due not only to tariffs but to production of EVs (using far fewer parts than gasoline-powered vehicles), increased automation and continued reliance on vulnerable Asian suppliers. (Employment decline)  By the end of the first quarter of 2026, exports of autos and auto parts to the United States declined by $4.87 billion, or by 11% compared to 2025. (Auto exports decline)

In my view, there is no reason to believe that major auto and auto parts producers for the U. S. market will fail to respond to the Trump administration's demands that they relocate at least some operations to the United States. Should USTR Jamieson Greer convince Mexico to accept a new 50% US content requirement for autos exported to the United States, Mexico's industry employment is expected to continue to decline, as Mexico currently accounts for 42% of all U.S. auto parts. (Fifty percent U.S. Content Requirement) Thus, a key factor is the extent to which it is feasible to shift and retraining auto workers for server production or other employment (which as noted below seems unlikely).

Unfortunately, reliable employment data are not available as of this writing. Fernando Alba, deputy economy minister in Chihuahua state, where Ciudad Juárez is located, is quoted in the Financial Times as observing, "One truck [of servers] is the equivalent of one thousand cars.” Unfortunately for the Mexican labor force, one thousand cars require far more workers than a truckload of servers produced in highly automated plants. Data provided by experts to the Financial Times’ reporters indicates that under current conditions that server production could realistically aspire to reach 3-7 per cent Mexican content, primarily regional direct labor, facility overhead, and localized logistics, compared to 39 per cent for autos. Nor does there appear to be any significant Mexican impact on product design. (Mexican content in servers) The director of the Claudio X. Gonzalez Center for the U.S. and Mexico, Tony Payan, recently shared with me his on-site observations that manufacturing in Ciudad Juárez began a decline before the beginning of the second Trump administration and is continuing despite expanding AI data server production.

Still, some auto industry employers might find employment in the data server industry. Workers with experience in assembly, production, quality control, maintenance, and supervisory roles could be transferrable, according to Luis Ricardo Rodriguez, managing director of Monarch Global Industries in Monterrey who counsels multiple foreign investors in Mexico. He notes that the capacity to absorb displaced automotive workers depends heavily on the type of operation. Server assembly, final-device manufacturing, and some back-end semiconductor activities are more labor-intensive, while semiconductor fabrication and other advanced manufacturing processes are significantly more capital- and skill-intensive. Many positions in data server production positions would require technical retraining, specialized skills, and, in some cases, English proficiency. As a result, these investments could absorb part of the displaced workforce, but probably not automatically or on a one-for-one basis.

Nevertheless, in time it seems reasonable to expect that higher Mexican component content could gradually be achieved, some of it much more labor-intensive than AI server assembly. A substantial share of the value incorporated into Mexican electronics exports is still generated abroad, with China, Taiwan, Korea, Malaysia, and Singapore remaining key suppliers, particularly of high-value components. But the domestic supplier ecosystem is gradually expanding, according to Rodriguez. Some multinational electronics companies already source a meaningful portion of their supply chain locally, in areas such as printed circuit boards, substrates, wiring, packaging materials, metal components, and other back-end manufacturing inputs.

As far as broader North American sourcing is concerned, it also seems reasonably possible that TSMC, a giant Taiwanese semiconductor producer with more than $265 billion of actual and planned investment in Arizona, will eventually be supplying some of the chips necessary for server production in Mexico from the United States, although currently most imports are from Asia. (https://www.taipeitimes.com/News/editorials/archives/2026/07/29/2003861544). Other American chip input to server assembly in Mexico could also be significant as eight of the world’s top chipmakers—Nividia, AMD, Intel, Alphabet, IBM, Meta, Broadcom and Qualcomm—are U.S. companies, although many of the chips they make are not currently produced in the United States. (U.S. computer chip producers) One estimate suggests that a single AI server rack contains at least 4,500 chips accounting for as much as 95% of the content value of a server ranging from inexpensive foundational chips to unique integrated circuits. (Chip value in servers)  (Chip varieties) Among other important considerations, sourcing more of the chips from the United States may be feasible if increased incentives for domestic production and for upstream materials are implemented. (U.S. Production Incentives) An added political and economic benefit from more U.S. sourcing would be reducing Mexico’s trade surplus with the United States.

Nor are U.S. policies regarding automotive industry imports likely to be extensively reversed in 2029, regardless of which political party gains the presidency. The U.S. auto unions, which have broad influence on many Republican as well as Democratic politicians in the United States, oppose Mexican imports in significant part because of much lower Mexican labor costs (e.g., at General Motors’ Mexican plants workers reportedly earn $25 per day compared to $18-$32 per hour in the United States. (Auto worker wages)  U.S. unions, which have opposed freer trade for decades, support higher tariffs on autos and auto parts imported from Mexico and Canada. (Union support for tariffs) Consequently, while server production can at least narrow or bridge the gap in the overall value of Mexican exports to the United States, it will probably not have a major impact on longer-term unemployment in the Mexican auto industry.

Risks for Mexico

The likelihood that the Trump administration would limit Mexican-sourced server imports seems low, given the importance of major AI enterprises such as Meta, Microsoft, Google, OpenAI and Nvidia to the administration and to the American economy. (AI providers' power)  However, it is not out of the question given U.S. concerns with the increasing trade deficit with Mexico and Mexican reliance on third-country inputs. U.S. actions detrimental to Mexican electronics exports, rational or not, could occur because of their continued reliance on parts and components imported from China and other Asian countries. (trade circumvention actions)

Another risk in my view is a decline in U.S. demand, either because of the bursting of an AI "bubble" warned against by some observers, or simply a decrease in demand in future years as new data centers are no longer being built. (AI Bubble) It is difficult to assess the chances of a bubble because of overbuilding, but overall, the installation of new AI data centers in the United States seems more likely to continue increasing than decreasing in the foreseeable future. It has been suggested that the AI facilities buildup “is the largest scale infrastructure build-out in the history of humanity,” compared in scope to the railroads in the 1800s and Roosevelt’s New Deal in the 1930s. AI investment in infrastructure is forecast to increase globally from $318 billion in 2025 to over $1 trillion in 2029. (AI Build-out)

Still, a temporary slowdown in U.S. AI data center construction could occur because of two factors. First, growing citizen opposition, reflecting NIMBY along with rising consumer electric bills and high center water usage, has led to moratoriums on new construction. New York was recently the first state to order a state-wide one-year construction moratorium. (Data Center Moratoriums) Secondly, it has also been predicted that data center power demand will increase from 31 GW in 2025 to 66 GW in 2027. (Data Center power demands)

Increasing AI server demand may also occur because of continuing upgrading of servers, for example to reduce electric power requirements or speed up computing operations. Thus, the situation is again quite different from the mature automotive industry, where US demand for new cars and small trucks is relatively stable at 15-17 million in recent years (rather than increasing), despite a small decline 2026 over 2025. (U.S. new car sales)

Any of these scenarios for future AI data server demand could have a negative impact on all players in the AI industry in the United States and on suppliers of the servers as well as the chips and other components. The impact on Mexico's exports could be material. The risk suggests that it is urgent for the Mexican government, working with Taiwanese investors, to take steps once again to try to develop a broader group of Mexican-based components suppliers, which presumably could supply electronic parts and components not only to server producers but to other producers of electronic products in Mexico. The fact that Taiwanese companies are investing in other electronic sectors, encouraging a "strategic axis in advanced manufacturing" between the two countries, is a positive factor for Mexico's economy. (https://mexicobusiness.news/trade-and-investment/news/mexico-taiwan-deepen-strategic-axis-advanced-manufacturing) This not only results in technology transfer and job creation, but provides an important alternative to mainland Chinese investment that is no longer politically feasible. (Mexico's China Problem)

Conclusions and Recommendations

The boom in AI server production and exports is not a solution to Mexico’s job losses because of Mr. Trump’s trade policies, even though it may replace much of the export revenue lost in the automotive sector. However, in my view it presents Mexico with a golden opportunity to move a large segment of Mexican manufacturing up the supply chain. As noted earlier, only 3-7% of the value of the AI data servers represents Mexican inputs; the rest are Taiwanese, Chinese, other Asian, or American. No one expects Mexico to become a manufacturer of chips in the foreseeable future; Mexico lacks the engineering talent and legal stability among other factors. Still, Taiwanese companies, working with the Mexican government, could in time raise the Mexican content of the servers, perhaps significantly if they believe the investment climate is sufficiently favorable. Mexico could also strongly encourage Taiwanese producers in Mexico to source their server chips from the United States rather than Taiwan or elsewhere in Asia, to the extent feasible. In this respect at least, American, and Mexican objectives are consistent: it is in the United States’ as well as Mexico’s interests to continue to encourage increased AI chip production in the United States, whether for use in AI servers or otherwise.

Whether the Sheinbaum administration has the willingness and the political ability to improve the business climate to the point where substantial additional direct foreign investment in component production, from Taiwan and elsewhere, occurs is problematic. In the view of some observers, President Sheinbaum’s predecessor, Andres Manuel Lopez Obrador, squandered many of the potential benefits of the nearshoring boom that began in 2017 or before by actively discouraging new direct foreign investment, particularly in the energy sectors. (Undermining Mexico's Favorable Investment Climate) After nearly two years as president, Sheinbaum in my view has mostly made things worse, replacing Mexico’s appointed judges with popularly elected ones, emasculating Mexico’s independent regulatory agencies, and accelerating Mexico’s movement toward becoming a one-party state, as it was before 2000. (Single-Party governance) Thus one can hope but not really expect the current Mexican government to be proactive in improving the investment climate, in this instance giving the AI server industry no viable alternative to importing most of the components from outside Mexico, as is the case today.