1. Introduction
Argentina–Financial Services Footnote 1 was the first financial services-related dispute to be brought to the WTO Appellate Body (AB) and the first in which the GATS carve-out for prudential regulation was invoked.Footnote 2 It is a rather distinctive case in that the dispute centred on measures taken by the Argentine government to combat tax evasion by nationals of Argentina. The Argentine economy has been characterized by recurring macroeconomic crises and weak economic governance. Following excessive spending and unsustainable fiscal deficits, Argentina defaulted on some US$91 billion in external debt in 2001, which led to a complicated and drawn out process of debt restructuring over the subsequent decade. During this period, the economy was effectively cut off from international capital markets. Argentina was only able to borrow internationally again in 2016. The balance of payments crisis motivated a variety of policies to control imports and retain foreign exchange, including formal and informal measures to limit foreign currency purchases for savings and investment abroad. The use of export taxes was expanded – rising to 15.5% of the value of total exports in 2012 (USTR, 2013). Stringent restrictions on imports were imposed, including extensive import licensing and policies requiring importers to balance what they brought into the country with an equivalent value of exports and foreign exchange.
The share of the public sector in total expenditures rose rapidly in the period pertinent to this dispute. General government primary spending increased from 26% to 39% of GDP between 2007 and 2015, with much of the increase going towards public sector wages, pensions, and subsidy programs. Government revenues also grew, from 28% to 34% of GDP during this period, with the tax burden rising to a quarter of GDP, one of the highest ratios in Latin America (IMF, 2016). The combination of a relatively large public sector, major fiscal deficits, lack of access to global capital markets and associated capital controls, inflationary financing of deficits, and efforts to raise tax revenues, all combined to create incentives for capital flight and tax evasion.Footnote 3 Policy efforts to control capital flight and tax evasion were not very successful. Most Argentine taxpayers evade taxes – 85% of respondents to surveys admit to evading some taxes, with half of respondents indicating they evade at least 20% of what is due. The collection rate for income tax due is below 50% (Bergman, Reference Bergman2009). This is despite substantial collection efforts by the tax administration – e.g., Argentina spends more than three times per unit of tax revenue collected than neighbouring Chile (Gómez-Sabaini and Jiménez, Reference Gómez-Sabaini and Jiménez2012).
In the dispute, Panama contested measures imposed by Argentina that penalized bilateral financial flows, justified on the basis that Panama was a ‘non-cooperating’ country because it did not exchange information with the Argentine tax administration. The measures comprised additional taxes on interest, dividends, and profits on transactions involving financial service suppliers based in Panama and restrictions on access to Argentine financial markets.
Panama is an offshore financial centre used by Argentine nationals, but is just one of many destinations for Argentine funds.Footnote 4 One estimate of the magnitude of funds held offshore by Argentine nationals in 2015 was US$210 billion, as compared to US$30 billion declared to the tax authorities. Much of these funds are deposited in high-income countries, including the United States, Luxembourg, Italy, and the UK. Together these four destinations account for over 80% of Argentine foreign portfolio assets (IMF, 2016).
The dispute arose during a period in which a concerted effort was being made at the global level to combat money laundering and tax evasion through international transparency and comity mechanisms. Both Argentina and Panama were part of this process and both referenced the international standards of tax transparency that had been agreed among participating countries. The dispute is therefore of more general salience for the WTO because it raises issues regarding the WTO/GATS compatibility of domestic measures if these are based on international standards. It is also of interest in shedding light on the scope that exists for invoking the general exceptions provision of the GATS (Art. XIV) as a justification for action against foreign services or service suppliers when this is deemed necessary for enforcing domestic regulation. Argentina invoked Art. XIV as a defence for its measures against Panama, arguing its measures were necessary to achieve a legitimate regulatory objective.
The Panel found the Argentine measures violated MFN and, because of this, could not be justified under GATS Art. XIV. It also ruled against a prudential regulation defence for some of the measures as it held they were not of a prudential nature. The AB overturned the Panel finding that the measures violated MFN, agreed with the Panel on its reasoning as regards Art. XIV, and rejected an appeal by Panama regarding the Panel's finding on the scope of measures falling under a prudential regulation defence. The upshot was that the Argentine measures were not found to violate its WTO commitments.
We will argue in what follows that the way the AB dealt with this case constitutes a missed opportunity to clarify three issues of central importance for the trading system: (i) determining the ‘likeness’ of services and service suppliers originating (based) in different jurisdictions; (ii) establishing the function of Art. XIV in matters pertaining to the enforcement of domestic law and regulations on national persons (natural or legal); and (iii) clarifying the role that internationally agreed good regulatory practices may play in justifying the use of services trade policies. Both the Panel and AB focused predominantly on the question of likeness and neglected the salience of Art. XIV in this case, given that the measures targeted Argentine tax nationals that sought to evade Argentine law, not foreign suppliers of services. An overemphasis on ‘likeness’ and the applicability of GATT case law confused rather than clarified matters, as the issues arising in this case were quite different from those that tend to characterize ‘standard’ GATT cases.
The plan of the article is as follows. In Section 2, we briefly discuss ongoing international cooperation efforts to establish agreed standards and criteria on good practices in the area of exchange of information on tax matters between jurisdictions. Section 3 summarizes the measures that were at issue in the dispute. Sections 4 and 5 discuss the Panel and AB findings. Section 6 steps back and provides a critique of the AB's reasoning and suggests an alternative approach that we believe would have better served the WTO membership. Section 7 concludes.
2. International standards on transparency and exchange of tax information
As the recommendations of the Global Forum on Transparency and Exchange of Information for Tax Purposes and the Financial Action Task Force (FATF) play a role in this dispute, we start with a brief discussion of these initiatives. The Global Forum was created in 2000 by the Organization for Economic Cooperation and Development (OECD) with a view to establishing high standards of transparency and exchange of information on tax matters. It complemented a contemporaneous Harmful Tax Competition initiative (OECD, 1998, 2000) that targeted the operation of tax havens and promoted exchange of information between tax administrations. The Global Forum was revitalized in 2009 following increased pressure at the G-20 level to do more to discipline tax havens (OECD, 2009), becoming the main multilateral forum where work on transparency and exchange of information takes place, including through the promulgation and implementation of voluntary standards (Hakelberg, Reference Hakelberg2016). It currently comprises 148 members. Two key standards adopted by the Forum concern the automatic exchange of information (AEOI) and the exchange of information on request (EOIR). Implementation of these standards is monitored through a public peer review process and periodic reports by the OECD secretariat (OECD, 2016).
Initially conceived as a global initiative to address money laundering, the FATF was created in 1989 by the G-7 (Freis, Reference Freis, Giovanoli and Devos2010). In a short period, the FATF adopted some forty Recommendations aimed at combating the misuse of financial systems by persons laundering drug money. The FATF subsequently updated its Recommendations to cover money laundering unrelated to drugs. In the aftermath of 9/11, FATF also adopted nine special Recommendations on Terrorist Financing. In 2003 and 2012, the FATF revised its 40+9 recommendations and adopted a set of international standards for anti-money laundering and countering the financing of terrorism (AML/CFT).Footnote 5 The recommendations constitute a non-binding instrument under international law but have been endorsed by over 180 countries. The FATF monitors the implementation of the recommendations, and members are assessed through a mutual evaluation (peer-review) process. The Financial Stability Board (FSB) recognizes the FATF recommendations as key components of sound financial systems in the field of institutional and market infrastructure.Footnote 6
The FATF has no legal personality. It is not an international organization. It has a temporary mandate, which is regularly renewed.Footnote 7 At the time of writing, the FATF comprises thirty-seven members, of which twenty-one are non-EU countries (including US, India, China, and Russia) or regional organizations. Not all EU member States are FATF members, but the European Commission, the EU's executive body and one of the leading forces in setting up the FATF, is a member. Argentina is a member of the FATF (it is also a G20 member), whereas Panama is not.
The FATF has a relatively rigorous process of identifying high-risk and non-cooperative jurisdictions. Assessments by the International Cooperation Review Group (ICRG) are used to identify jurisdictions that have strategic AML/CFT deficiencies. Panama was subject to the FATF monitoring process until February 2016, at which point it was deemed to have become compliant.Footnote 8 Thus, when WTO consultations were requested by Panama at the end of 2012 and at the time of the establishment of the Panel in 2013, Panama was still considered a high-risk jurisdiction by the FATF. Argentina ceased to be subject to the FATF's AML/CFT compliance process in October 2014, several months after the initiation of the dispute.
3. The measures at issue
Panama challenged eight measures (Table 1). Four relate to taxation (measures 1–4). The remaining measures concerned access by foreign suppliers to segments of the Argentine financial markets: the reinsurance sector (measure 5), the capital market (measure 6), registration of branches of foreign companies (measure 7), and access to foreign exchange (measure 8). To give a flavour of the tax-related policies, Measure 1 pertains to payments made to creditors in non-cooperative countries for credits, loans, or the placement of funds in a foreign location. If the country to which payments are being made is deemed cooperative, a capital gains tax of 35% is imposed on 43% of the amount transferred. If the country is not cooperative, the tax imposed is 35% of the total amount transferred – i.e., the total value of the financial flow is considered to be profit (a capital gain). Measure 2 pertains to financial inflows, with a presumption that inflows from non-cooperative countries constitute capital gains; Measure 3 presumes that transactions with non-cooperative countries are between related parties and are affected by transfer pricing manipulation; while Measure 4 addresses the allocation of expenditures with a view to prevent underreporting of profits. All four measures essentially have to do with provisions that increase the base for capital gains (profit or income) taxation for flows involving non-cooperative jurisdictions. The other four measures essentially raise the cost or prohibit access to the Argentine financial markets for firms from non-cooperative jurisdictions.
Table 1. Measures challenged by Panama
Source: Panel Report, Argentina–Financial Services, para. 2.9.
Panama argued before the Panel that, by classifying Panama as a non-cooperative country, all eight measures violated the MFN obligation (Article II GATS). For measures 2, 3, and 4, Panama also claimed a violation of national treatment (NT) commitments (Article XVII GATS). Argentina invoked the general exception provision of Article XIV GATS to defend all measures, except for its measures relating to the reinsurance market and access to the Argentine capital market, for which it invoked the prudential regulation carve-out (paragraph 2(a) of the GATS Financial Services Annex).
The key criterion in the application of the measures used by Argentina, defined in Decree No. 589/2013 (hereinafter ‘the Decree’), was whether entities providing financial services were based in ‘cooperative’ or ‘non-cooperative’ countries for tax transparency purposes. According to the Decree, a ‘cooperative’ status is granted to those jurisdictions that: (a) have concluded an agreement with Argentina on exchange of tax information or a double taxation avoidance agreement with an information exchange clause (provided that such exchange is effective); or (b) agree to negotiate such agreements.
Crucially, Argentina classified Panama as a cooperative economy as of January 2014, immediately following the establishment of the Panel in November 2013. This ‘upgrade’ in status occurred even though none of the conditions laid down in the Decree was met and no agreement for effective exchange of tax-related information had been agreed upon.Footnote 9 The result of this decision was to blur the distinction between cooperative and non-cooperative jurisdictions that is the key criterion in the Decree. It led the Panel to conclude that Argentina granted cooperative status to economies (such as Panama) with which no effective exchange of tax information occurred and thus that there was discrimination against services and service suppliers of non-cooperative countries.
4. The Panel findings
In assessing Panama's claim that all eight measures were inconsistent with the MFN rule, the Panel found that the difference in treatment between cooperative and non-cooperative countries is due to origin. In this respect, the Panel underlined the Argentine failure to prove that factors, other than origin, such as the exchange of tax information affected the competitive relationship between services and service suppliers of cooperative and non-cooperative countries (for instance, by showing how such factors influence the characteristics of the service and consumers’ preferences).
The Panel found that inconsistencies in the way that Argentina designated and treated different cooperative and non-cooperative countries made it impossible to evaluate potentially relevant ‘other factor(s)’ in addition to their origin that might impact on assessing the likeness of services and service suppliers from different home countries (p. 74). As a result, all eight measures were deemed to violate MFN (Article II:1 GATS) because services and service suppliers of non-cooperative countries do not benefit immediately and unconditionally from treatment no less favourable than that which services and service suppliers of cooperative countries enjoy – and that differences in treatment were not based on whether Argentina had access to tax information, because some countries treated as cooperative had not agreed on mechanisms to exchange tax information.
Furthermore, with respect to measures 2 (a rebuttable presumption of unjustified increase in wealth), 3 (an obligation to apply transaction valuation based on transfer pricing rules), and 4 (a rule on the allocation of expenditures (costs) for transactions between Argentine taxpayers and persons of non-cooperative countries), the Panel found that Argentine services and service suppliers are like services and service suppliers of non-cooperative countries. However, in the Panel's view, the Argentine measures were in line with the NT obligation set out in Article XVII GATS, as they did not provide for more favourable treatment for Argentine services and service suppliers relative to foreign ones. Rather, according to the Panel, these three measures were designed to ensure that the competitive relationship among services and service suppliers from cooperative and non-cooperative countries was on equal footing. By finding so, the Panel emphasized the importance of measures promoted by (consistent with) internationally agreed standards in addressing competitive distortions – neutralizing an unintended competitive advantage enjoyed by non-cooperative jurisdictions due to the lack of tax transparency (exchange of tax information).Footnote 10
Argentina sought to justify measures 1 (an irrefutable presumption of higher capital gains taxes on payments of interest or remuneration), 2, 3, 4, 7, and 8 by invoking the affirmative defence of Article XIV(c). The Panel noted that in order for Argentina to successfully meet the legal standard of this subparagraph, Argentina needed to show that the measures were designed to secure compliance with relevant Argentine laws and regulations that are not in themselves inconsistent with the GATS; and are necessary to secure such compliance. The Panel found that Argentina met its burden of proof required for each measure. By making transactions with a high risk of tax evasion subject to a higher tax base or subject to additional information requirements or closer scrutiny, the measures at issue were designed to achieve the objective of discouraging tax evasion/ensuring tax collection on all earnings obtained in Argentina.
However, it found that the measures failed to meet the strict standard of the chapeau of Article XIV GATS because Argentina appeared to grant cooperative status to countries with which it is negotiating a tax information exchange agreement without however effectively gaining access to tax information by those countries. Thus, because the application of these measures appears to defy the objective of the Argentine authorities to have effective access to tax information and blurs the distinction between cooperative and non-cooperative jurisdictions, the Panel found that the Argentine measures constituted arbitrary or unjustifiable discrimination within the meaning of the chapeau of Article XIV GATS.Footnote 11
The Panel went on to exercise judicial economy with respect to the Argentine claim under Article XIV(d), as it was conditional on a finding of inconsistency of measures 2, 3, and 4 with the NT obligation. As the Panel found no such inconsistency, it refrained from discussing the application and scope of Article XIV(d), which relates to the imposition and collection of direct taxes.
Finally, with respect to other GATS-related dimensions of the dispute, the Panel had to rule on the Argentine claim that measures 5 and 6 (relating to requirements for the supply of reinsurance services by suppliers of non-cooperative countries and requirements for stock market intermediaries when engaging in transactions ordered by persons of non-cooperative countries, respectively) can be justified by paragraph 2(a) of the GATS Financial Services Annex. After finding that the measures at stake were measures affecting the supply of financial services, the Panel examined whether these measures were taken for prudential reasons. In this inquiry, the Panel found that Argentina failed to demonstrate that a rational relationship existed between the objectives pursued and the measures adopted. Based on this finding, the Panel refrained from examining under the second sentence of paragraph 2(a) whether, by promulgating these measures, Argentina wanted to avoid its GATS commitments and obligations.
Panama also challenged measures 2 and 3, claiming that they are inconsistent with the MFN obligation of Article 1.1 GATT. The Panel found that Panama did not meet its burden of proof and thus rejected Panama's claim. The Panel also dismissed Panama's alternative claim that measure 3 is covered by Article XI GATT because it found that it was of a fiscal nature. Having rejected Panama's GATT-related claims, the Panel also denied to rule whether these measures are covered by Article XX(d) GATT.
5. The AB findings
The AB had to address essentially four claims brought by Argentina and Panama:
1. In regard to measures 1 to 8, whether the services and service suppliers from cooperative and non-cooperative countries are ‘like’ within the meaning of Article II:1 GATS (MFN).
2. With respect to measures 2–4, whether the Argentine services and service suppliers are ‘like’ services and service suppliers of non-cooperative countries within the meaning of Article XVII:1 GATS (brought by Argentina) and whether the Panel erred in finding that less favourable treatment (LFT) did not occur (raised by Panama).
3. Whether Argentine measures 1–4, 7, and 8 were designed and necessary to secure compliance with the relevant Argentine laws and regulations;
4. Whether measures 5 and 6 were covered by the prudential carve-out set out in paragraph 2(a) of the GATS Financial Services Annex and, more broadly, whether this carve-out covers all measures affecting the supply of financial services.
We discuss the AB reasoning regarding these four claims in what follows. In Section 6, we argue that the approach taken to the likeness question is misconceived and that the key issue here is item (c) under GATS Art. XIV.
5.1 Likeness
The AB did not examine separately likeness under the MFN obligation and likeness under the NT obligation. Confirming a degree of convergence in the determination of likeness within and across WTO agreements, the AB analysed the criteria and peculiarities of determining likeness of services and service suppliers at an abstract level that spanned both MFN and NT. The AB also emphasized the connecting line that transcends MFN and NT in both GATT and GATS, i.e., that they address discriminatory measures and centre on the competitive relationship among services and service suppliers or among products. This approach is sensible given the absence of import tariffs on services, which diminishes the salience of services product classifications and favours a criterion relating to competitive relationships.Footnote 12
The WTO case-law on likeness
It is established WTO case law that any criteria that could be used as proxies to establish likeness are neither treaty-mandated nor listed in an exhaustive manner in any one WTO legal provision.Footnote 13 Indeed, the GATT and GATS texts are silent on how likeness among two different products, services or service suppliers is to be established. Since the Border Tax Adjustments (BAT) Working Party Report,Footnote 14 we know that four general criteria can guide the determination of likeness: physical characteristics (properties, nature, and quality); end-uses (describing the possible functions of a product); consumer preferences (describing how consumers appreciate these functions); and tariff classification. We also know that the WTO adjudicating bodies, following GATT practice, give varying weight to the four criteria depending on the GATT provision involved. In Japan–Alcoholic Beverages, the AB noted that such discretion on the side of Panels is ‘unavoidable’.Footnote 15 In Argentina–Financial Services, the AB confirmed this:Footnote 16
As in the context of trade in goods, however, we equally consider that the criteria for analysing ‘likeness’ of services and service suppliers are simply analytical tools to assist in the task of examining the relevant evidence, and that they are neither a treaty-mandated nor a closed list of criteria that will determine the legal characterization of services and service suppliers as ‘like’.
Broadly speaking, we know from WTO case law on likeness that:
(a) There is no closed list of criteria but the four criteria noted above more often than not will give a strong indication whether Panels and the AB will judge two products to be ‘like’.
(b) The four criteria are interrelated and not mutually exclusive, so that evidence can be used for more than one criterion.Footnote 17
(c) A determination on a case-by-case basis is warranted.
(d) A market-specific analysis of the competitive relationship between and among products is required,Footnote 18 which can mean that two products that are like in a given market may not be like in another market.
(e) Regulatory concerns may play a role in determining likeness if they have an impact on the competitive relationship between and among the products concerned.
(f) A certain degree of convergence in interpreting the concept of likeness has occurred in that the four likeness criteria, along with the concept of competitive relationship, transcend the WTO agreements;Footnote 19 Such convergence is observed for both fiscal and non-fiscal measures falling under the various WTO agreements, linking the concept of ‘likeness’ and ‘directly competitive or substitutable products’ through the overarching construct of ‘competitive relationship’ (or the degree and extent of competition in the marketplace) or competitiveness (Hudec, Reference Hudec, Cottier and Mavroidis2000).
(g) A close relationship exists between likeness and LFT, with the AB clarifying that the latter informs the former, suggesting that likeness is about the nature and extent of a competitive relationship between and among products.
Likeness of services and service suppliers
Establishing likeness in the services realm is more challenging than it is for goods. For one, both MFN and NT require likeness of both services and service suppliers. There is no requirement for a determination of like producers in the GATT. Service suppliers are relevant in the GATS because trade in services includes establishment by service suppliers and because service suppliers often (usually) are the focus of regulation (as opposed to the specific services they provide). Physical characteristics are not helpful for services because of their intangibility, thus removing one of the GATT criteria for establishing likeness. Moreover, GATS makes no distinction between like and directly competitive or substitutable products, or between taxes and regulations, which implies that the consistency of measures affecting trade in services with NT commitments in determinations of likeness need to be examined in a flexible manner. This is even more so given that different modes of supply may be used to contest a market and that these modes often are treated differently in Members’ schedules of commitments under the GATS. The fact that NT is not an unconditional obligation in the GATS but comprises specific commitments through a positive listing or sectors combined with a negative listing of limitations in schedules of commitments further complicates matters (Delimatsis, Reference Delimatsis2007).
In Argentina–Financial Services, the AB appeared to take issue with the EC–Bananas III Panel finding that service suppliers providing like services are like. Instead, the AB followed the reasoning of the China–Electronic Payment Services PanelFootnote 20 that in certain circumstances a determination of likeness for both services and service suppliers is necessary before proceeding to discuss LFT. At the same time, the AB underlined that no separate findings with regard to the likeness of services and that of service suppliers is warranted: depending on the case, the relative weight of one may prevail over the other. The AB noted that this flexible approach is in line with the holistic analysis of likeness that it advocates.Footnote 21
While the AB suggested that criteria for likeness developed in GATT case law may be relevant as analytical tools, it limited its guidance on how these criteria should be interpreted in the services realm to just one paragraph of its report.Footnote 22 The AB started with the criterion of ‘characteristics’, which is reminiscent of the first likeness criterion under GATT (physical properties, nature, and quality). It also referred to the criterion of tariff classification suggesting that, under the GATS, the UN central product classification (CPC) system can play a role similar to the Harmonized System in GATT cases. Finally, it referred to the China–Electronic Payment Services Panel to note that services-specific considerations may require combinations of the different likeness criteria. After noting – correctly in our view – the potential relevance of modes of supply for the determination of likeness, the AB abruptly finished the discussion on criteria noting (with apparent relief) that it was not called upon to reflect on the relevance and weight of the likeness criteria in this case.
A final conceptual issue the AB dealt with under the rubric of likeness related to the so-called ‘presumption’ of likeness.Footnote 23 Established WTO case-law suggests that a complainant, instead of discussing the four BAT criteria to establish likeness, can alternatively establish that two products are like by demonstrating that the measure at issue distinguishes between two products based exclusively on origin. If so, likeness should be presumed to exist and WTO Panels can proceed to the analysis of LFT. The AB confirmed that such an approach is relevant for likeness in the case of services, but warned of the complexities that such transposition entails due to the characteristics of services mentioned earlier, which render more challenging the evaluation of an origin-based distinction. The AB avoided any further statement on the matter, noting that the nature, configuration, and operation of the measure affects the application of the presumption approach.
The AB noted that the complainant bears the burden of demonstrating that an origin-based distinction between services and service suppliers is drawn by the regulating State. The respondent may rebut by showing that origin is not the exclusive basis for the distinction drawn and/or that certain factors affect the four BAT criteria for establishing likeness and thus have an impact on the competitive relationship between the services and service suppliers at issue. If successful, WTO Panels are obliged to analyse all four likeness criteria before discussing whether LFT is granted to foreign services and service suppliers.
‘Due to origin’ or ‘based exclusively’ on origin?
A key question for the ultimate outcome of the case was whether the Panel erred by finding that the services and service suppliers at stake are like because the difference in treatment is due to origin. The AB found that the Panel should have continued its analysis because it was clear that it was not origin itself but the regulatory framework linked to such origin (cooperative vs. non-cooperative jurisdictions) that allowed for such distinctions and differences in treatment. The Panel had recognized and considered this but concluded that the implementation by Argentina of the criteria used to differentiate between cooperating and non-cooperating jurisdictions was inconsistent, if not incoherent. The AB clarified that it understands the presumption approach in a narrow sense to only cover origin-based distinctions. It seems that the measures where such a presumption would be relevant would typically encompass de jure discriminatory measures distinguishing among services and service suppliers of different origin. Thus, different treatment of services and service suppliers of non-cooperative jurisdictions in terms of tax purposes was regarded as too generic a category to benefit from the presumption.
The AB argued that the Panel failed to discuss the possibility that access to tax information is a decisive relevant factor that impacts on the competitive relationship between services and service suppliers of cooperative and non-cooperative countries. This was a crucial claim by Argentina. The AB held that the Panel should have evaluated the arguments and evidence brought forward by Argentina showing how access to tax information affects the competitive relationship, in particular the choices made by consumers.Footnote 24
Based on these considerations, the AB reversed the Panel's threshold finding that services and service suppliers of non-cooperative countries are like services and service suppliers of cooperative countries. Because this finding was critical for the Panel's findings on both MFN and NT, the AB found that the Panel erred in its likeness analysis for both MFN and NT under the GATS. The AB concluded its analysis on likeness by underlining that it takes no view as to whether services and service suppliers of non-cooperative countries are like services and service suppliers of cooperative countries or those originating in Argentina.
5.2 Less favourable treatment
The AB could have ended its analysis here, as all other Panel's findings were based on the finding relating to likeness and thus became moot and of no legal effect. Instead, the AB decided to continue its analysis and review the other findings made by the Panel.Footnote 25 While this was a welcome development in view of the fact that GATS disputes are a rarity, it is questionable whether the AB had the necessary authority to do so, as no request for completing the analysis was made by the parties to the dispute.
Following the logic of its analysis under likeness, the AB discussed the LFT standard enshrined in MFN and NT together. The AB confirmed that the role of the LFT standard under both provisions is to ensure equality of competitive opportunities in the marketplace, and thus involves assessing whether the measure at issue modifies the conditions of competition without any additional inquiry into the regulatory objectives or concerns underlying the measure at hand.Footnote 26 The AB concluded that non-exchange of tax information does not affect the determination of less favourable treatment, thereby reversing the Panel's previous finding in this regard. In the view of the AB, such regulatory aspects of concerns that could potentially justify discriminatory measures are more appropriately addressed by the GATS general exceptions provision.Footnote 27 At the same time, the AB did not exclude the possibility of taking into account evidence relating to regulatory aspects, as long as they have a bearing on the conditions of competition.
5.3 Designed and necessary to secure compliance
The third matter the AB was asked to consider, appealed by Panama, was whether measures 1–4, 7, and 8 found by the Panel to violate MFN and NT under the GATS were measures designed and necessary to secure compliance with GATS-consistent Argentine laws and regulations within the meaning of Article XIV(c) GATS. Surprisingly, Argentina did not appeal the Panel's finding that its measures were not justified by Article XIV(c) because Argentina failed to meet the non-discrimination requirement of the chapeau of Article XIV.
In discussing the first part of Panama's appeal, the AB referred to the by now well-established case law on measures to secure compliance under Article XX(d) GATT.Footnote 28 The AB reiterated that measures can be considered as securing compliance if their design reveals that they do so even if it cannot be guaranteed that the result aimed will be achieved. In this regard, Panels must assess the content and expected operation of the measures at issue as well as review their relationship with the specific rules, obligations, or requirements contained in the GATS-consistent laws and regulations as put forward by the respondent who bears the burden of the affirmative defence. Once found that the measure is designed to secure compliance, the necessity analysis follows based on the ‘weighing and balancing’ process identified in Korea–Beef by the AB.Footnote 29 The AB noted that the two steps of analysis are conceptually distinct but related aspects of the overall inquiry under Article XIV(c).
Regarding the design of the measures and their effect of securing compliance, the AB agreed with the Panel that the measures were designed to secure compliance and noted that this requirement would be met even if there were several inconsistent measures as long as those measures were used by the respondent to secure compliance with just one GATS-consistent law or regulation. Partial compliance would most likely also meet the requirements of Article XIV(c) (Cottier et al., Reference Cottier, Delimatsis, Diebold, Wolfrum, Stoll and Feinäugle2008).
With respect to the necessity of the measures at issue, the AB criticized the Panel because it made a collective weighing and balancing exercise for all measures at issue rather than each one of those measures individually. Furthermore, the AB took issue with the rather rudimentary analysis of the extent of the contribution of each one of the measures to the overall objective of protecting the Argentine tax collection system against the risks posed by competitive tax practices of countries classified as non-cooperative for tax transparency purposes.
However, the AB ultimately agreed with the Panel's findings that the Argentine measures met the standard set by Article XIV(c) GATS. Overall, the AB confirmed the broad scope of Article XIV(c) and the rather deferential approach to be adopted by the WTO judiciary when examining the legal requirements of the subparagraphs of Article XIV, echoing the relevant GATT case law under Article XX. A relevant issue in this regard would be whether WTO Members can use non-WTO law – in this case the FATF and/or the Global Forum standards – to justify discrimination or other measures that deviate from WTO commitments. Whereas in most cases there would be a transposition of such standards into the domestic legal order, in the type of case at hand the matter is more one of comity: cooperation by trading partners. Unfortunately, this matter was not addressed by the AB – we return to this question in Section 6.
Additionally, and in line with the overall deference and judicial restraint called for under the analysis of the subparagraphs of the general exceptions provisions, it is questionable whether it is really necessary to require a weighing and balancing, holistic exercise for each one of the measures at issue, in particular when many if not all of the measures form an inextricable part of an overall regulatory (here tax evasion control) policy. Finally, it is quite ironic and regrettable from a jurisprudential development viewpoint that, in a tax-related dispute, any discussion and analysis of Article XIV(d) is virtually absent.
5.4 The prudential carve-out
Argentina did not appeal the Panel's finding that measures 5 (requirements regarding access to reinsurance services) and 6 (requirements conditioning access to the Argentine capital market) were not covered by paragraph 2(a) of the Financial Services Annex (prudential carve-out).Footnote 30 Panama, however, appealed the Panel's finding that the prudential carve-out covers all types of measures affecting the supply of financial services and instead requested the AB to find that only domestic regulation measures of the type described in Article VI (typically non-discriminatory measures aimed at ensuring the quality of the service at issue) fall within the scope of the prudential carve-out. In Panama's view, the prudential carve-out could encompass inconsistencies with any provision of the GATS on condition that the prudential measure imposed was a ‘domestic regulation’, as opposed to, for example, a market access restriction.
The prudential carve-out is a key provision in the Financial Services Annex of the GATS. The legal discipline, entitled ‘Domestic Regulation’, reads:
Notwithstanding any other provisions of the Agreement, a Member shall not be prevented from taking measures for prudential reasons, including for the protection of investors, depositors, policy holders or persons to whom a fiduciary duty is owed by a financial service supplier, or to ensure the integrity and stability of the financial system. Where such measures do not conform with the provisions of the Agreement, they shall not be used as a means of avoiding the Member's commitments or obligations under the Agreement. (Emphasis added)
This is the first dispute in which a WTO member invoked the prudential carve-out. The disagreement between Argentina and Panama concerned the scope of this exception. The GATS, like trade agreements in general, has little to say on the substantive content of prudential rules or the policy rationales behind their enactment. While WTO Members agreed on the need to insert such an exception into the Financial Services Annex, discussions about the actual scope of the provision under the auspices of the Committee on Financial Services have been few (Marchetti, Reference Marchetti, Delimatsis and Herger2010).
Derogations from any obligation enshrined in the GATS, not just Article VI GATS, can be justified if the conditions laid down in Paragraph 2 are met.Footnote 31 The carve-out covers a potentially broad range of measures affecting the supply of financial services as long as it can be shown that they are adopted for prudential purposes. The initial burden of proof under the prudential carve-out lies with the complaining party, who must adduce evidence of a violation of a substantive GATS obligation. Once established, the burden of proof shifts to the respondent to establish the affirmative prudential defence.Footnote 32
Although the Argentina – Financial Services Panel and much of the literature consider the carve-out to be an exception (Leroux, Reference Leroux2002; von Bogdandy and Windsor, Reference Von Bogdandy, Windsor, Wolfrum, Stoll and Feinäugle2008; Marchetti, Reference Marchetti, Delimatsis and Herger2010), Cantore (Reference Cantore2018) argues that this misconceives the role of financial sector regulation. The main reason for negotiating the carve-out in the first place was not great concern that this was an area lending itself to abuse (discrimination; concession erosion), but rather serious worries by financial regulators that they might be constrained by a trade agreement in taking whatever actions they deemed necessary to ensure financial stability. This line of argument also implies that putting the burden of proof on the respondent is inappropriate – and very unlikely to be enforceable given the broad regulatory mandates and autonomy in many jurisdictions of financial sector regulators (e.g., Ministries of Finance; central banks).
Regulating authorities have a broad margin of manoeuvre in regulating the financial sector. Unlike other public policy exceptions enshrined in Article XIV GATS, all measures taken for prudential reasons are considered necessary. Members are not limited in any way regarding the prudential measures authorities may wish to use to protect the safety and soundness of the financial system, the integrity of financial markets, and the financial interests of investors and consumers. The language of the Annex differs from that in Article XIV GATS in not requiring that measures be ‘necessary’ to achieve the stated objectives. Nor is there language that measures not be applied in a manner that amounts to arbitrary or unjustifiable discrimination or a disguised trade restriction. The only constraint is that measures be taken for prudential reasons and not be used to circumvent GATS obligations or nullify scheduled commitments.
The analysis of both the Panel and the AB is in line with the discussion above and offers no grounds to justify the very narrow approach advocated by Panama with respect to the substantive scope of the prudential carve-out. The AB noted that the introductory sentence of paragraph 2(a) does not limit its application to a particular GATS provision but rather creates an exception vis-à-vis any other GATS provision. In addition, the second sentence provides support to such a broad interpretation by referring to the scope of the provision as covering measures which ‘do not conform with the provisions of the Agreement’. It is true that many prudential measures, notably those relating to the integrity and stability of the financial system, will be non-discriminatory and aimed at ensuring the quality of financial services, thereby falling under Article VI GATS. However, discriminatory measures of the type covered by the GATS market access, national treatment, and MFN provisions (Articles XVI, XVII, and II) could also be taken for prudential reasons, e.g., to protect domestic depositors or investors from fraud, or to prevent money laundering and financing of terrorist organizations. Nothing precludes the discriminatory application of prudential regulation measures.
Argentina justified measures 5 and 6 as being prudential in nature. A generous interpretation of prudential as spanning avoidance of money laundering and terrorism would conclude the measures fall under the carve-out. More salient however is to consider that these measures, as for the other measures, emphasized the tax regime prevailing in the partner country. These measures were part of the broader policy set used by Argentina to fight tax evasion by its nationals – and perhaps also to increase incentives for non-cooperating jurisdictions to cooperate.Footnote 33
More generally, it would be rather convoluted to have to invoke the prudential carve-out to take measures that have been agreed internationally pertaining to the transfer of illicit funds obtained from illegal trade in drugs, slavery or the financing of terrorism. As discussed below, a more straightforward approach would be to invoke Art. XIV to justify such regulatory measures and to clarify that the non-discrimination rule is not inconsistent with measures that use international standards to determine which jurisdictions they apply to. On the other hand, taking into account the absence of a necessity test and the implicit call for judicial restraint embedded in paragraph 2, one cannot deny the appeal of invoking the prudential carve-out in such circumstances.
6. Broader policy considerations and alternative approaches
To recap, the issue at hand is an effort by Argentina to combat tax evasion by its citizens. It is not a case that is about protection of domestic service suppliers. The problem is offshore jurisdictions that do not share information on financial holdings and operations of Argentine nationals. In order to reduce the incentives of its nationals to evade taxes, Argentina distinguished between cooperating and non-cooperating countries and imputed and collected higher taxes on transactions and flows between Argentina and non-cooperators. From an economic (indeed, common sense) perspective, the approach taken by Argentina is sensible: as a small country with limited market power, there is not much that Argentina by itself can do to change the behaviour of offshore financial centres.
In a situation where (some) countries play non-cooperatively (i.e., sovereignty rules), abstracting for the moment from WTO commitments and disciplines, Argentina does the best it can, given what Panama and other non-cooperating countries do. Conversely, Panama will determine its optimal strategy in light of what other financial centres do and what other high tax countries do. Imposing high penalties on suspect transactions – ideally, such that the expected return from attempts at evasion becomes smaller than simply paying taxes due and reporting financial assets held abroad – is a rational strategy.
Economic models of international ‘tax games’ suggest that some dimensions of the Argentine measures are what economic analysis suggest is called for in the prevailing situation, notably a high (penalty) tax on flows to non-cooperating countries. The economic literature has identified a potential trade-off between beneficial international economic integration and actions to protect against tax havens that results in too little action against tax havens (Johannesen, Reference Johannesen2012). Argentina's constrained access to international capital markets and scarcity of foreign exchange attenuate such trade-offs and provide further support for the measures from an economic perspective.
This takes as given the policy objectives of the Argentine government and the prevailing highly distorted incentive framework. In practice, there are many alternative policy sets that could mobilize greater tax revenues and higher rates of tax collection – including, for example, changes in tax bases, tax levels, improving governance, and reducing public corruption. Clearly, Argentine policy could be doing a better job at promoting national welfare (aggregate real income; growth). This raises the possibility that the existence of tax havens could be a ‘force for good’ in helping to induce the government to take actions to improve economic governance and domestic policies.Footnote 34 Of course, such considerations are not pertinent to the WTO dispute.
Assuming for purposes of discussion that Argentine policy is justifiable on economic grounds, the main question is whether the WTO constrains Argentina's ability to impose penalty measures on its citizens that transact with non-cooperating offshore centres to combat tax evasion. Clearly, this is detrimental to the offshore financial centres. But equally clearly, the policy is not designed and does not have the effect of benefitting domestic service suppliers. A tax compliant citizen can do business with cooperating offshore centres and there are many of those. The measures target taxpayers and seek to enforce domestic regulation and law to Argentine nationals. The problem was that Argentina could not demonstrate in a compelling and unambiguous fashion that it did not discriminate in its treatment of cooperative as opposed to non-cooperative jurisdictions, affording some non-cooperative ones effective cooperative treatment. This was an unnecessary own goal. If it had implemented its decree consistently in distinguishing between cooperative and non-cooperative jurisdictions, the Art. XIV defence presumably would have been effective.
The services that were targeted are ‘plain vanilla’ financial instruments and payments and income associated with them: interests, dividends, profits, capital gains, equity capital, lending, and underwriting risks. In this regard, the services from different foreign jurisdictions are clearly ‘like’: what country A suppliers offer will be very close to that of country B suppliers – the offshore financial services markets are competitive. The difference of course is that from the perspective of an Argentine tax evader, a jurisdiction with strong secrecy rules will not be a close substitute for an offshore centre that shares information on account holders with Argentine tax authorities.
This brings us to the question of ‘likeness’. Are services from non-cooperative jurisdictions unlike those of cooperative ones because of the absence of cooperation? Again, from the perspective of an Argentine tax evader, the answer is clear. But this is not because of anything specific as regards the substance of the financial products involved. Nor is it because regulatory regimes and tax rates differ across countries. What was at issue was an absence of comity and cooperation between governments. Argentina was requesting other governments to help it to enforce its tax laws. In Argentina–Financial Services, the AB missed an opportunity to clarify these matters.
The intangible and often non-storable nature of services, the resulting need to consider four different modes of supply and the fact that GATS concessions are for the most part regulatory commitments clarifying how services trade is to be conducted and regulated (including supplier-related characteristics) are essential features that should discourage any automatic transposition of the GATT case-law relating to likeness to GATS disputes. Instead, in our view these characteristics call for a narrow test, as the WTO judiciary should avoid second-guessing national regulatory preferences – a point on which the AB clearly agrees based on its case law (Howse, Reference Howse2016).
The very nature of services reduces the value of criteria such as physical characteristics or product classification. The former can at best only be applied when examining the characteristics of service suppliers. The low level of disaggregation of services statistics (the UN Central Product Classification) prohibits the detailed type of examination of products that is possible under the Harmonized System for goods. Thus, end-uses and consumer preferences are the most relevant criteria for assessing the likeness of services and service suppliers (Mattoo, Reference Mattoo1997).
In our view, as a rule of thumb, regulatory concerns such as access to tax information (or lack thereof) and combatting tax evasion should not play a role in determining likeness but be assessed under the general exceptions provision of Article XIV. In the case at hand, unlike the AB, we would agree with the Panel's ultimate finding of likeness. However, rather than misinterpreting the presumption approach regarding likeness, the Panel's analysis should have focused on end-uses and consumer preferences. It is an open question what the result of such an approach would have been.
As noted above, from a narrow perspective that focuses on the technical dimensions of the financial services concerned, similar consumer needs (investments, etc.) are satisfied. The service suppliers describe their business in very similar terms and compete with each other in the same business sector.Footnote 35 Clearly, consumer preferences towards tax evasion will differ, but a refusal by an offshore centre to cooperate with other jurisdictions for tax purposes is exogenous to the services and the service suppliers in question, as are the concerns and goals of the regulating State. An assessment of likeness should not encompass these considerations. Going further and considering the effects of regulatory regimes per se would make matters even worse, opening the door to governments to define for themselves what aspects of foreign policies they deem to influence the ‘substance’ of a specific service product or supplier. This is a slippery slope.
In this respect, we disagree with the AB's finding that factors that affect the competitive relationship such as effective access to tax information should be discussed under likeness insofar as such factors do not relate strictly to the services or service suppliers in question. Any attempt to identify overly detailed supplier-related criteria such as the legal form, the skills and certifications or the commercial characteristics of the service supplier would lead to a situation whereby no set of two service suppliers could ever be deemed like.Footnote 36 It would be more consistent with the overall logic of the GATS and more efficient from an economic viewpoint if regulatory concerns of the type at issue in this dispute were to be considered under Article XIV GATS, leaving consideration of likeness to cases where the matter at hand concerns actions that change the competitive landscape in a manner so as to afford protection to domestic services and service suppliers. This suggestion applies even more to cases relating to taxes, as Article XIV specifically refers to justifying measures aimed at the equitable imposition of taxes (Cottier et al., Reference Cottier, Delimatsis, Diebold, Wolfrum, Stoll and Feinäugle2008).
Another option could be to discuss regulatory concerns after the likeness test, under the LFT analysis. Note that, unlike the GATT, the GATS explicitly incorporates the LFT standard in both MFN and NT. In such a case, no less favourable treatment of like services and service suppliers would occur if it does not modify the conditions of competition in favour of certain foreign services or service suppliers (with respect to MFN) or in favour of domestic services and service suppliers (NT). A problem with this option is that taking into account regulatory concerns in order to discriminate will modify the conditions of competition and thus fail to meet the proviso of Articles II and XVII GATS. In addition, it may lead to absurd results whereby a violation of MFN is found but no violation of NT is established, as in the Argentina–Financial Services Panel Report.
All this suggests that a narrow likeness test and by implication a narrow LFT standard, coupled with a discussion of regulatory concerns under Article XIV, might support greater convergence around the concept of competitive relationship under the GATS. This does not imply an excessive burden on the respondent or a re-balancing of rights and obligations under the DSU. Even in the approach ultimately adopted by the AB in this dispute, under the substantive obligation of Articles II and XVII GATS the respondent would still need to rebut based on regulatory concerns. In any event, requesting this type of input from the regulating State makes sense as the respondent, more than anyone participating in the dispute, has the necessary information to motivate its regulatory concerns and objectives.
The foregoing discussion assumes a situation where there is no international cooperation on the regulatory matter at issue. It presumes a government unilaterally decides the criteria it will use to decide which countries are cooperative. This is clearly problematical, both because the criteria used may be inappropriate or inefficient, and because of the extraterritorial dimension and unilateral nature of the process. While such concerns are limited in the case at hand, as the target of the measures were nationals of the respondent state and the issue was not the substance of foreign countries’ regulatory and tax regimes but (lack of) exchange of information, there are nonetheless significant issues relating to privacy, data security, due process, and human rights, to name just a few with a unilateral determination of what constitutes cooperation. The international effort to define good practices in sharing of financial information and associated standards are a possible means of addressing this problem.
Standards resulting from international cooperation and agreements regarding good regulatory practices require underlying processes that are inclusive and transparent in order to be legitimate. The extent to which this was the case with the Global Forum process is somewhat of an open question – the fact that most tax havens participated may have more to do with concerns about the outside option or threat point than a reflection of their true preferences (Hakelberg, Reference Hakelberg2016). Whatever the case may be, the extant international standards, the mutual peer review process, and certification of jurisdictions provide a much stronger basis for differentiating between countries than does a unilateral determination. In our view, the existence of relevant international standards that have been adopted by a critical mass of nations could serve as the basis for legitimate discrimination against non-compliant (non-cooperative) countries.
The GATS does not have the equivalent of a Technical Barriers to Trade (TBT) agreement with its presumption that use of international standards implies WTO consistency of the measure. Hoekman and Mavroidis (Reference Hoekman, Mavroidis, Sauvé and Roy2016) argue that the absence of TBT type norms in the GATS is due in part to a dominance of market access concerns and differences in view regarding the application of a ‘necessity test’ under GATS Art. VI:4 but also to a relative absence of international standardization of services, reducing the utility of a TBT-like reliance on international standards as a way of satisfying the national treatment principle. The case at hand illustrates that there may be value in greater effort by the WTO Membership to utilize international standards where these emerge. A clear signal by the AB that this will be considered in cases where a WTO invokes Art. XIV to take action to achieve legitimate domestic regulatory objectives could increase incentives to participate in such international cooperation. More generally, explicit adoption of international standards by WTO members into their GATS schedules of commitments, using GATS Art. XVIII (‘additional commitments’) could help move the GATS incrementally towards the GATT acquis.
7. Conclusion
Argentina–Financial Services raises numerous interesting issues. To some extent, the case is sui generis as there were strong incentives for Argentine citizens to get their money out of the country given the macroeconomic environment, absence of growth prospects, populist politics, and endemic corruption. From an economic perspective, any solution to the Argentine problems in collecting tax revenue extends well beyond the measures involved in the dispute – it must encompass far-reaching improvements in governance to re-establish trust in the public sector and structural reforms to improve the investment climate and reduce the extent to which incentives are distorted. Weaknesses in governance and public administration may have played a direct role in the case at hand – reflected in the inability of the government to distinguish in a consistent manner between different offshore financial centres, its treatment of cooperative as opposed to non-cooperative jurisdictions, and the non-appeal of the Panel finding regarding its Art. XIV defence. Argentina did itself no service in the way it implemented its measures and defended itself.
Abstracting from these country-specific idiosyncrasies, the dispute raises issues that are important for the trading system. We have argued that the AB missed an important opportunity in this case – although this is in part a function of the non-appeal by Argentina of the Panel finding regarding its Art. XIV defence. We will not repeat ourselves here but conclude with a general remark. This case illustrates a basic tension in trade agreements, not limited to the GATS, in dealing with regulation. Trade agreements and trade negotiators are focused on market access, not on achieving regulatory objectives more efficiently or in supporting regulatory cooperation. This market access mind set influences the approach that is taken towards efforts to discipline the use of measures that may restrict trade in services but are necessary to achieve domestic regulatory objectives – including in this case enforcement of national legislation that applies to national persons only. The application of a GATT-inspired ‘likeness’ framework in cases that centre on the enforcement of domestic regulation, and the willingness to consider arguments that regulation – which inherently plays a significant role in the operation of many services markets – may influence determinations of whether services are like illustrates the problem.
How to deal with the trade effects of domestic regulatory measures is nothing new. The challenge of balancing trade vs. regulatory goals will only become more salient as WTO Members become more services-intensive economies over time. In the specific dispute considered in this article, the matter was rather straightforward and further facilitated by the existence of international standards that permitted an objective delineation to be made between countries on the basis of whether they had been certified as being in compliance with the standards. It is unfortunate that the opportunity was not taken to use this dispute to begin to develop a conceptual framework that can help guide WTO Members in dealing with the tensions that may arise between trade integration and domestic regulation.
Open access
