Inclusion of aviation in EU ETS

Legal assistance paper

All reasonable efforts have been made to ensure the accuracy of this information at the time the advice was produced. However, the materials have been prepared for informational purposes only and may have been superseded by more recent developments. They do not constitute formal legal advice or create a lawyer- client relationship. To the extent permitted any liability is excluded. Those consulting the database may wish to contact LRI for clarifications and an updated analysis.

Print Friendly, PDF & Email

Date produced: 17/09/2011

1. Is the EU Emissions Trading Scheme (ETS) implemented by EU Directive 2003/87/EC, as amended by Commission Decision 2011/149/EU (incorporating the aviation sector within the ETS), compatible with the WTO regime? Could the amended directive be classified as an environmental exception (as defined in WTO texts)?

2. Is it possible to class aviation as a ‘service sector’, rather than a ‘trade sector’ (as defined by the WTO) and would this then exclude it from the terms of the WTO agreement?

3. Are there any other international agreements or legal instruments which may be affected by the amended directive or which could be used to make a legal challenge to the amended directive?

1. Introduction
1.1. The EU’s scheme
The aviation industry will be included in the EU’s emissions trading scheme (ETS) from 1 January 2012. Airlines will have to acquire and ‘surrender’ allowances for the carbon emissions produced by their flights. The scheme is comprehensive: it applies to EU and non-EU airlines (subject to a potential exemption), to passenger and cargo flights, and to flights between EU airports and between EU and non-EU airports. An airline that fails to surrender allowances is fined €100 per allowance and must make up the shortfall the following year.
The EU ETS – of which the aviation scheme is a part – is a ‘cap and trade’ scheme for reducing emissions of carbon dioxide (and some other gases). These schemes set a ‘cap’ on total overall emissions by establishing a fixed number of emissions ‘allowances’, distribute these to industries according to a given benchmark, and permit industries to trade these allowances according to their needs. In the case of the EU ETS, the allowances are distributed initially by a combination of free allocation and auction. The EU ETS also envisages agreements for the mutual recognition of allowances issued by other countries participating in the Kyoto Protocol system.
For aviation operators, the EU has created emissions allowances corresponding to 97 percent of a benchmark calculated as the industry’s average carbon emissions during the three years 2004-2006. Of these allowances, 85 percent in 2012 are allocated for free (according to the airlines’ respective 2010 market shares), and the remaining 15 percent are available for purchase by auction. From 2013, when the so called ETS Phase III commences, the total quantity of allowances drops to 95 percent of the benchmark, and 3 percent of this new total will be reserved for ‘new entrants’ and rapidly growing airlines. Airlines are also able to purchase a certain number of additional allowances from other industries covered by the EU ETS but this is not reciprocal: operators of stationary installations are not permitted to purchase allowances issued to airlines.
The full economic implications of the scheme for the air transport industry are uncertain. In the first place, new allowances will need to be purchased to cover any increase in emissions due to industry growth since 2004-6 or an airline’s own increase in market share since 2010. On the other hand, it is theoretically possible that the scheme will come at no cost, if airlines manage to stay within their free allowance by developing greater fuel efficiency. It is relevant in this regard that biofuel emissions are not counted for purposes of the scheme. The result is that estimates of the costs vary wildly. Some claim the costs are likely to be small. On the other hand, US airlines claim that costs for them could run to more than US$3.1 billion over the next eight years. Regardless of the sums involved, there is at least a risk that the scheme will produce an initial cost for airlines, some (if not all) of which will be passed on to passengers and freight operators. Indeed, if it does not, it will have failed in its purpose.

1.2. The ICAO dimension
The EU’s scheme has for some years been the subject of fractious debate within the International Civil Aviation Organization (ICAO). In 2007, the ICAO Assembly adopted Resolution A36-22 which ‘[u]rged Contracting States not to implement an emissions trading system on other Contracting States’ aircraft operators except on the basis of mutual agreement between those States’. However, in a 2010 Resolution A37-19 the Assembly recognized that ‘some States may take more ambitious actions prior to 2020, which may offset an increase in emissions from the growth of air transport in developing States’. It also implicitly endorsed unilateral measures, ‘[u]rg[ing] States to respect the guiding principles listed in the Annex, when designing new and implementing existing MBMs [market-based-measures] for international aviation’, even as it also urged them ‘to engage in constructive bilateral and/or multilateral consultations and negotiations with other States to reach an agreement’.
The ICAO heralded Resolution A37-19 as a ‘historic breakthrough’. However, it has not been uncontested. A number of ICAO Contracting States lodged reservations expressly denying that unilateral measures were permitted. Perhaps most belligerently, the Russian Federation warned that it ‘does not rule out the introduction of adequate retaliatory measures by other Contracting States in respect of the operators of Contracting States which introduce market-based measures unilaterally.’
Furthermore, even to the extent that Resolution A37-19 can be said to endorse unilateral measures, it is not clear on the question whether unilateral measures may be applied to non-national airlines. Obviously those countries that do not accept the premise deny that this is possible. But the lack of clarity was evidently sufficiently uncertain to prompt the EU and 44 European states to lodge a reservation setting out their view:
The Chicago Convention contains no provision which might be construed as imposing upon the Contracting Parties the obligation to obtain the consent of other Contracting Parties before applying … market-based measures … to operators of other States in respect of air services to, from or within their territory.
The very fact that the EU and these other states felt it necessary to stress this point in a reservation, of all things, indicates that the issue is not as straightforward as one might otherwise be led to believe. And this is also supported by the fact that, on 2 November 2011, the 36 member ICAO Council – by a vote of 16 to 8 (all EU Member States) and with 2 abstentions – endorsed a working paper presented by 26 ICAO members which ‘urge[d] the EU and its Member States to refrain from including flights by non-EU carriers to/from an airport in the territory of an EU Member State in its emissions trading system’. It is reported that the next step might be dispute settlement against the EU Member States under Article 84 of the ICAO Convention.

1.3. Challenges to the EU’s scheme
These disagreements have spilled over into diplomatic and legal confrontation outside of the ICAO as well. China, India and the US have now publicly opposed the EU’s scheme; and China has allegedly followed this up by freezing US$4bn worth of orders from Airbus. Domestically, there is also discontent. On 24 October 2011, the US House of Representatives voted in favour of a bipartisan bill banning US airlines from participating in the scheme. This legislation now awaits approval in the US Senate.
The EU’s scheme has also provoked litigation. In 2010, a consortium of US airlines, supported by the International Air Transport Association (IATA) and the National Airlines Council of Canada, initiated a legal action in which they argued that the EU violated its obligations under customary international law and various international agreements, including the Chicago Convention. On 6 October 2011, Advocate General Kokott delivered an Opinion, finding that only certain provisions in the Open Skies Agreement had direct effect in EU law, such that the applicants could rely on it. But she also found, in the alternative, that the EU’s scheme would not violate these obligations in any case. It is difficult to predict how the ECJ will decide, but it usually sides with the Advocate General.

1.4. The WTO dimension
It is against this background that this paper considers another aspect of the EU’s scheme, namely, its legality under WTO law. This is based on the assumption that any compliance costs imposed on aviation are likely to result in new burdens for WTO Member exports of both goods and services to the EU (particularly tourism involving EU consumers). The fact that these burdens may be slight, or that in some cases foreign products and services may even end up in a better situation than before, is not material. Nor does it necessarily matter that, in broad terms, the ETS affects the EU economy more than it does other economies. WTO law takes a myopic view, for the most part focusing on how measures affect any given ‘like product’ or ‘like service’.
Another preliminary point concerns the various WTO carve-outs for measures affecting air transport services. As the analysis in this paper demonstrates, these are more limited than is often assumed. In particular, while the General Agreement on Trade in Services (GATS) Annex on Air Transport Services establishes certain substantive and procedural carve outs for measures affecting trade in air transport services, it also does not entirely rule out the application of WTO law to the EU’s scheme. This particular point is analysed at length below.

2. Trade in Goods
2.1. Is the EU’s scheme a tax or charge, or a regulatory measure?
The GATT treats measures differently according to whether they are considered as taxes, charges or general regulations. It is therefore helpful to begin by determining into which of these categories the EU’s scheme falls.
The EU’s scheme as a tax
Carbon emissions schemes that simply require industries to pay for their carbon emissions can be considered a form of taxation. However, ‘cap and trade’ schemes such as the EU’s scheme, in which allowances are created, distributed and traded, are different. As Advocate General Kokott said in the ATAA case:
It would be unusual, to put it mildly, to describe as a charge or tax the purchase price paid for an emission allowance, which is based on supply and demand according to free market forces, notwithstanding the fact that the Member States do have a certain discretion regarding the use to be made of revenues generated.
This paragraph makes two points. The first is that a tax is imposed by and accrues to the state; the second is that the EU’s scheme involves a ‘price’ for an allowance. It is suggested that the first point is of lesser importance. If a state imposes a requirement to purchase motorcycle helmets, this does not become a ‘tax’ just because the state also sells these helmets. It is suggested, rather, that what characterises both the EU’s scheme and this example as regulatory measures is the fact that, for a ‘price’, private actors obtain property with independent value. Who obtains this ‘price’, and whether it is fixed or variable, are secondary considerations.
The EU’s scheme as a charge
But even if the EU’s scheme is not a ‘tax’, does it impose a ‘charge’? Article II:2(c) GATT shows that that ‘charges’ may be levied for non-fiscal reasons: this includes ‘charges’ for services supplied by the state. For Advocate General Kokott, this was decisive:
Charges are levied as consideration for a public service used. The amount is set unilaterally by a public body and can be determined in advance. Other charges too, especially taxes, are fixed unilaterally by a public body and laid down according to certain predetermined criteria, such as the tax rate and basis of assessment.
An emissions trading scheme such as the EU scheme, however, is a market-based measure. No provision is made for fees or charges for the acquisition of emission allowances … If emission allowances are subsequently traded in the market after their allocation by the competent authorities, the price will in that case also be governed by supply and demand and is not fixed in advance.
This seems correct, although, as just noted, what seems more significant is that, as with taxes, a ‘charge’ does not give the affected individual anything of tradable value. One cannot resell a service.
If this analysis is correct, it would follow that the EU’s scheme should not be seen as imposing a tax or a charge on individual airlines. Rather, the costs of the EU’s scheme should be seen as the costs of complying with a regulatory measure.

2.2. Internal and non-internal aspects of the EU’s scheme
Article III:4 GATT is a non-discrimination provision guaranteeing that imported products are accorded no less favourable treatment than ‘like’ domestic products. It applies to measures ‘affecting the internal sale, offering for sale, purchase, transportation, distribution or use’. It has long been understood that the word ‘affecting’ means that Article III:4 applies to measures affecting conditions of competition for imported products in the marketplace. However, this cannot mean that Article III:4 applies to all measures with this effect: if it did, it would also govern quantitative restrictions. There must be another way to distinguish measures that are internal from those that are not.
It is suggested that the correct test for determining this question is to ask whether a measure regulates (or most directly affects) an act involving products once they have been imported. This is for several reasons. First, it respects the wording of Article III:4, which refers to measures affecting internal acts. Second, it parallels the Appellate Body’s test for taxes and charges, which are subject to Article III:2 only when they ‘accrue’ on the basis of an internal condition or event. Third, it explains the jurisprudence on the point.
There are two cases of importance. In the first, EC – Bananas III, the Appellate Body decided that a measure allocating import licences to domestic distributors is to be assessed under Article III:4. The Appellate Body said that this was because the rules were intended to have an effect on the sales of competing domestic products. But this is not convincing. As already noted, the same can be said of any quantitative restriction. The better explanation is that the measure directly regulated an act (distribution) that was only relevant to imported products. The second, US – FSC (Article 21.5 – EC) involved a tax rebate contingent on the use (ie purchase) of domestic products. The Appellate Body decided that this was an internal measure because it affected ‘the “internal … use” of imported products, within the meaning of Article III:4 of the GATT 1994, as compared with like domestic products.’ For the reasons given, it is suggested that the first part of this sentence is correct; the second irrelevant.
Enforcement measures under the Note Ad Article III
Article III:4 also has an extended application to certain non-internal measures, by virtue of the Note Ad Article III. This Note states, relevantly, that a measure that is enforced at the time of importation is still to be treated as an internal measure. It permits WTO Members to employ border measures, which need not have any formal correlation to the internal measure being enforced, where this can be justified on the grounds of administrative efficiency as part of a domestic regulatory scheme.
Sometimes the Note Ad Article III is overlooked. This occurred, for example, in the Panel Report in China – Auto Parts, which, relevantly, involved administrative requirements imposed at the border on imported products, with a view to enforcing an internal charge. Citing EC – Bananas III and US – FSC (Article 21.5 – EC), the Panel considered these measures to be internal measures because they affected the conditions of competition of the relevant products once they had been imported. For the reasons suggested here, it is suggested that the result was correct, but the reasoning flawed. The measures were internal measures, but because they enforced an internal charge. This issue was not appealed, and the Appellate Body seems to have thought that the Panel’s approach was appropriate.
Application to the EU’s scheme
On this analysis, the EU’s scheme is properly considered an internal measure to the extent that it applies to products once they have been imported. In practice, this means that the EU’s scheme is internal insofar as it applies to imported products on intra-EU flights. Second, the EU’s scheme can be considered an internal measure to the extent that it applies to air transport over EU airspace prior to importation. This aspect of the EU’s scheme qualifies as an enforcement measure under the Note Ad Article III. It relieves the EU authorities of requiring importation formalities to be conducted at the border. However, the application of the EU’s scheme to flights outside of EU airspace cannot be justified in the same way. This aspect of the EU’s scheme is not necessary, or even linked, to the internal aspects of the EU’s scheme. This aspect of the scheme is not therefore an internal measure under Article III:4. Whether or not is qualifies as a quantitative restriction under Article XI:1 GATT will be considered below.

2. 3. The application of the EU’s scheme to EU airspace
Article III:4 GATT requires that the EU’s scheme not accord ‘less favourable treatment’ to imported products than to ‘like’ domestic products. In at least some cases, this is unlikely to pose any difficulties. Within EU airspace, oranges flown from Cadiz to Frankfurt have no advantage over oranges flown from Miami to Frankfurt. But this does not hold if the majority of Cadiz oranges are transported to Frankfurt by some other means (aviation being the only means of transport covered by the EU’s ETS). Should this occur, there is a possibility that the Miami oranges are accorded ‘less favourable treatment’ under Article III:4.
Relevance of the second sentence of Article III:4
The second sentence of Article III:4 offers a certain degree of flexibility in connection with charges for internal transportation. It states that:
The provisions of this paragraph shall not prevent the application of differential internal transportation charges which are based exclusively on the economic operation of the means of transport and not on the nationality of the product.
It seems unobjectionable to interpret the phrase ‘economic operation of the means of transport’ in light of the ‘polluter pays’ principle. It would follow that a WTO Member could internalise the costs of carbon emissions by means of a charge on internal transportation. However, strictly speaking this sentence does not apply in the present case. It is limited to transportation ‘charges’, and according to the above analysis the EU’s scheme is not a ‘charge’. It is, however, considered below whether this sentence might yet have some interpretive value in other contexts.
The meaning of ‘less favourable treatment’ in Article III:4
The meaning of ‘less favourable treatment’ in cases of formally equal treatment of imported and domestic products is still unclear, as recent cases have demonstrated. Nonetheless, two main approaches may be distinguished.
Two approaches
A ‘disproportionate impacts’ approach focuses on the relative effects of the measure on products of a certain origin. This approach has two variations. Using an ‘asymmetry’ test, one compares the groups of ‘like’ imported and domestic products. There is discrimination if the measure accords ‘less favourable treatment’ to a larger proportion of the group of imported products than it does to the group of domestic products. This test is to be distinguished from a stricter ‘diagonal’ test, according to which there is discrimination if any individual imported ‘product’ is treated less favourably than any ‘like’ domestic ‘product’. This ‘diagonal’ test has mainly been applied to taxes under the first sentence of Article III:2 GATT, and does not appear relevant to the context of regulations under Article III:4.
A second approach focuses on whether the favourable treatment has, in practice, been ‘denied’ to the imported product. A classic example, from the most favoured nation context, is a 1904 German measure granting market access for all cows that grazed at Alpine altitudes: in law this treatment was available to Danish cows; in reality it was denied to these cows. The evident idea was to ensure that the favourable treatment was only available to Swiss cows.
Generally, when imported products are denied favourable treatment they also suffer a disproportionate impact. A finding that favourable treatment has been ‘denied’ then serve as a proxy for a ‘disproportionate impacts’ approach. But the corollary does not hold: there can be a disproportionate impact even when favourable treatment is not denied to products because of their origin. An example is Dominican Republic – Cigarettes, involving a measure according to which imported cigarettes incurred a higher per unit bond cost than domestic cigarettes. According to the Appellate Body, this had nothing to do with the fact that the imported cigarettes were imported. It had to do with the fact that these cigarettes had a low market share. In other words, there was no causal link between the origin of the affected products and their treatment.
In the Swiss cows example, the denial of benefits approach is a useful shortcut to a finding of less favourable treatment. In DR – Cigarettes, it counteracts a disproportionate impact, to the opposite effect. But these are both unusual cases. The difficult cases are when the more favourable treatment is not denied but just made more expensive.
This distinction is important, as can be seen from the recent US – Tuna II Panel Report, in the related context of the national treatment obligation under Article 2.1 of the TBT Agreement. In this Report, the Panel decided that an ‘advantage’ (access to a label) had been denied to the imported products because Mexican tuna processors were able to source their tuna from non-Mexican fleets. However, in coming to this decision, the Panel did not consider the costs to the processors of changing their sources of supply. Moreover, when it came to the Panel’s consideration of the Mexican fleets themselves, it said expressly that ‘the existence of adaptation costs, in itself, does not, in our view, imply the existence of less favourable treatment.’
This is unrealistic, and certainly not a principle of general application. In US – Tuna II, the Panel did not think that adaptation costs were relevant; in fact, this is the fundamental question. After all, with sufficient resources, it might have been possible for the Danish cow farmers in 1904 to build a mountain. Something of the sort is now proposed in the Netherlands.
An alternative approach: ‘reasonably available’ favourable treatment
It is for this reason that another test is suggested for these cases. This test asks whether the favourable treatment is ‘reasonably available’ to private actors concerned with the imported products.
There are several advantages to such a test. It recognizes – expressly – that private actors often have a choice whether to obtain the favourable treatment, and focuses on the reasonableness of the costs of that choice. It is responsive to legitimate policy considerations, and yet not flawed by the evidentiary problems of a test that looks at protectionist intent (whether subjective or manifested in the ‘structure’ of the measure). It is also flexible enough to deal with exceptional situations, such as that which arose in US – Clove Cigarettes. The Panel in this case struggled with the parties’ different views on ‘less favourable treatment’ in the context of Article 2.1 of the TBT Agreement, but found a way out, based on the unusual facts of the case. The US had decided not to apply the measure to like domestic products in order to save costs. According to the Panel, this was impermissible: the US ‘must not accord less favourable treatment to imported clove cigarettes than that it accords to the like domestic menthol cigarettes for reasons of avoiding potential costs’.
The ‘reasonably available’ treatment test would say that even when the more favourable treatment is available to imported products, and, implicitly, even if otherwise it might be reasonable for these costs to be borne by the imported products, it is not reasonable when the costs have been intentionally waived for the like domestic products.
Application to the EU’s scheme
How, then, does this apply to the EU’s scheme? If proportionately more Miami oranges than Cadiz oranges are flown to Frankfurt, the EU’s scheme will have a disproportionate impact on Miami oranges, for the reasons given. On a disproportionate impact test, and on these facts, there would be less favourable treatment. On a denial of favourable treatment test, it is more difficult to say in the absence of further facts. If Cadiz airport has good ground freight facilities, perhaps the favourable treatment is reasonably available to Miami oranges. An airplane from Miami could land in Cadiz and from there on obtain the same treatment as the Cadiz oranges. For one reason or another, this might not be a realistic option for the private actors involved in the importation of Miami oranges. But even if the favourable treatment is not ‘denied’, it may not be ‘reasonably available’, according to the test proposed here. Again, this depends on the facts, and on a judicial assessment of what is ‘reasonably available’ in the circumstances. But there is at least a chance that the favourable treatment enjoyed by the Cadiz oranges is not ‘reasonably available’ to the Miami oranges, and that there would be a violation of Article III:4.

2.4. The application of the EU’s scheme outside of EU airspace
Application of Article XI:1 GATT
To the extent that the EU’s scheme does not fall under Article III (either directly or via the Note Ad Article III), it might be a quantitative restriction under Article XI:1 GATT. This article states as follows: No prohibitions or restrictions other than duties, taxes or other charges, whether made effective through quotas, import or export licences or other measures, shall be instituted or maintained by any contracting party on the importation of any product of the territory of any other contracting party or on the exportation or sale for export of any product destined for the territory of any other contracting party.

WTO panels have interpreted the term ‘other measures’ as a broad residual category covering ‘any form of limitation imposed on, or in relation to importation’. In Colombia – Ports of Entry, the Panel considered whether this covered a measure that restricted the ports that could be used by importers. The Panel decided that it would if the measure affected the cost of shipping products from the port of origin to the place of sale (evidence was lacking on this point); or if it was applied in an arbitrary manner, thereby increasing the uncertainty of private actors involved in the importation of the product (it was).

Application to the EU’s scheme
Insofar as it applies to pre-import transportation over non-EU airspace, the EU’s scheme shares certain features with the measure in Colombia – Ports of Entry. It has a direct impact (and even directly regulates) the means by which products are imported into the EU. It is true that Advocate General Kokott, in her Opinion in the ATAA case, denied that the EU’s scheme was not ‘[a] concrete rule regarding [foreign airlines’] conduct within airspace outside the European Union.’ But it is difficult to see how an EU Directive that imposes fines of €100 for any allowance not obtained and ‘surrendered’ to the EU can be seen as anything but a ‘concrete rule’.
The EU’s scheme also has similar effects to the measure in Colombia – Ports of Entry insofar as the cost of complying with the measure is likely to increase transportation costs. The EU’s scheme is even similar insofar as its impacts on imported products vary, unpredictably, with the carbon market.

While a firm conclusion is difficult, there is reason to believe that Article XI:1 applies to the EU’s scheme insofar as this scheme imposes compliance costs on non-EU aircraft for carbon emissions produced in non-EU airspace. Further, if so, there is reason to believe that this element of the EU’s scheme amounts to a ‘restriction … on the importation’ of products of other WTO Members contrary to Article XI:1 GATT.

2.5. The most favoured nation obligation (Article I GATT)
Article I:1 GATT requires that:
any advantage, favour, privilege or immunity granted by any contracting party to any product originating in or destined for any other country shall be accorded immediately and unconditionally to the like product originating in … the territories of all other contracting parties.

Application of Article I:1 to the EU’s scheme
This obligation clearly applies to the internal aspects of the EU’s scheme; in this respect, however, it is difficult to see how there could be any violation. Once products are imported, and even between arriving in EU airspace and undergoing import formalities, the EU’s scheme does not seen to have any disproportionate impact on imports from any given origin. Nor does there appear to be any ‘Swiss cow’-style denial of treatment, or a failure to accord treatment that is not reasonably available, to imported products from all sources.
A more difficult question, with additionally complicated consequences, is whether Article I:1 applies to the EU’s scheme insofar as it accounts for emissions produced on flights outside of EU territory.
This raises a question fundamental to the reality of international trade, and yet rarely, if ever, addressed: does the most favoured nation obligation, the ‘cornerstone’ of the GATT, apply to the international transportation of products? There are good reasons to believe that it should. To choose some straightforward examples: a country might impose regulatory compliance costs on all products arriving by sea, to the advantage of its continental neighbours; or another might impose costs on all products coming descending from the Alps. In both cases, products of a particular origin are denied the more favourable treatment. Nor is such a situation covered by Article V:6 (discussed below), which requires non-discrimination between different routes from the same origin.
But just because Article I:1 should apply does not mean that it does. An answer to this question depends upon its precise wording. Article I:1 describes the measures to which it applies as, relevantly, ‘rules and formalities in connection with importation’. On a narrow reading, this phrase is limited to rules regulating the actual act of importation, such as customs formalities. As discussed in the context of Article XI:1, however, there is a ‘connection’ between international transport costs and the importation of products: for example, variability in the calculation of these costs could provide an incentive to import at a particular location. This is a particular possibility in the case at hand. Switzerland does not participate in the EU’s scheme. As a result, a flight of Miami oranges could land in Geneva, and reroute to Frankfurt by road, thereby avoiding any compliance costs whatsoever. The likely effects of international transport costs on import decisions should suffice for Article I:1 to be applicable to international transportation prior to importation.
According an ‘advantage’ to all ‘like’ products
On the basis that Article I:1 does apply, it is necessary to consider whether the EU’s scheme accords an ‘advantage’ to ‘like’ products from different WTO Members, not only formally but also de facto. The following considers two examples, both involving (as a hypothetical) oranges exported to Frankfurt from Miami and Marrakesh respectively. In the first example, the oranges are air freighted; in the second, the oranges from Miami are currently air freighted, and the oranges from Marrakesh are sea freighted.
Identifying the ‘advantage’
The first question is how to identify the ‘advantage’ in these example. Should it be the actual compliance cost per shipment of oranges, or should it be the formula used to calculate the compliance cost? If it is the actual cost, there will often be a disproportionate impact: Miami oranges, in the present example, will be subject to greater absolute compliance costs than Marrakesh oranges. If it is the cost formula, the answer may be different.
The tendency in the jurisprudence is to consider the ‘advantage’ in terms of market access. Usually, this involves a lower duty per product, although in US – Tuna II (a case under Article 2.1 of the TBT Agreement) the Panel framed the ‘advantage’, in a way that seems similar, as the availability of a particular label for a given product. These cases make sense. To frame an ‘advantage’ in terms of a formula is to invite de facto discrimination. Formulas can easily conceal origin-specific factors: it one could structure the 1904 German measure to be structured as a duty based on grazing altitude. It is safer to consider the ‘advantage’ as the actual cost or other burden borne by any given product.
There is however another consideration. The second sentence of Article III:4, discussed already, allows for the possibility of differential charges for internal transport according to the economic costs of the means of transport. Might it be argued that the same principle should apply to regulatory compliance costs involving international transport? Given the wording of Article I:1, which speaks of ‘advantages’ being accorded, the only way to operationalise this result would be to frame the ‘advantage’ in terms of the formula for calculating the differential economic costs of international transport. It would follow that the Miami and Marrakesh oranges would be accorded the same advantage to the extent that they are both transported by air, though not if the latter are transported by sea.
According the advantage to all like products
Having identified the ‘advantage’ is only the first step; it is next necessary to determine when the advantage is accorded to like products of WTO Members.
Much of the jurisprudence on this question deals with denial of advantage cases. In EC – Tariff Preferences, a panel rejected a measure according to which products received different duties depending on whether their country of origin had difficulties in regulating drugs; and in EC – Fasteners, a panel rejected a measure applying different duties to products depending on whether their country of origin had a ‘market economy’. In neither case was the more favourable treatment available to the affected products.
US – Tuna was a more complicated case, and also more helpful to the present question (although it involved Article 2.1 TBT). In this case, the affected producers were able to change their fishing grounds, in which case they would have been more likely to have had access to the favourable treatment. The question which should have been asked was whether this favourable treatment was ‘reasonably available’. Unfortunately, the Panel asked the wrong question: it said that a restriction that de facto applied predominantly in a particular geographical area affected Mexican fleets and other fleets equally. This was like saying that because the 1904 German measure treated Dutch cows as badly as Danish cows, there was no discrimination in favour of Swiss cows. It is a simple logical error.
There is a reason why the jurisprudence on Article I:1 does not squarely confront this question in these terms. It is because of a widespread analytical error, according to which this question is considered in terms of the requirement in Article I:1 to accord advantages ‘immediately or unconditionally’. Thus, Panels are now engaged in an argument as to whether all conditions not related to a product are prohibited or whether ‘conditions’ applied to private parties are permitted unless they are discriminatory. These questions have even carried over to cases of de jure discrimination. In fact, the premise of the argument is erroneous. Unconditional most favoured nation treatment (historically, an alternative to conditional most favoured treatment) is concerned with conditions addressed to states, such as whether or not adopt a certain regulatory system, or enter into a treaty. It is not concerned with ‘conditions’ (or other burdens) addressed to individuals.

Application to the EU’s scheme
It is not entirely certain that Article I:1 GATT applies to the EU’s scheme, given that the EU’s scheme is not necessarily a measure ‘in connection with importation’. But if it does apply, the EU will not be able to exempt airlines from its scheme on a selective basis, as is envisaged in the Directive, which permits a waiver if ‘a third country adopts measures for reducing the climate change impact of flights departing from that country which land in the Community’. Beyond this, it is likely, though not certain, that there will be discrimination in favour of products that are transported by sea (or other non-air means). Whether or not there is a violation in connection with groups of products transported by air depends on whether the ‘advantage’ is defined in terms of absolute costs or the applicable cost formula. In the former case, there is more likely to be a violation, because products from some origins (eg Miami oranges) will also be liable to greater compliance costs than like products from others (eg Marrakesh oranges). In the latter case, however, the result might be different.

2.6. Freedom of transit (Article V GATT)
The EU’s scheme also involves goods in transit, in two ways: first in relation to products that transit across the EU, and second in relation to products that have been in transit before they arrive in the EU as a final destination. Article V GATT sets out obligations in relation to both scenarios.
Application of Article V: the carve-out in Article V:7
It is appropriate to comment first on the carve-out in Article V:7 GATT. While this paragraph states that ‘[t]he provisions of this Article shall not apply to the operation of aircraft in transit’, it continues to stipulate that it ‘shall apply to air transit of goods (including baggage).’ Article V:1 which defines ‘traffic in transit’ to include ‘[g]oods (including baggage), and also vessels and other means of transport’. By simple subtraction, it is clear that it is only the operation of the aircraft themselves that is exempt from Article V. Goods and baggage – the items of present interest – are fully covered.

EU as a transit territory
Article V applies in the first instance to the EU in its capacity as a transit territory. Article V:3 states, relevantly, that:
… traffic coming from or going to the territory of other contracting parties shall not be subject to any unnecessary delays or restrictions and shall be exempt from … all transit duties or other charges imposed in respect of transit, except charges for transportation …
If the EU’s scheme is considered a transportation ‘charge’, it is then subject to Article V:4, which specifies that ‘[a]ll charges and regulations imposed by contracting parties on traffic in transit … shall be reasonable, having regard to the conditions of the traffic.’ However, for the reasons given above, it is difficult to conceive of the EU’s scheme as a ‘charge’ at all, whether on transportation or otherwise. It is better seen as a regulatory scheme which imposes compliance costs on airlines, and therefore on their customers.
The question is then whether or not the scheme constitutes an ‘unnecessary restriction’ (Article V:3) or unreasonable regulation’ (Article V:4). It is difficult to know what this means, in particular because there is no benchmark against which the ‘necessity’ or ‘reasonableness’ of the measure could be tested. The argument, presumably, would be that the EU’s scheme is ‘necessary’ to implement the ‘polluter pays’ principle in connection with transport and a ‘reasonable’ regulation for the same reason. Some support for this approach might be found in the second sentence of Article III:4, which, as mentioned, permits differential charges on internal transportation corresponding to its real economic costs. If so, the EU’s scheme would appear safe under Article V:3 and Article V:4.
EU as destination
Most of the obligations set out in Article V relate to products while they are in transit. However, the Panel in Colombia – Ports of Entry confirmed that Article V:6 imposes obligations on the WTO Member of final destination in relation to products previously in transit.
Article V:6 is a most favoured nation obligation that compares different transit routes from the same origin to the same destination. It states:
Each contracting party shall accord to products which have been in transit through the territory of any other contracting party treatment no less favourable than that which would have been accorded to such products had they been transported from their place of origin to their destination without going through the territory of such other contracting party.
Because allowances are only required for the final leg of any journey, the EU’s scheme frequently imposes different requirements on like products from the same origin depending on their transit routes. This is the case whether one compares different indirect flights, indirect flights with direct flights, or flights with other means of transport. But does this mean that ‘favourable treatment’ is not being accorded to products taking different journeys?
The meaning of this phrase has been discussed above. Based on that discussion, the treatment can be identified as the compliance costs borne by any given product, and the favourable treatment as the opportunity of being subject to the lowest available compliance costs. On this basis, it is theoretically possible, but practically unlikely, that any favourable treatment is ‘denied’ to any given product from any given origin. Presumably if one box of Miami oranges can travel to the EU by ship, others can as well. The same result is true if the appropriate test is whether the treatment is ‘reasonably available’. The EU’s scheme would appear safe under Article V:6.

2.7. Exceptions
The foregoing analysis indicates that the EU’s scheme may be inconsistent with at least some of its obligations under the GATT, principally Article XI:1 GATT (insofar as the EU’s scheme applies outside of EU airspace) and Article I:1 (if the EU grants a selective exemption to certain airlines). Whether it is non-discriminatory in other ways depends on the facts. However, as other means of transport are treated more favourably than air transport, there is likely to be discrimination if that other means of transport is, for one reason or another, not equally available to products from all WTO Members. Nonetheless, it is possible that the scheme might be justified under the general exceptions set out in Article XX GATT.
Conservation of exhaustible natural resources (Article XX(g) GATT)
The most relevant exception is Article XX(g), which permits WTO Members to take measures ‘in relation to the conservation of exhaustible natural resources’, provided that such measures are ‘made effective in conjunction with restrictions on domestic production or consumption’. Another potentially relevant exception is Article XX(b) GATT, which permits WTO Members to adopt measures necessary for the protection of human or animal or plant life or health. However, the ‘necessity’ requirement is a higher standard. Where measures can fall under both paragraphs, they are typically argued under Article XX(g).
The present measure, adopted to mitigate climate change, is clearly in relation to the conservation of the atmosphere, almost certainly an ‘exhaustible natural resource’ and is made effective in conjunction with similar domestic measures (EU ETS more generally, the EU ETS more specifically as it applies to EU airlines). Territoriality also does not appear to present a problem. In US – Shrimp, the Appellate Body held that turtles (as a species) were an essentially migratory species, and therefore sufficiently within US territory to provide a ‘jurisdictional nexus’ for the regulation. The ‘atmosphere’ that the EU seeks to protect has, if anything, an even closer ‘jurisdictional nexus’ to the EU, as Advocate General Kokott suggested in her Opinion. There can be little doubt that the measure would fall within the scope of Article XX(g).
Discrimination under the Chapeau to Article XX
A somewhat more difficult question is whether the measure would survive the conditions set out in the Chapeau of Article XX. It may be assumed that the measure is not a disguised restriction on international trade. But does it amounts to arbitrary or unjustifiable discrimination between countries where the same conditions prevail, or between countries where different conditions prevail?
‘Conditions prevailing’
In US – Shrimp the Appellate Body gave the impression that the purpose of a measure is relevant only to its provisional justification under one of the subparagraphs of Article XX. It implied that the Chapeau is relevant to the ‘application’ of a measure, as somehow distinct from the measure itself. However, as at least one author has pointed out, this makes no sense. The purpose of the challenged measure must be relevant under Chapeau. It is the only relevant benchmark for determining the relevant ‘conditions prevailing’ for the purposes of a discrimination analysis, and for determining whether any such discrimination is unjustifiable or arbitrary.
This is also the understated premise of most cases on the Chapeau, although it is rarely relevant in the case of measures focused on domestic activities. Brazil – Retreaded Tyres is an example. Here, the measure restricted the disposal, and correspondingly the importation, of retreaded tyres. The acts causing the risks were solely domestic, and so for all tyre-producing countries the ‘conditions prevailing’ were identical.
But the question is more pressing when the challenged measure focuses on acts taking place in other countries. For example, in US – Gasoline the acts allegedly causing the risk were both domestic (gasoline consumption) and foreign (gasoline production methods). The relevant question was whether the ‘conditions prevailing’ in Venezuela and the United States were the same (they were). Likewise, in relation to a quarantine measure, ‘conditions prevailing’ in a country with rinderpest will differ from those in a country without rinderpest. And for a measure concerning migratory resources the ‘conditions prevailing’ will be any acts relevant to those resources, such as fishing practices.
The EU’s scheme seeks to protect the EU from the effects of climate change by internalising carbon emissions produced by airlines both in the EU and elsewhere. It also takes into account the different ‘conditions prevailing’ in different countries by offering the possibility of exempting airlines from countries that have adopted ‘equivalent measures’. This avoids ‘double counting’, as urged by an ICAO resolution on the subject. So long as it is not discriminatory in some other way, the EU’s exemption mechanism seems therefore to be justified under the Chapeau.
However, as between non-exempt countries the situation is not so simple. This is because the EU’s scheme does not, in fact, internalise all carbon emissions produced by airlines. It only internalises carbon emissions produced on flight sectors beginning or terminating at an EU airport. As a result, there is a potential difference in treatment between non-exempt countries, depending on the routes and means of transport chosen. A flight from Hong Kong to Frankfurt is almost the same as from Guangzhou. However, if the flight from Hong Kong transits in Dubai, it will only have to surrender allowances for the final flight sector from Dubai to Frankfurt. This may amount to discrimination if the Guangzhou flight is not able to land in Dubai (or equivalent), or only with difficulty. Nor is this unforeseeable, given the highly regulated system of air transport services. Moreover, this problem is exacerbated when one compares air transport with other means of transport which are not covered by the EU’s scheme.
Unjustifiable or arbitrary discrimination
Even if there is discrimination – and this might be a remote possibility – this does not mean that the discrimination is unjustifiable or arbitrary. This depends on the rationale for the discrimination. In Brazil – Retreaded Tyres, the Appellate Body said this:
[W]e have difficulty understanding how discrimination might be viewed as complying with the chapeau of Article XX when the alleged rationale for discriminating does not relate to the pursuit of or would go against the objective that was provisionally found to justify a measure under a paragraph of Article XX.
Both of these factors are relevant here. To the extent that the EU scheme discriminates between two non-exempt countries, the reason is that it does not apply its scheme to flights that do not terminate (or begin) at an EU airport, and it does not apply its scheme to other means of transport. This is presumably because the EU considers that it lacks jurisdiction, under international law, to apply its scheme to emissions generated in this way.
This might be overly modest. In her Opinion in the ATAA case, Advocate General Kokott took the view that the EU was able to protect itself against climate change because, inter alia, ‘greenhouse gases … can have effects on the environment and climate in every State and association of States, including the European Union.’ It would follow that the EU could in fact include all flights and other means of transport in its scheme. But even if the EU were somehow constrained, on the current state of the law this would not be relevant under the Chapeau. In Brazil – Retreaded Tyres the Appellate Body said that Brazil’s international law obligations were no defence. In that case, the rationale for the discrimination was Brazil’s need to comply with a decision of a Mercosur tribunal, but this failed on both counts. It had nothing to do with the purpose of the measure, which was to protect domestic health, and the effect of complying with this decision was to worsen the risk to public health, due to potential increases in imports of retreaded tyres from Uruguay. Nor did it matter that the worsened risk was only to a ‘small degree’.
This reasoning has direct application in the present case. Not only is this reason for discriminating unrelated to the objective of the EU’s scheme, it even contradicts it, and not even necessarily to a ‘small degree’. Lufthansa calculates that a direct flight from Hong Kong to Frankfurt emits 251t, while an indirect flight via Dubai emits 296t, 18 percent more. Effectively, the EU’s scheme gives airlines an incentive to produce more emissions. This cannot be justified in terms of the rationale for the scheme.

3. Trade in services
The GATS applies to all measures affecting trade in services. As mentioned, by imposing costs on air travel, the EU’s scheme has an inhibiting effect on trade in services, especially services delivered outside the EU. However, aviation transport services are largely exempt from the GATS by virtue of the GATS Annex on Air Transport Services. Does this mean that the GATS does not protect services dependent on air transport?

3.1. The Annex on Air Transport Services
Scope of the Annex
Paragraph 1 of the Annex states that it applies to ‘all measures affecting trade in air transport services, whether scheduled or non-scheduled, and ancillary services’. The language used is reminiscent of the phrase ‘measures affecting trade in services’ in Article I:1 GATS, and it seems appropriate to interpret both in a similar way. Speaking of Article I:1, the Appellate Body has said that this is a broad term which covers any measures which have an effect on trade in services.
But does this phrase also cover all measures affecting trade, or only those measures affecting conditions of competition for foreign services and service supplier. It seems that the more limited interpretation is more common, even among complaining parties in litigation. But this cannot be correct. This would lead to the duplication of an inquiry properly conducted in the context of relevant non-discrimination obligations. In addition, the GATS contains provisions, such as those on domestic regulation in Article VI, which have nothing to do with discrimination. The answer must therefore be that Article I:1 GATS applies also to measures that have no effect on conditions of competition, or – to put it another way – non-discriminatory measures.
This has a direct bearing on paragraph 1, where similar considerations also apply. As will be seen, the Annex contains provisions that apply also to non-discriminatory measures. The phrase ‘measures affecting trade in air transport services’ must therefore also be understood to mean measures affecting the quantity and type of services provided by foreign service suppliers, not just measures affecting their conditions of competition, which might exclude non-discriminatory measures.
Paragraph 2 ATS
The main substantive carve out for measures affecting trade in air transport services is set out in paragraph 2 ATS. This paragraph states as follows:
2. The Agreement [GATS], including its dispute settlement procedures, shall not apply to measures affecting:
(a) traffic rights, however granted, or
(b) services directly related to the exercise of traffic rights, except as provided in paragraph 3 of this Annex
Both of these paragraphs are relevant to the EU’s scheme.
Paragraph 2(a) ATS
Paragraph 2(a) exempts ‘measures affecting traffic rights’ from GATS obligations. ‘Traffic rights’ are defined in paragraph 6(d) as follows:
‘Traffic rights’ mean the right for scheduled and non-scheduled services to operate and/or to carry passengers, cargo and mail for remuneration or hire from, to, within, or over the territory of a Member, including points to be served, routes to be operated, types of traffic to be carried, capacity to be provided, tariffs to be charged and their conditions, and criteria for designation of airlines, including such criteria as number, ownership, and control.
The most likely way in which the EU scheme might be deemed a ‘measure affecting traffic rights’, as per the definition of such measures in paragraph 6(d), is if the scheme affects ‘tariffs to be charged and their conditions’. Tariffs are commonly defined in air services agreements as:
The prices to be paid for the carriage of passengers, baggage and freight and the conditions under which those prices apply, including prices and conditions for agency and other auxiliary services, but excluding remuneration or conditions for the carriage of mail.
This language must be understood in the context of the original system for tariff setting, in which tariffs were negotiated within the International Air Transport Association (IATA) and approved by governments. Since 1979, these negotiations have been largely superseded by deregulated private fares.
A measure require a maximum or a minimum tariff to be charged would clearly affect the right to set tariffs. However, a measure of general application imposing regulatory costs on airlines is in a different category. No doubt for this reason, in none of the various ICAO disputes to date has there been any suggestion that EU’s scheme affects tariffs. Indeed, the airlines challenging the EU’s scheme in the European Court of Justice do not even allege that the scheme affects their right to charge prices under Article 11 of the US-EU Open Skies Agreement. And the EU itself states that even though there might be a ‘minor’ impact on prices, ‘[i]ncluding aviation in the EU ETS will not directly affect or regulate air transport tickets’. This all supports the view that the EU’s scheme does not affect ‘tariffs to be charged and their conditions’ within the meaning of paragraph 6(d).
If, accordingly, the EU’s scheme is not a measure affecting tariff rights within the meaning of paragraph 6(b) of the Annex, it is also not covered by the exemption in paragraph 2(a).
Paragraph 2(b) ATS
Paragraph 2(b) ATS establishes another substantive carve out for ‘measures affecting services directly related to the exercise of traffic rights’. These services are undefined, but correlate broadly to the so-called ‘soft rights’ involving currency exchanges, ground and baggage handling, catering, marketing, and airport usage. It is possible that the EU’s scheme might affect these services, as a result of airlines changing routes to minimise their compliance costs under the EU’s scheme. To the extent that it does, paragraph 2(b) would be applicable and the EU’s scheme would be exempt from scrutiny under the GATS. However, this is by no means certain, and it is therefore still appropriate to pursue an analysis under the GATS.
Paragraph 4 ATS
Paragraph 4 of the Annex establishes a procedural carve-out for measures affecting trade in air transport services. It states that, in relation to the measures defined in paragraph 1, WTO dispute settlement is only available ‘where … dispute settlement procedures in bilateral and other multilateral agreements or arrangements have been exhausted.’
When, then, are the conditions in paragraph 4 satisfied? The point of this paragraph, and the point of the Annex more generally, is to ensure the primacy of the ICAO system over the WTO system in cases of regulatory overlap, and perhaps also to prevent true conditions of competition in the market for air transport services. But primacy can be applied in different ways. On a narrow view, primacy would apply in relation to matters prohibited by an ICAO agreement. More generally, it might be thought that paragraph 4 applies also to matters governed in some way by the ICAO, including by positive authorization. But at least the matter would have to fall within ICAO competence to some degree.
In the case at hand, there is good reason to believe that the EU’s scheme does not violate any ICAO obligations. There is no definitive ICAO ruling on the matter, but Advocate General Kokott has decided that the EU’s scheme does not violate any relevant ICAO obligations, and this echoes decisions to similar effect by the UK High Court and the Dutch Supreme Court with respect to ‘ticket taxes’. In practical terms, it is also unlikely that the EU, the UK, the Netherlands and perhaps other governments would argue in WTO dispute settlement proceedings that the EU’s does (or even might) violate their ICAO obligations. This is particularly true for the UK, which has argued (successfully) that the Chicago Convention does not even have any ‘application’ to its Air Passenger Duty. If the narrow view is taken, the result would be that the conditions in paragraph 4 are not satisfied, and the EU’s scheme can be challenged in WTO dispute settlement proceedings.
However, the question is likely to be resolved differently if the broader view is taken that paragraph 4 applies if a matter is governed by the ICAO. In this case, it is. In Resolution A37/19, in a paragraph not subject to reservations, the ICAO Assembly ‘request[ed] the Council to ensure that ICAO exercise continuous leadership on environmental issues relating to international civil aviation, including GHG emissions’. It is true that some countries have claimed that the ICAO should cede this primary role to the United Nations Framework Convention on Climate Change (UNFCCC). However, on the present state of affairs, this should not change the conclusion that, for purposes of paragraph 4, the ICAO continues to govern the matter.
If this analysis is correct, then even if one of the substantive carve outs in paragraph 2 does not apply, it is likely that a WTO Panel would lack jurisdiction to determine whether there is a GATS violation until ICAO remedies have been exhausted. However, this does not mean that the WTO Member would be complying with its WTO obligations. It just means that dispute settlement is not available. For this reason, and also in the event that the preceding analysis is incorrect, the following considers the applicable GATS obligations and exceptions.

3.2. The most favoured nation obligation (Article II:1 GATS)
Article II:1 GATS, inspired by Article I:1 GATT, requires that any ‘advantage’ accorded by the EU to any service or service provider must be accorded immediately and unconditionally to the like service or service of any other WTO Member.
Unlike Article I:1 GATT, there is no doubt that Article II:1 GATS applies to the EU’s scheme. By virtue of Article I:1 GATS, Article II:1 applies to all measures with an effect on services. Clearly this measure has such an effect, most notably on services supplied to EU consumers travelling outside the EU, such as tourism. It seems also relatively clear that the EU’s scheme has a disproportionate effect on services and service suppliers in certain countries, because of the fact that the cost of compliance is calculated on the basis of distance. One might expect tourism in the United States to be proportionately more affected than tourism in Israel.
However, this does not necessarily mean that the measure fails to accord an ‘advantage’ to like services and service providers. This depends on the meaning of ‘advantage’ in the specific context of Mode 2 trade in services (consumption abroad). The critical question is whether the advantage is an actual compliance cost, or a formula for calculating this cost. As in the context of Article I:1 GATT, the present state of the law does not allow for a definitive answer to this question, although one would be inclined to decide in favour of the first option, for the reasons stated above. What can be said, however, is that any benefits flowing from the exemption of certain airlines from the EU’s scheme would need to be extended immediately to all services and service providers of all WTO Members.

3.3. Obligations applicable to commitments on service sectors
Unlike Article II:1, most of the other obligations under the GATS only apply to the extent that a WTO Member has made specific commitments in relation to those services. The EU has made full commitments in Mode 2 (consumption abroad) in relevant tourism and recreational services. The question arises whether the EU’s scheme violates any obligations with respect to these service sectors.
Market access (Article XVI GATS)
In the first place, one might consider whether the scheme violates Article XVI GATS. In respect of scheduled services, this forbids the measures described in Article XVI:2 GATS. Relevantly, these require that the measure set a maximum number of suppliers or various elements of services, whether in their form or (according to the Appellate Body, in cases of a zero quota) in their effect. The EU’s scheme does not, however, set any maximum limits, even if it has a restrictive effect on the supply of services. Article XVI does not therefore apply.
National treatment (Article XVII GATS)
The remaining question, then, is whether the EU’s scheme discriminates in favour of domestic services and service suppliers in these (and other) sectors, contrary to Article XVII:1 of GATS. This provision reads as follows:
In the sectors inscribed in its Schedule, and subject to any conditions and qualifications set out therein, each Member shall accord to services and service suppliers of any other Member, in respect of all measures affecting the supply of services, treatment no less favourable than that it accords to its own like services and service suppliers.
It seems clear that the EU’s scheme has the effect of modifying the conditions of competition in favour of EU services and service suppliers, compared to those of other WTO Members. As tickets become more expensive, it is foreseeable that EU residents will prefer to holiday at home. But this disproportionate impact of the EU’s scheme may not be sufficient for the scheme to amount to ‘no less favourable treatment’ for those services and service suppliers.
Based on the different interpretations given to this test above, it can be said that the measure comes close to a denial of the favourable treatment, in the sense that the foreign service suppliers are unable to do anything to obtain the more favourable treatment. Indirectly, of course, they could subsidise the compliance costs of EU tourists. In such a case, the question might be whether this is a reasonable burden for them to assume. Arguably, it is not.

Footnote 10
This is not quite the end of the analysis. Article XVII is subject to a footnote 10, which states that:
[s]pecific commitments assumed under this Article shall not be construed to require any Member to compensate for any inherent competitive disadvantages which result from the foreign character of the relevant services or service suppliers.
It might appear that footnote 10 protects the EU’s scheme. However, as the Panel in Canada – Automobiles said, footnote 10 ‘does not provide cover for actions which might modify the conditions of competition against services and service suppliers which are already disadvantaged due to their foreign character’. In the context of Mode 2 services, footnote 10 protects the EU from having to subsidise the costs of international transportation of consumers. However, it does not, of itself, permit the EU to add to these costs. It is of no further relevance to the current issue.

The result of this analysis is that there are at least certain risks that EU’s scheme provides less favourable treatment to foreign services and service suppliers in sectors in which the EU has made commitments, in violation of Article XVII:1 GATS.

3.4. Exceptions for environmental reasons (Article XIV(b) GATS)
Even if the EU’s scheme encounters the legal difficulties described, its GATS-illegal aspects may be justified under Article XIV GATS. This provision is similar to Article XX GATT in most respects, although, notably, it does not include an equivalent to Article XX(g) GATT. This means that the only cover for the EU’s scheme is Article XIV(b) GATS, which permits measures ‘necessary’ ‘to protect human, animal or plant life or health’ (like Article XX(b) GATT).
It may be assumed that the EU’s scheme makes a ‘material contribution’ to the protection of ‘human, animal or plant life or health’, which is a threshold requirement for this provision. But whether it is otherwise ‘necessary’ is a more difficult question. According to the traditional understanding of this test, the EU would have to show that it could not achieve its objective by an alternative measure that is (a) reasonably available and (b) less trade restrictive. In a recent case, the Appellate Body said that it is the WTO-inconsistent aspects of the measure that must be judged ‘necessary’ according to this test, rather than the measure as a whole.
The question is therefore whether there is another measure, or version of the EU’s scheme, that is reasonably available and that would achieve these given objectives in a manner less inconsistent with the GATS. It is at this point that alternatives would need to be discussed, such as an international air passenger levy, or a per-plane tax. One would imagine that it would also be relevant, in determining whether an alternative measure is reasonably available, whether that measure is permitted under international law. It is difficult to give an answer to this question without a full consideration of these alternatives.
But even if the EU’s scheme were considered ‘necessary’, it would need to meet the criteria in the Chapeau of Article XIV. For the reasons mentioned above in the context of Article XX, if the possibility of indirect flights, which reduce the compliance costs, varies according to destination, there may be discrimination between countries where the same conditions prevail. Nor would such discrimination be justified, for the reasons mentioned above. On the other hand, any exemption for flights from countries that have adopted ‘equivalent measures’ could be justified on the basis that the EU is treating countries where the same conditions prevail (itself and those countries) in an equal manner.

4. Conclusions
The overall conclusions of this analysis are somewhat uncertain. In part, this is because the law has not yet evolved clear standards to deal with the legal questions generated by the EU’s scheme; in part it is because the impacts of the measure, and its final design are unknown. Nonetheless, the EU’s scheme may be predicted to have an impact, even if modest and even if only potential, on exports of goods and services (particularly tourism) to the EU, and this is sufficient to engage the EU’s WTO obligations. It is irrelevant at this stage whether the EU’s scheme has an environmental purpose.
Trade in goods under the GATT
This paper first considered the nature of the EU’s scheme in the terms relevant for much of its later analysis under the GATT. The key question at this stage was whether the scheme should be seen as a tax, a charge, or a regulatory measure. On the basis that the scheme functions by granting tradable rights to airlines, and supported by various other considerations, including a recent EU Advocate General’s Opinion to the same effect, it was concluded that it is best seen as a regulatory measure entailing various compliance costs.
The next question concerned the question whether it could be seen as an internal measure governed by Article III:4 GATT. The answer to this question was that it could, insofar as the scheme applied to flights on which imported products were transported; and also under the Note to Article III, to flights once they entered EU airspace. However, to the extent that the scheme applied to flights outside of EU airspace, it could not be considered an internal measure solely applicable to imported products. The paper further concluded that, to the extent that it was applicable, and depending on certain legal and factual variables, there could be a violation of Article III:4. It also concluded that external aspects of the EU’s scheme were both governed by and inconsistent with Article XI:1.
The paper also considered the application to the EU’s scheme of the unconditional most favoured nation obligation in Article I:1. It concluded that this provision did apply, and while it was difficult to draw general conclusions, there would be a certain violation of this provision if the EU granted selective exemptions to airlines from countries that have adopted ‘equivalent measures’. On the other hand, the paper also found that Article V does not pose any problems for the EU’s scheme.
Finally, the paper considered the application of the exception in Article XX(g) GATT to the EU’s scheme, to the extent that it does violate any substantive GATT provisions. The conclusions were that the scheme is provisionally justified under Article XX(g), but it may violate the Chapeau due to the fact that it does not cover all flights, nor all means of transport. It does not matter that the EU’s scheme might be designed in this way to comply with its international law obligations. First of all, it is not clear that these obligations constrain the EU from regulating more generally, and second, this would not be a valid excuse in any case. This reasoning does not however apply to any violation of Article I:1 GATT resulting from exemptions for airlines from countries that have adopted ‘equivalent measures’, which would be consistent with the Chapeau.

Trade in services under the GATS
The paper next considered the legality of the EU’s scheme under the GATS. In this context, a threshold question concerned the application of the GATS to the scheme, given the Air Transport Annex, which establishes certain carve-outs for measures affecting air transport services. A detailed analysis of this showed that, in fact, the substantive carve outs necessarily apply to the EU’s scheme. The EU’s scheme does not appear to affect ‘traffic rights’, and it is very uncertain whether it affects trade in ancillary services, such as ground handling. On the other hand, even if the GATS applies, it is likely that dispute settlement is precluded by the exhaustion of remedies rule in paragraph 4 of the Annex.
The paper proceeded to consider the substantive legality of the EU’s scheme in light of its GATS obligations, and the relevant exceptions. It began by considering the most favoured nation obligation in Article II:1 GATS, and arrived at similar results to those in connection with Article I:1 GATT. It also found a possible, though not certain, inconsistency with the national treatment provision in Article XVII GATS. Furthermore, if there is any such violation, because the EU’s scheme does not apply to all flights or other means of transport, the scheme would probably not be justified under Article XIV, with the exception of exemptions for airlines from countries that have adopted ‘equivalent measures’. Nonetheless, because dispute settlement under the GATS is unlikely, in practice this will not be a significant constraint on the EU’s scheme.

Implications for the WTO
The EU’s scheme raises a number of relatively novel and difficult questions under WTO law. To some extent, it seems likely to violate various WTO provisions. However, by and large the scheme would be protected under the general exceptions to GATT and (if this agreement applies) to GATS. Its main vulnerability – ironically – stems from the fact that it both goes too far, and not far enough. If the EU’s scheme were limited to EU airspace, it would most likely be justified; the same would be true if it applied to all means of transport around the world. The former option is of course possible, and may even be required as a result of diplomatic and legal pressure in other forums. But it is the latter option that raises more difficult questions for the WTO. The WTO Appellate Body has said that international law constraints are no excuse for discrimination: in some cases this can mean that a WTO Member is effectively required to break the law. This is unsustainable and, given the work of the International Law Commission on these issues, largely unnecessary. If the EU’s scheme does end up being challenged in WTO dispute settlement proceedings, it is to be hoped that the Appellate Body takes the opportunity to ensure that legitimate international law obligations are taken into account in determining whether a WTO Member has violated its WTO obligations.