16.3 PP 115 195 Disciplines On Subsidies

4
Disciplines on subsidies
Evidently, and importantly, the SCM Agreement does not condemn all
specific subsidies covered under Articles 1 and 2. The Appellate Body,
recognizing the broad subsidy definition, stressed:
The granting of a subsidy is not, in and of itself, prohibited under the
SCM Agreement. Nor does granting a ‘subsidy’, without more, constitute
an inconsistency with that Agreement. The universe of subsidies is vast.
Not all subsidies are inconsistent with the SCM Agreement.1
Articles 1 and 2 merely define the concepts of subsidy and specificity,

serving as a threshold for the application of the disciplines prescribed by
Parts II, III, IV, and V of the SCM Agreement.2 The SCM Agreement aims at
targeting those subsidies that are trade-distorting. To this end, the agree-
ment as originally implemented grouped subsidies in three categories, each
imposing different disciplines.3 This so-called ‘traffic light approach’ proved
pivotal to finding a compromise between the harsh stance of the United
States and the looser stance of other countries on disciplining subsidies. At
one end of the scale, two types of subsidy were principally prohibited per se
(red light) because of their direct trade-distorting effect, namely export
subsidies and local content subsidies (Part II of the SCM Agreement).4 At
the other end, three types of subsidy, namely for research activities, for
disadvantaged regions, and for the adaptation to environmental require-
ments, were deemed non-actionable (green light); these were in principle
allowed (Part IV of the SCM Agreement). All other subsidies were
1
Appellate Body Report, Canada – Aircraft (Article 21.5 – Brazil), para. 47.
2
See also Appellate Body Report, US – FSC (Article 21.5 – EC), paras. 85–87. Hence
Articles 1 and 2 cannot be considered to be ‘violated’, as the panel in US – Export
Restraints (para. 9.1) seemed to suggest.
3
The origins of the traffic light approach date back to US proposals during the Tokyo
Round. See Hufbauer and Shelton Erb, above Chapter 2 n. 12, at 22; Croome, above
Chapter 2 n. 49, at 62.
4
See, for instance, Panel Report, US – Large Civil Aircraft, paras. 7.1808–7.1810.
115
https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

116 subsidies and countervailing measures
actionable subsidies (amber light), meaning that they could be challenged if

they caused adverse effects (Part III of the SCM Agreement).
However, the category of green light subsidies expired at the end of
1999.5 Since then, just two categories have been in place: a ‘line [is] drawn’6
between those subsidies that are prohibited per se (red light) and all
others, which can be challenged if they cause adverse effects on other
members (amber light).7 Specific subsidies previously granted green light
status are thus currently actionable if they cause adverse effects. The
category of prohibited subsidies circumvents the difficult proof of adverse
effects, and more powerful tools are provided to challenge such subsidies.
The analysis starts with those subsidies that are captured under this
prohibition set out under Part II of the SCM Agreement.
4.1 Prohibited subsidies

Article 3 of the SCM Agreement prohibits export subsidies and local
content subsidies because they are considered trade-distorting by their
very nature.8 Both types are examined in turn.
4.1.1 Export subsidies

Article 3.1 of the SCM Agreement reads in the relevant part:
Except as provided in the Agreement on Agriculture, the following
subsidies, within the meaning of Article 1, shall be prohibited:
(a) subsidies contingent, in law or in fact, whether solely or as one of
several other conditions, upon export performance, including those
illustrated in Annex I;9
5
Article 31 of the SCM Agreement.
6
Appellate Body Report, EC – Large Civil Aircraft, para. 1054.
7
In addition to these substantive disciplines, members have to notify all specific subsidies
(Article 25 of the SCM Agreement), but this requirement is increasingly left unfulfilled.
The percentage of members failing to make any notification increased from 28 per cent in
1995 to 46 per cent in 2011 (G/SCM/W/546/Rev.4, 16 April 2013, para. 8).
8
Expansion of the red light category beyond the Illustrative List of Export Subsidies of the
Subsidies Code was one of the most contentious issues during the Uruguay Round, with
the United States as its main proponent (MTN.GNG/NG10/W/20, June 15, 1988, at 3–4).
In the Doha negotiations, the United States and the EU unsuccessfully proposed an expansion
of the red light subsidies beyond the existing two types. Whereas the United States mainly
aimed to prohibit the so-called ‘dark amber’ types of domestic subsidy, the EU aimed at
prohibiting dual pricing strategies (TN/RL/W/78, 19 March 2003; TN/RL/GEN/94, 16
January 2006; TN/RL/GEN/146, 5 June 2007; TN/RL/GEN/135, 24 April 2006).
9
Footnotes in original omitted.

disciplines on subsid ies 117
The opening part of Article 3.1 installs an exception on agricultural export

subsidies that are in conformity with the Agreement on Agriculture. In
Chapter 6, whether such export subsidies are still exempted from this
prohibition set out in the SCM Agreement will be examined.10 Turning to
non-agricultural subsidies, these are prohibited if ‘contingent . . . upon
export performance, including those illustrated in [the Illustrative List]’.11
Two undisputed conclusions can be derived from the text of Article
3.1(a) regarding its relationship with the Illustrative List (Annex I).12
First, the Illustrative List, by definition, only lists ‘examples – illustra-
tions’13 of export subsidies. Some subsidies thus fall outside the Illustrative
List’s scope but are covered under the export subsidy definition stipulated in
Article 1 in conjunction with Article 3.1(a) of the SCM Agreement.14
Second, a subsidy that falls within the scope of the Illustrative List is
deemed to be a prohibited export subsidy.15 Consequently, the com-
plaining party does not first have to demonstrate that these subsidies are
contingent on export performance within the meaning of Article 3.1(a)
of the SCM Agreement. Complainants could bypass the export contin-
gency demonstration and directly base their claim on one of the items
of the Illustrative List. But could complainants use the Illustrative List
not only to jump the ‘export contingency’ test of Article 3.1(a) of the SCM
Agreement, but also to bypass the ‘subsidy’ test under Article 1 of the SCM
Agreement? Recall that Article 3.1 reads that ‘the following subsidies, within
the meaning of Article 1, shall be prohibited’.16 Further, the subsidy
definition of Article 1 also leaves no doubt that it applies ‘for the purpose
of this Agreement’, which implies, according to the Appellate Body, that
this ‘applies wherever the word “subsidy” occurs throughout the
SCM Agreement and conditions the application of the provisions of that
10
See below Chapter 6, section 6.2.3.1.
11
S&D treatment on export subsidies is discussed below in Chapter 6, section 6.1.1.1.
For the normative discussion on the disciplines on export subsidies see below Chapter 16,
section 16.2 (developed countries) and Chapter 17, s. 17.3.
12
Panel Report, Brazil – Aircraft (Article 21.5 – Canada), paras. 6.29–6.31.
13
Appellate Body Report, Brazil – Aircraft (Article 21.5 – Canada), para. 61.
14
This follows from the use of the word ‘including’. Some countries had suggested during
the Uruguay Round that it be converted into an exhaustive list, but this was opposed by
others such as the United States and the EC. Stewart, above Chapter 2 n. 49, at 888.
15
Panel Report, Brazil – Aircraft (Article 21.5 – Canada), paras. 6.30–6.31. This was
implicitly endorsed by the Appellate Body Report in Brazil – Aircraft (Article 21.5 –
Canada), para. 61.
16
Emphasis added.

Agreement regarding prohibited subsidies in Part II . . . ’.17 This suggests

that only subsidies within the meaning of Article 1 could be prohibited
under the SCM Agreement and that complainants should not, as a result, be
allowed to jump directly to the Illustrative List.18 On the other hand, this
reading would strongly reduce the relevance of the Illustrative List. Indeed,
the Illustrative List, pre-dating the subsidy definition, provides more guid-
ance on what GATT contracting parties considered to be a ‘subsidy’ than on
the ‘export contingency’ element. By including the Illustrative List in the
SCM Agreement, members apparently agreed that this list should still be
relevant.19 This could explain why several panels decided that complainants
could very well rely directly on the Illustrative List, without the need for
demonstrating the subsidy element.20 Similarly, panels and the Appellate
Body (as well as all parties) in all US – Upland Cotton proceedings agreed
that measures that fall under item (j) of the Illustrative List are per se export
subsidies, without the need for a separate analysis under Articles 1 and 3.21
In sum, complainants can make use of two different tracks to dem-
onstrate that a subsidy is a prohibited export subsidy under the SCM
Agreement: on the basis of Article 1 in conjunction with Article 3 of the
SCM Agreement or by reference to the Illustrative List if the measure
is listed under one of its items. Because the subsidy definition under
both approaches might differ, the question of the correct implementa-
tion could arise: does implementation differ depending on whether
the existence of an export subsidy is demonstrated on the basis of
Article 1 in conjunction with Article 3 or the Illustrative List? Part II
17
Appellate Body Report, US – FSC, para. 93; see also Panel Report, US – Softwood Lumber
IV, para. 7.104.
18
Unless, of course, all items in the Illustrative List are, by definition, beneficial financial
contributions (or income or price support) in the meaning of Article 1.
19
The cost-to-government approach under the Illustrative List (advocated by the EU) is
generally considered to be more flexible than the benefit-to-recipient approach elabo-
rated under Article 1 of the SCM Agreement.
20
Panel Report, Brazil – Aircraft (Article 21.5 – Canada), paras. 6.43–6.44. See also Panel
Report, Korea – Commercial Vessels, para. 7.204. The panel in Korea – Commercial
Vessels observed that this ‘perhaps reflects the historical context of the Illustrative List, in
the sense that it was first drafted before the definition of “subsidy” set forth in the SCM
Agreement was introduced’ (n. 126).
21
Panel Report, US – Upland Cotton, paras. 7.802–7.803, 7.946–7.948 and 8.1(d)(i); Appellate
Body Report, US – Upland Cotton, paras. 666, 720–733 and 763(e)(iv); Panel Report, US –
Upland Cotton (Article 21.5 – Brazil), para. 14.52; Appellate Body Report, US – Upland
Cotton (Article 21.5 – Brazil), paras. 322–323. See also Panel Report, Canada – Dairy (Article
21.5 – New Zealand and US II), para. 5.154.

will explain in detail how this question was solved in the US – Upland
Cotton case.22
Both tracks to meet the export subsidy definition are further explored.
Importantly, as will be explained subsequently, the Illustrative List could
in some instances be used a contrario as well, namely to demonstrate that
some forms of export subsidies (in the meaning of Article 1 in conjunc-
tion with Article 3 or the Illustrative List) are not prohibited under the
SCM Agreement.
4.1.1.1 Export subsidies in the meaning of Article 1

in conjunction with Article 3 of the SCM Agreement:
subsidies ‘contingent’ on exportation
Under this first track, an export subsidy is present if two elements are
demonstrated: a subsidy that is contingent on exportation. The first
element has been explained above. The complainant should demonstrate
the presence of a ‘subsidy’ within the meaning of Article 1: a financial
contribution (or income or price support) by the government confers a
benefit.23 In contrast to some items of the Illustrative List, this subsidy
definition applies a benefit-to-recipient standard and not a cost-to-
government standard. The relevant question under Article 1 is not
whether the government incurred a cost when making the contribution,
but whether the financial contribution was beneficial to the recipient.
Next, the complainant should demonstrate that such a subsidy is
‘contingent’ on exportation. The SCM Agreement explicitly provides
that such contingency could operate not only in law but also in fact,
which prevents governments from circumventing the prohibition by
linking subsidies to export performance without prescribing it explicitly
in their laws.24 In case of de jure export conditionality, the granting of the
subsidy is made legally contingent on export performance, which should
be ‘demonstrated on the basis of the words of the relevant legislation,
regulation or other legal instrument’.25 De facto export conditionality
should demonstrate on the basis of three different elements26 that (i) the
granting of a subsidy27 is (ii) in fact tied to28 (iii) actual or anticipated
22
See below Chapter 10, section 10.2.3.2; Chapter 11, section 11.1.1.2.2.
23
Recall that the specificity test should not be passed because ‘export subsidies’ are deemed
to be specific.
24
Appellate Body Report, Canada – Aircraft, para. 167.
25
Ibid., para. 167. See also Appellate Body Report, Canada – Autos, para. 100.
26
See Article 3.1(a), n. 4 of the SCM Agreement.
27
Appellate Body Report, Canada – Aircraft, para. 170. 28 Ibid., para. 171.

exportation or export earnings.29 The mere fact that a subsidy is granted

to export-oriented industries is considered a relevant, but in itself an
insufficient element, to meeting this standard.30 On the other hand, a
subsidy could qualify as an export subsidy if export performance is one
among other conditions for receiving the subsidy. In other words, export
performance does not have to be a sufficient condition for receiving the
subsidy: other conditions might in addition be required.31 The Appellate
Body explained that the required level of conditionality is similar for de
jure and de facto export subsidies. The only difference between both
types is what evidence may be employed to prove the tie between the
subsidy and export performance.32
The required strength of the tie between the subsidies and export
performance, which is ‘at the very heart of the legal standard’, has been
further clarified in the case law.33 Exploring de facto export contingency,
the Appellate Body in Canada – Aircraft elaborated on the three ele-
ments of the legal standard. First, the ‘granting of the subsidy’ element
shows that the inquiry is on the granting authority and not on the
recipient. Hence the relevant question is whether the granting authority
imposed a condition based on export performance in providing the
subsidy.34 Second, the ‘tied to’ element (i.e., contingency) shows that ‘a
relationship of conditionality or dependence must be demonstrated’.35
Third, regarding actual or anticipated exportation, the dictionary mean-
ing of ‘anticipated’ is ‘expected’, but it does not imply that a subsidy is
export-contingent merely if the government expects future exports.
In EC – Large Civil Aircraft the Appellate Body further explained how
to establish that a subsidy is ‘in fact tied to . . . anticipated exportation’.
First of all, whereas ‘actual exportation’ refers to exportation that has
occurred at the time a subsidy is granted, ‘anticipated exportation’ means
exportation that is expected by the granting authority (and thus not the
recipient) at the time a subsidy is granted. It is, therefore, not required
that the anticipated exportation effectively takes place after the subsidy is
granted. To demonstrate that subsidies are ‘tied to’ such anticipated
exportation, the following test has to be conducted: ‘is the granting of
29
Ibid., para. 172.
30
Article 3.1(a), n. 4 of the SCM Agreement; Appellate Body Report, Canada – Aircraft,
para. 173.
31
Article 3.1(a), (b) of the SCM Agreement; Appellate Body Report, US – FSC (Article
21.5 – EC), para. 111.
32
Appellate Body Report, Canada – Aircraft, para. 167. 33 Ibid., para. 171.
34
Ibid., para. 170. 35 Ibid., para. 171.

the subsidy geared to induce the promotion of future export perform-

ance by the recipient?’36 This standard is only met if the subsidy is
designed to increase the firm’s ratio of export sales to domestic sales.
Assuming that the recipient’s existing export/domestic ratio at the time
of the subsidy is 2:3, a subsidy designed to increase this firm’s future
production by five units would not be tied to anticipated exportation if
the firm is not expected to export more than two of the additional five
units.37 Hence the mere fact that the subsidy increases the recipient’s
export sales is insufficient to label it as an export subsidy.38 Put differ-
ently, export-contingent subsidies should ‘favour a recipient’s export
sales over its domestic sales’, which occurs if they are geared to induce
a recipient ‘to export in a way that is not simply reflective of the
conditions of supply and demand in the domestic and export markets
undistorted by the granting of the subsidy’.39 The Appellate Body further
explained how this standard has to be demonstrated:
[It] ‘must be inferred from the total configuration of the facts constituting
and surrounding the granting of the subsidy’, which may include the
following factors: (i) the design and structure of the measure granting the
subsidy; (ii) the modalities of operation set out in such a measure; and
(iii) the relevant factual circumstances surrounding the granting of the
subsidy that provide the context for understanding the measure’s design,
structure, and modalities of operation.40
If relevant evidence is present, it could involve a comparison between

the ratio of anticipated export and domestic sales of the subsidized
product and the ratio of such sales in the absence of the subsidy. The
latter export/domestic sales ratio could either be derived from historical
sales by the recipient or from the sales of a hypothetical profit-
maximizing firm in absence of the subsidy. An export subsidy is present
if the anticipated export/domestic sales ratio is higher than the one
derived from historical/hypothetical sales.41 The Appellate Body appears
to suggest that the anticipated export/domestic sales ratio could not
be derived from the actual sales ratio after subsidization because the
36
Appellate Body Report, EC – Large Civil Aircraft, para. 1044. 37 Ibid., para. 1048.
38
The standard is not met merely because ‘the granting of the subsidy is designed to
increase a recipient’s production, even if the increased production is exported in whole’
(ibid., para. 1045).
39
Ibid., paras. 1053, 1045, 1098.
40
Ibid., para. 1046 (emphasis added; footnote in original omitted). 41 Ibid., para. 1047.

assessment must be based on the information available to the granting

authority at the time the subsidy is granted.42
In developing this new standard, the Appellate Body firmly rejected
the broader export contingency interpretation advanced by the panel, in
which an export subsidy would present if the subsidy was granted simply
because the granting authority anticipated export performance.43 In
substance this standard was too broad, because it would capture subsi-
dies that might boost export sales without giving an incentive to export
rather than to sell domestically. Regarding the demonstration of export
contingency, the Appellate Body equally faulted the panel’s reliance on
the subjective motivation of the granting authority. Even if the granting
government is motivated to boost exports at the expense of domestic
sales, this is insufficient to satisfy the de facto export contingency stand-
ard. The conditional relationship must be based on ‘objective evidence,
rather than subjective intent’.44 At the same time, the Appellate Body
acknowledged that ‘objectively reviewable expressions of a government’s
policy objective for granting a subsidy’ may constitute relevant evi-
dence.45 But ascertaining the reasons for granting the subsidy is neither
sufficient nor strictly necessary to demonstrate export contingency.46
In sum, the Appellate Body advanced a rather narrow interpretation
of de facto export contingency that seems to resonate with the basic
definition of an export subsidy in economic theory. Such subsidies
induce the recipient to export rather than to sell domestically and thus
create a gap between the subsidized export price and domestic price.
Supporting an export-oriented firm to boost its general sales level would
therefore not be prohibited under the SCM Agreement but would be
actionable if it causes adverse effects. The Appellate Body deemed its
approach consistent with previous jurisprudence, in particular with
Canada – Aircraft.47 The original Technology Partnerships Canada
(TPC) programme was correctly found to be export-contingent by the
panel because it contained selection criteria relating to exportation (e.g.,
whether projects would increase export sales of the funded firms) and
stated export promotion as the objective. The revised TPC programme,
in which both the selection criteria and objectives were removed, did not
qualify as an export subsidy because it was no longer geared to induce the
promotion of future export performance. At the same time, the new
export contingency standard seems stricter than the standard adopted in
42
Ibid., para. 1049. 43 Ibid., paras. 1062–1063. 44 Ibid., para. 1050.
45
Ibid., paras. 1050, 1051, 1064. 46 Ibid., para. 1051. 47 Ibid., para. 1055.

some previous case law. For instance, the panel in Australia –

Automotive Leather II (Article 21.5 – US) concluded that Australia’s
grant to Howe was an export subsidy partly because Howe had to
continue, and perhaps even increase, exports in order to reach the sales
performance targets, which in the panel’s view turned these targets into
export performance targets.48 Paraphrasing the Appellate Body, if ‘the
granting of the subsidy is designed to increase a recipient’s production,
even if the increased production is exported in whole’,49 this is insuffi-
cient to meet the export contingency standard: this boost in production
must favour export sales over domestic sales.50
Previous case law on de jure export contingency of subsidy pro-
grammes seems to have detected export subsidy programmes that
boosted export sales in favour of domestic sales. In Canada – Autos,
a manufacturer could receive a certain amount of import duty exemp-
tions (subsidies) even if he or she did not export. Yet this subsidy was
de jure export-contingent because ‘the more motor vehicles a manufac-
turer exports [export performance], the more motor vehicles that
manufacturer is entitled to import duty-free [subsidy]’.51 In US – FSC
(Article 21.5 – EC), the United States contended that export contingency
should be understood as a necessary condition, and therefore argued that
the ETI measure was export-neutral. After all, the ETI measure granted a
tax exemption (subsidy) when goods are sold for use abroad and was thus
available with respect to goods not produced in the United States.52 Yet
the Appellate Body reached a different conclusion, by distinguishing two
situations. First, when goods are produced domestically and sold for use
abroad, exporting is obviously a necessary condition for receiving the tax
exemption and the tax benefit is thus contingent on export.53 The
Appellate Body did not decide on the second situation, whereby goods
48
Panel Report, Australia – Automotive Leather II (Article 21.5 – US), para. 9.67.
49
50
This case raises the interesting question of whether the existing export/domestic sales
ratio at the time the subsidy is granted is still relevant if it is affected by past export
subsidies. Due to previous export subsidies, Howe exported the ‘overwhelming majority’
of its sales. Is the challenged subsidy, which replaced these export subsidies, contingent
on exportation if it is designed not to improve further this inflated export/domestic sales
ratio but merely to keep it stable? The panel’s ultimate conclusion was arguably correct if
the ratio of export/domestic sales undistorted by previous export subsidies would be
considered as benchmark.
51
Appellate Body Report, Canada – Autos, para. 106 (emphasis added).
52
Appellate Body Report, US – FSC (Article 21.5 – EC), para. 110.
53
Ibid., paras. 116–118.

are produced abroad and sold for use abroad, which clearly does not
involve any exportation from the United States.54 Importantly, the fact
that the subsidies granted in the second situation might not be export-
contingent ‘does not dissolve the export contingency’ arising in the first
situation55 because it concerns two different factual situations.56 The
crux of the argument seems to be that both situations were different
because they were mutually exclusive. For property produced in the
United States, exportation is the only option in order to benefit from
the tax exemption. Whereas export contingency is thus not a necessary
condition for receiving the subsidy in the light of the ETI measure as a
whole, the ETI measure with respect to the first situation was considered
export-contingent.57
4.1.1.2
Export subsidies in the meaning of one
of the items of the Illustrative List
Alternatively, measures are prohibited ‘export subsidies’ if they are
covered by one of the items of the Illustrative List, which is basically
the same as the list included in the Subsidies Code.58
Three listed examples of export subsidies do not need much explan-
ation. They either do not add anything to the Article 1 in conjunction
with Article 3 export subsidy definition or are nowadays insignificant.
An example of the first category is the opening item of the list, which
refers to the provision by governments of ‘direct subsidies’ to a firm or an
industry contingent on export performance (item (a)). Likewise, the last
item of the list, which was included as a residual category in the Subsidies
Code, seems redundant because it refers to any other charge on the
public account constituting an export subsidy in the sense of Article
XVI of the GATT (item (l)).59 Lastly, currency retention schemes or
any similar practice that involves a bonus on exports are prohibited
(item (b)).60 Such multiple currency practices, for which IMF approval
54
Ibid., paras. 109 and 120. 55 Ibid., para. 119. 56 Ibid., paras. 113–115, 119.
57
See also Appellate Body Report, US – Upland Cotton, paras. 556–584.
58
The list was based on the 1960 Declaration (see above Chapter 2, section 2.2).
59
Luengo explained that this item was included in the Subsidies Code because of the
different interpretation of ‘subsidy’ between the US (focusing on benefit) and the EU
(focusing on the cost to government or charge on the public account). Luengo, above
Chapter 3 n. 139, at 151.
60
Siegel indicated that a currency retention scheme ‘usually involves allowing certain
exporters to retain a portion of their foreign exchange earnings notwithstanding a
general rule for residents to surrender receipts of foreign exchange to local banks, or
the central bank, in exchange for local currency’. As Dam explained, such a currency

is required for IMF members, were prevalent when the list was originally
drafted in the 1950s but are almost non-existent today.61
The remaining items on the Illustrative List could be grouped into
three categories, each relating to one of the three types of financial
contribution singled out under Article 1 of the SCM Agreement. This
typology is made only for analytical purposes and does not imply that
those items would ipso facto be covered under Article 1.
4.1.1.2.1 The provision of goods or services favourable to

exporters Two items of the Illustrative List relate to the provision of
goods or services favourable to exporters. First, the government is not
allowed to subsidize exports by providing or mandating internal trans-
port and freight charges on export shipments on terms more favourable
than for domestic shipments (item (c)).62,63 The second type is broader
and requires some further discussion: item (d) in essence prohibits
governments from providing products (or services) on more favourable
terms when these are used in export production. The panels in Canada –
Dairy distinguished three elements that should be present in order for a
violation of item (d) to be found: ‘(1) the provision of products for use in
export production on terms more favourable than for provision of like
products for use in domestic production; (2) by governments either
retention scheme could form an important incentive to export because, depending on

the details of the scheme, such foreign exchange earnings could either be sold by
exporters on the free market, where, because of the scarcity of foreign exchange, they
could charge a higher price than under the official exchange rate, or they could be used to
import goods that were otherwise unavailable under the applicable exchange control
regulation. Such currency retention schemes involving a bonus on exports are thus
prohibited export subsidies. D. E. Siegel, ‘Legal Aspects of the IMF/WTO Relationship:
The Fund’s Articles of Agreement and the WTO Agreements’, 96:3 American Journal of
International Law (2002), 561–99, at 596; K. W. Dam, The GATT: Law and International
Economic Organization (University of Chicago Press, 1970), at 137–8.
61
The relevance of this item was already being questioned during the 1970s (Spec(72)61,
12 July 1972, at 2).
62
See Dam, above n. 60, at 140. Because the exporter receives a service in return for its
contribution, such a ‘charge’ should not be considered a ‘tax’, and item (c) is therefore
not discussed as an example of revenue forgone. The concept of ‘tax’ is not defined in any
of the WTO Agreements. For a definition of the terms ‘tax’ and ‘charge’ in the context of
the OECD, see Note by the Chairman, Definition of Taxes (DAFFE/MAI/EG2(96)3, 19
April 1996).
63
The comparison is thus made between charges for export shipments and those for
domestic shipments. As the panel in Brazil – Aircraft underscored, this is ‘irrespective
of whether those charges are higher, lower or equal to the charges paid with respect to
the shipments of competing products from other Members’. Panel Report, Brazil –
Aircraft, para. 7.25.

directly or indirectly through government mandated schemes; and

(3) on terms more favourable than those commercially available on
world markets’.64
Item (d) thus explicitly covers the indirect provision of goods or services
through government-mandated schemes. It equally covers situations where
the government delegates authority to its agency, ‘which, in turn, sets up a
“government-mandated” scheme’.65 Yet one might wonder whether the
required government nexus under item (d) is looser than under the entrust-
ment/direction standard stipulated in Article 1.1(a)(1)(iv) of the SCM
Agreement.66 In Canada – Dairy (Article 21.5 – New Zealand and US II),
this seemed to be the parties’ assumption, since the complainants based
their argumentation on item (d) of the Illustrative List, whereas Canada
argued that the scope of item (d) should be interpreted in the light of
Article 1.1(a)(1)(iv) of the SCM Agreement.67 The panel, however, followed
the complainants’ view because the Illustrative List contains per se viola-
tions. The panel therefore disregarded the entrustment/direction standard
of Article 1.1(a)(1)(iv).68 It concluded that ‘the provision of goods is made
or mandated by government for export as a result of the prohibition on
diversion of [Commercial Export Milk (CEM)] back into the domestic
regulated market and the exemption which gives processors for export
access to the lower CEM prices’.69 In response to previous rulings,
Canada had attempted to remove government action from the export
transactions by introducing a new category of milk for export processing,
CEM. This CEM was exempted from the pricing regulations applicable to
other types of milk. Price and volume of CEM were negotiated directly
between processor and producer, and the diversion of CEM on to the
domestic market was prohibited. Canada’s argument, based on Article 1.1
(a)(1)(iv), was that ‘there is no provision of products through a government
mandated scheme because the government does not command or direct
producers to produce CEM’, but this was simply considered irrelevant.
64
Panel Report, Canada – Dairy (Article 21.5 – New Zealand and US II), para. 5.157. Panel
Report, Canada – Dairy, paras. 7.128–7.131.
65
Panel Report, Canada – Dairy, para. 7.130.
66
Regarding the scope of Article 1.1(a)(1)(iv) of the SCM Agreement, see above Chapter 3,
section 3.1.2.
67
For background to this case see below Chapter 6, section 6.2.1.2.1.1.2.
68
The panel relied on item (d) of the Illustrative List to interpret the meaning of ‘export
subsidy’ under Article 10.1 of the Agreement on Agriculture. Panel Report, Canada –
Dairy (Article 21.5 – New Zealand and US II), paras. 5.153, 5.154, 5.159, 5.160.
69
Ibid., para. 5.160. The panel also found support in Article 9.1(c) of the Agreement on
Agriculture.

Although the Appellate Body did not go into the interpretation of item
(d),70 it seemed to qualify the panel’s wide interpretation of this item of the
Illustrative List. The Appellate Body explicitly contrasted the loose govern-
ment nexus under the listed types of agricultural export subsidy (i.e., Article
9.1(c) of the Agreement on Agriculture) with both Article 1.1(a)(1)(iv) of
the SCM Agreement and item (d) of the Illustrative List.71 Drawing this
comparison, the Appellate Body observed that, even though Article 9.1(c) of
the Agreement on Agriculture ‘certainly covers situations where govern-
ment mandates or directs that payments be made, it also covers other
situations where no such compulsion is involved’.72 A contrario, some
kind of compulsion of a third party seems required not only under Article
1.1(a)(1)(iv) but also under item (d) of the Illustrative List.73 Whereas the
Appellate Body did not equate the required government nexus under Article
1.1(a)(1)(iv) and item (d) of the Illustrative List, its conclusion implies that
both have at least in common the fact that they necessitate some form of
compulsion in case a subsidy is provided through a third party.74
4.1.1.2.2 Border tax adjustments and duty drawback systems Five

examples of export subsidies in the Illustrative List deal with fiscal
incentives (items (e), (f), (g), (h), (i)) by which the government forgoes
revenue otherwise due by exporters. As the following two sections
explain, these items address the conditions under which taxes as well
as import duties could be rebated on exportation.
4.1.1.2.2.1 Border tax adjustments

4.1.1.2.2.1.1 Introduction Border tax adjustments (BTAs) encom-
pass the imposition of a tax on imported products equal to a correspond-
ing tax on domestic products (import side) and the rebate of such a tax
70
Appellate Body Report, Canada – Dairy (Article 21.5 – New Zealand and US II), para. 158.
71
The Appellate Body also referred to other items of the Illustrative List.
72
Appellate Body Report, Canada – Dairy (Article 21.5 – New Zealand and US II), para.
128 (emphasis added).
73
Arguably, the Canadian scheme was such a situation in which no compulsion was
involved. Hence it would seem to fall outside the scope of item (d) of the Illustrative
List and Article 1.1(a)(1)(iv). The required nexus between the government and the
payment was sufficient to bring it within the scope of Article 9.1(c) of the Agreement
on Agriculture (see below Chapter 6, section 6.2.1.2.1.1).
74
Referring inter alia to item (d) of the Illustrative List, the Appellate Body stated that ‘In
these provisions, some kind of government mandate, direction, or control is an element
of a subsidy provided through a third party.’ Appellate Body Report, Canada – Dairy
(Article 21.5 – New Zealand and US II), para. 128, n. 113. Recall that the concept of
‘direction’ is interpreted broadly by the Appellate Body (see above Chapter 3 n. 113).

on exported products (export side). Such tax adjustments put into effect
the destination principle, in which products are taxed only in the country
of consumption.75 These adjustments would generate trade neutrality:
they level the playing field between products from different countries
with regard to such taxes. Adjustments by all trading countries applied
symmetrically on the import and export side ensure that products are
not double taxed but could likewise not take advantage of low taxes in
their country of origin.76
However, the GATT/WTO disciplines only allow for BTAs on indirect
taxes, which are imposed directly or indirectly on products, but not on
direct taxes, which are considered to be imposed on the producer.
Framed differently, the destination principle is installed with regard to
indirect taxes, whereas the origin principle is applied to direct taxes.
Direct taxes are imposed in the country of production and cannot be
adjusted on either the import or export side. The rationale underpinning
this distinction is twofold. First, indirect taxes are considered to be
shifted forward fully by the taxpayer into the final price charged to
consumers (i.e., consumption tax), whereas direct taxes would be com-
pletely shifted backward and thus be borne by the producer (i.e., income
tax). Direct taxes would therefore not be reflected in the final price of the
product.77 Already during the GATT era, however, GATT contracting
parties understood that this proposition does not entirely hold because,
for instance, direct taxes could very well be reflected in the final price.78
Second, this dichotomy might be partly explained on the basis of tradi-
tion: it codified GATT contracting parties’ practices to adjust only for
indirect taxes and not for direct taxes.79
Despite its questionable economic underpinning, voiced particularly
by those countries such as the United States relying predominantly on
direct taxes, this dichotomy is still in place. Members have the right – but
not the obligation – to adjust indirect taxes on the import and/or export
side, and not even symmetric application on both sides is required. On
75
See, e.g., L/3464, para. 4.
76
See, e.g., P. Demaret and R. Stewardson, ‘Border Tax Adjustments under GATT and EC
Law and General Implications for Environmental Taxes’, 28:4 Journal of World Trade
(1994), 5–64, at 6.
77
See WT/CTE/W/47, 2 May 1997, para. 36.
78
See, e.g., L/3464, 20 November 1970, para. 8; J. Bhagwati and P. Mavroidis, ‘Is Action
against US Exports for Failure to Sign Kyoto Protocol Legal?’, 6:2 World Trade Review
(2007), 299–310, at 306.
79
See, e.g., L/3464, para. 8; Dam, above n. 60, at 139.

the import side, Article II:2(b) in conjunction with Article III:2 of the
GATT allows members to impose an equivalent indirect tax on imported
products.80 More relevantly for our discussion, several items of the
Illustrative List elaborate on the Note Ad Article XVI of the GATT,
which allows adjustments of indirect taxes on the export side.81 During
the Doha Round negotiations, the United States urged once more the
stronger equalization of the treatment of direct and indirect taxation, but
it seems unlikely that this call will be answered.82
4.1.1.2.2.1.2 Direct taxes Item (e) installs the origin principle for direct
taxes and social welfare charges.83 Their full or partial exemption ‘specifi-
cally related to exports’ qualifies as a prohibited export subsidy.84
Nonetheless, this prohibition does not intend to constrain a member from
taking measures ‘to avoid the double taxation of income earned by a
taxpayer of that member in a “foreign” State’.85,86 The United States
invoked this exception formulated in footnote 59 as a defence with regard
80
The import side of border tax adjustments is not relevant for our discussion on export
subsidies.
81
For an overview of the negotiating history, see G. C. Hufbauer, Fundamental Tax Reform
and Border Tax Adjustments (Washington, DC: Institute for International Economics,
January 1996), at 47–50; Hufbauer and Shelton Erb, above Chapter 2 n. 12, at 51–7.
82
See TN/RL/W/78, 19 March 2003, at 5. The Draft Consolidated Chair Text (above
Chapter 2 n. 113) does not amend the relevant provisions.
83
Direct taxes cover, by virtue of footnote 59 of the SCM Agreement, ‘taxes on wages,
profits, interests, rents, royalties, and all other forms of income, and taxes on the
ownership of real property’. At the time of the Working Party Report on Border Tax
Adjustments (the 1970s), it was still being debated whether taxes on property and social
security charges could be rebated. Yet the SCM Agreement clarifies that both are
excluded from adjustment on exportation.
84
A deferral is not an export subsidy where, for example, appropriate interest charges are
collected. See footnote 59 of the SCM Agreement; Appellate Body Report, US – FSC,
para. 97. Also considered prohibited is ‘the allowance of special deductions directly
related to exports or export performance, over and above those granted in respect to
production for domestic consumption, in the calculation of the base on which direct taxes
are charged’ (item (f), emphasis added).
85
See footnote 59 of the SCM Agreement, in fine as interpreted by the Appellate Body in US –
FSC (Article 21.5 – EC), paras. 137–138. Double taxation occurs ‘when the same income, in
the hands of the same taxpayer, is liable to tax in different States’, and foreign-source income
‘refers to income generated by activities of a non-resident taxpayer in a “foreign” State which
have such links with that State so that the income could properly be subject to tax in that
State’. Appellate Body Report, US – FSC (Article 21.5 – EC), paras. 137, 145.
86
In footnote 59, members also ‘reaffirm the principle that prices for goods in transactions
between exporting enterprises and foreign buyers under their or under the same control
should for tax purposes be the prices which would be charged between independent enter-
prises acting at arm’s length’. See also Appellate Body Report, US – FSC, paras. 96–100.

to the ETI measure. The Appellate Body clarified that this constitutes an
‘affirmative defence’ that may justify a prohibited export subsidy and that
the burden of proof is on the defending party.87 Yet the United States failed
to meet this burden because the ETI exemption improperly combined
domestic-source income and foreign-source income.88 If the measure
were ‘confined to those aspects which grant a tax exemption for “foreign-
source income”, it would fall within footnote 59’.89
4.1.1.2.2.1.3 Indirect taxes Regarding indirect taxes, the conditions

for applying BTAs on the export side are specified in items (g) and (h) of
the Illustrative List. Item (g) sets out the general rule: ‘the exemption or
remission, with respect to the production and distribution of exported
products, of indirect taxes’ is considered a prohibited export subsidy only
if it is ‘in excess of those levied with respect to the production and
distribution of like products when sold for domestic consumption’.90
Implementing the destination principle, this provision prohibits only
excessive adjustment of indirect taxes on exported goods. Indirect taxes
form the residual category of taxes under the SCM Agreement, covering
‘sales, excise, turnover, value added, franchise, stamp, transfer, inventory
and equipment taxes, border taxes and all taxes other than direct taxes
and import charges’.91 Evidently, not only sales taxes, collected at the
point of final sale, but also value added taxes (VAT), which are collected
at every stage of production and distribution on the value added, could
be adjusted on exportation within the limits set out under item (g).92
87
Appellate Body Report, US – FSC (Article 21.5 – EC), paras. 124, 126, 127, 133, 134.
88
Ibid., paras. 184–186. The panel had reached the same conclusion, but seemed to put
more emphasis on the purpose of the measure, namely, the avoidance of double taxation.
Panel Report, US – FSC (Article 21.5 – EC), paras. 8.94–8.108. See R. E. Hudec,
‘Industrial Subsidies: Tax Treatment of “Foreign Sales Corporations”’, in E.-U.
Petersmann and M. A. Pollack (eds.), Transatlantic Economic Disputes: The EU, the
US, and the WTO (Oxford University Press, 2003), 175–205, at 201.
89
Reading some flexibility into footnote 59, the Appellate Body recognized that ‘avoiding
double taxation is not an exact science and we recognize that Members must have a
degree of flexibility in tackling double taxation’. But this flexibility is not interpreted so
widely as to allow members ‘to adopt allocation rules that systematically result in a tax
exemption for income that has no link with a “foreign” State’. Appellate Body Report,
US – FSC (Article 21.5 – EC), para. 185.
90
Emphasis added.
91
Footnote 58 of the SCM Agreement (emphasis added). Concerning machinery and stamp
taxes, there was also disagreement in the Working Party on BTAs (1970) on whether these
could be adjusted (see also above n. 81). As ‘indirect taxes’, they are eligible for BTAs.
92
See footnote 60 of the SCM Agreement.

Under a system of BTAs on exportation, no VAT is thus charged on

exported products, and the exporter could deduct the VAT paid on its
inputs, implying that there is no residual VAT reflected in the export
price. Observe that today almost 150 countries have set up a VAT
system, including all OECD countries except for the United States.93
Item (h) implements the destination principle with regard to a specific
type of indirect taxes, namely ‘prior-stage cumulative indirect taxes’
(PSCI taxes). These PSCI taxes may ‘be exempted, remitted or deferred
on exported products’, even when no such adjustment is made when sold
for domestic consumption, insofar as they ‘are levied on inputs that are
consumed in the production of the exported product (making normal
allowance for waste)’.94 The SCM Agreement hereby enlarged the scope
of taxes that could be rebated under the Subsidies Code. Inputs ‘con-
sumed in the production process’ encompass, according to footnote 61,
not only ‘inputs physically incorporated’95 but also ‘energy, fuels and oil
used in the production process and catalysts which are consumed in the
course of their use to obtain the exported product’.96 The scope of item
(h) is confined to prior stage indirect taxes (i.e., taxes levied on goods or
services used directly or indirectly in making the product)97 that are also
cumulative in nature. Such cumulative taxes are ‘multi-staged taxes
levied where there is no mechanism for subsequent crediting of the tax
if the goods or services subject to tax at one stage of production are used
in a succeeding stage of production’.98 The prototypical example of such
PSCI taxes is a cascade tax system. Whereas under VAT systems only
taxes on value-added are levied on each stage of production, cascade
taxes are sales taxes levied on the actual output value at each stage of the
production process. Compared with VAT systems, such cascade systems
therefore bear a double disadvantage. They are not neutral on the length
of the production process chain as each stage is taxed on the actual value.
And they do not allow, at the final stage before exportation, the accurate
calculation of the tax amount levied during various stages of production,
93
OECD, ‘Consumption Taxes: The Way of the Future?’, OECD Observer – Policy Brief
(October 2007).
94
Emphasis added.
95
It is further clarified that inputs are physically incorporated ‘if such inputs are used in
the production process and are physically present in the product exported’ (Annex II,
section II, para. 3 of the SCM Agreement).
96
Contrast item (h) in conjunction with footnote 61 of the SCM Agreement with item (h)
of the Annex to the Subsidies Code.
97
Footnote 58 of the SCM Agreement. 98 Ibid.

generating the risk that rebates on exports would be in excess of taxes

incurred. For these reasons, OECD countries – but gradually developing
countries as well – have replaced their cascade system by a VAT system.
Some developing countries, however, still operate a cascade system
because of a lack of administrative capacity to set up a more complex
VAT system. The progressive conversion of cascade systems has reduced
the relevance of item (h) of the Illustrative List.99
By virtue of footnote 1 of the SCM Agreement, adjustments on
indirect taxes in conformity with item (g) or (h) of the Illustrative List
are neither countervailable nor actionable under the SCM Agreement.
Hence these rebates are not only exempted from the prohibition on
export subsidies but from the scope of the SCM Agreement as a whole.
In the concluding part of this volume the topical debate on whether
carbon taxes could be rebated upon exportation by virtue of item (g) or
(h) will be explored.100
4.1.1.2.2.2 Duty drawback systems In addition to indirect taxes, the

destination principle is equally applied with regard to import charges on
inputs.101 This means that these could also be rebated on exportation
insofar as they are consumed in the production of the exported prod-
uct.102 Consequently, by virtue of item (i), such a duty drawback system
is only prohibited if it results in excess drawback of the import charges
initially levied on the imported input.103,104 The justification for such
rebates is similar to that singled out for indirect taxes: import charges are
borne by the product, and most developed countries had such a draw-
back system in place. If the aim is to level the playing field between
exporters (i.e., trade neutrality), not only indirect taxes but also import
99
Also take into account the fact that certain low-income members are exempted from the
prohibition on export subsidies (see below Chapter 6, section 6.1.1.1). Hence they do not
have to rely on item (h) where they have a cascade tax system in place rebating taxes on
exports. Still, conformity with item (h) in conjunction with footnote 1 would place their
cascade taxes outside the reach of CVD action and actionable subsidy claims. Hence item
(h) might still be relevant for developing countries benefiting from S&D treatment.
100
See below Chapter 16, section 16.1.1.2.3.
101
Footnote 58 of the SCM Agreement stipulates that import charges ‘mean tariffs, duties, and
other fiscal charges not elsewhere enumerated in this note that are levied on imports’.
102
See also above Chapter 3 n. 82.
103
Again, ‘normal allowance for waste’ is treated as consumed in the production process.
The concepts of ‘waste’ and ‘normal’ are defined in respectively para. 4 and para. 5 of
Section II of Annex II of the SCM Agreement.
104
See also Article VI:4 of the GATT. The US already inscribed a duty drawback provision
in 1789. See Hufbauer and Shelton Erb, above Chapter 2 n. 12, at 63.

charges should be rebated on exports. After all, import charges are in

essence a combination of a consumption tax with a production sub-
sidy.105 Because such drawback systems often went beyond ensuring a
level playing field, excess drawbacks have been explicitly condemned as
export subsidies since the 1960 Declaration. This prohibition on excess
drawbacks is currently incorporated in item (i).
To ease the administrative burden on exporters, such an excess draw-
back is, nonetheless, allowed under item (i) of the Illustrative List if a
firm uses a quantity of home market inputs as a substitute for imported
inputs.106 To prevent disguised subsidization, such substitution draw-
back systems have to respect three conditions: the quantity of domestic
and imported inputs should be equal; domestic and imported inputs
should have the same quality and characteristics; and the import and
corresponding export operation should both occur within a reasonable
period of time not exceeding two years. The (outdated)107 idea behind
such a substitution drawback system is that exporters are not required to
ensure that a particular amount of imported inputs is effectively used in
the production of the exported product for which a drawback is
requested. Overall, however, the conditions imposed on such substitu-
tion drawback system should guarantee that no drawback will occur in
excess of levied import charges on inputs. Annex III of the SCM
Agreement spells out specific guidelines for determining whether a
substitution drawback system qualifies as an export subsidy under item
(i) of the Illustrative List.
Like BTAs as regards indirect taxes, drawback systems are confined to
inputs that are ‘consumed in the production of the exported products’ (item
(i) in conjunction with Annex II of the SCM Agreement), which covers
‘inputs physically incorporated, energy, fuels and oil used in the production
process and catalysts which are consumed in the course of their use to obtain
the exported product’ (n. 61 of the SCM Agreement). Consequently, import
charges on capital inputs could not be rebated since these are not physically
incorporated in the processed products. To remove this discrepancy, India
has put forth a proposal to expand the scope of drawbacks to import duties
105
See above Chapter 1, section 1.1. See also WT/TF/COH/15, 14 February 2003, at 7–8.
106
This was already inscribed in the Subsidies Code. The US adopted a domestic substitution
provision in the 1930s. See Hufbauer and Shelton Erb, above Chapter 2 n. 12, at 64.
107
Already in the 1980s, Hufbauer and Shelton Erb questioned the legitimacy of domestic
substitution systems because modern accounting methods made it much easier to
separate inventories on domestic and imported inputs. Hufbauer and Shelton Erb,
above Chapter 2 n. 12, at 64–5.

paid on capital inputs.108 Observing only administrative difficulties in

implementation, the IMF concurred that on economic grounds this pro-
posal is indeed solid. No economic rationale is available on why drawback
systems should be restricted to inputs consumed in the production proc-
ess.109 After all, physical capital goods that are used up (i.e., depreciate) in
the production process and other inputs such as computer software add as
much to the final cost of the product as do inputs physically incorporated.110
In addition to practical objections, there are two obvious reasons why
developed countries are not keen to accept this proposal. First, as their
import charges on capital goods are much lower, their exporters using
such capital inputs compete on a better than level playing field with export-
ers from developing countries.111 Second, allowing duty drawbacks on
capital inputs would ease the internal pressure of developing countries’
exporters on their government to lower import charges on capital inputs.
Hence developed countries tactically remarked that developing countries’
concerns could simply be addressed through lowering import charges on
capital goods. India refuted this suggestion by pointing to the importance of
import duty revenue for developing countries’ budgets.112 Given the
opposed interests, it comes as no surprise that the Draft Consolidated
Chair Text does not make capital inputs eligible for drawback along the
lines suggested by India.113
In addition to expanding the scope of inputs available for drawback,
India has proposed that a uniform drawback rate imposed on exporters
collectively be explicitly allowed for, so that an individual rate for each
and every exporter should no longer be calculated;114 especially for small
exporters, the costs of keeping track of inputs eligible for drawback
108
For the latest version of this proposal, see TN/RL/GEN/153/Rev.2, 23 February 2011.
109
WT/TF/COH/15, 14 February 2003, at 10–11. 110 Ibid., at 10.
111
Because certain low-income countries benefit from S&D treatment (see below
Chapter 6, section 6.1), rebating import charges upon capital inputs is not prohibited
for these countries but is still vulnerable to CVD action and actionable subsidy claims.
112
TN/RL/GEN/153/Rev.1, 13 March 2008, para. 6; G/SCM/W/430, 9 March 2001,
para. 3.2.
113
In response to administrative concerns articulated by members, India clarified that its
proposal is for the purpose of Annex II and not Annex III of the SCM Agreement.
However, the scope of this clarification is hard to grasp (since Annex II deals generally
with the definition of guidelines on inputs consumed and Annex III only spells out
guidelines on substitution drawback systems). It seems to suggest that members would
not be allowed to use import duties on capital goods in a substitution drawback system,
but that would not deal with other members’ concerns. TN/RL/GEN/153/Rev.2, 23
February 2011, para. 9; TN/RL/W/254, 21 April 2011, para. 32.
114
TN/RL/GEN/153/Rev.1, 13 March 2008; G/SCM/W/430, 9 March 2001.

would be too high. Once more, the IMF agreed with the rationale of this
proposal.115 The requirements for an individual drawback rate might be
onerous for developing countries, potentially leading to delays in calcu-
lating the drawback rate and thus hurting the competitive position of
their exporters. The IMF also saw merit in the proposal for political-
economy reasons because a collective rate would reduce the risk of rent-
seeking behaviour by individual exporters. The IMF therefore proposed
a hybrid system. An individual rate could be calculated for large export-
ers, since they have to keep track of the necessary records for VAT
purposes anyway, whereas for small exporters a uniform drawback rate
should indeed be an option. But, again, India’s proposal for a collective
drawback rate has not found its way into the Draft Consolidated Chair
Text, and, equally, failed to gather support during later negotiations.116
Finally, India fruitfully proposed to clarify that only the excess amount
of a tax or duty rebate shall be treated as a subsidy in a CVD inves-
tigation. Although support has been expressed for the principle of this
amendment, some members considered that it would pose ‘significant
practical problems’ for CVD investigating authorities.117
4.1.1.2.3 The (potential) direct transfer of funds: export credit

support Finally, two items of the Illustrative List deal predominantly
with official export credit support, and in essence spell out a cost-to-
government standard to determine whether such support is prohibited
(items (j) and (k)). Both items discipline different forms of export credit
support: item (j) deals with so-called ‘pure cover support’ (i.e., export credit
guarantees or insurance), and item (k) sets restrictions on ‘official financing
support’ (e.g., direct export credits). These items will be discussed in depth
in Part II and are therefore only briefly introduced below.
First, government guarantees or insurance programmes covering
potential default on credits extended to foreign buyers are considered
prohibited export subsidies under item (j) if they are offered at premium
rates inadequate to cover the long-term operating costs and losses of the
programmes. The same standard is prescribed with regard to insurance
115
WT/TF/COH/15, 14 February 2003, at 8–9.
116
It was argued that India’s proposal was too closely tailored to its own situation, left
insufficient discretion to authorities, and, ‘more fundamentally’, based the assessment
on industry averages rather than on actual inputs consumed by a particular firm. TN/
RL/W/254, 21 April 2011, para. 32.
117
Ibid., paras. 31–32.

or guarantee programmes against increases in the cost of exported

products and exchange risk programmes.
Second, item (k) prohibits governments from offering direct export
credits if these are granted at rates below those the government in
question actually has to pay for the funds so used or, alternatively,
below those the government would have to pay if it borrowed similar
funds on international capital markets. Such direct export credits would
only violate item (k) insofar as they are ‘used to secure a material
advantage in the field of export credit terms’. Additionally, government
payments of the cost incurred by exporters or financial institutions in
obtaining credits likewise qualify as export subsidies under item (k).
Again, this only kicks in when these payments are used to secure a
material advantage in the field of export credit terms.
Notwithstanding the benchmarks for export credit practices articulated
under item (j) and paragraph 1 of item (k), some of these practices are not
prohibited by virtue of paragraph 2 of item (k). This paragraph carves out a
safe haven for certain export credit conforming to the Arrangement on
Guidelines for Officially Supported Export Credit (OECD Arrangement).
To benefit from this safe haven, export credit support should be subject
to the OECD Arrangement’s interest rate obligations and conform to its
general disciplines. Because export credit guarantees or insurances are
currently not subject to the interest rate provisions, such support cannot
rely on this exemption. This means that this safe haven cannot be invoked
to justify ‘pure cover’ export credit support violating item (j) and/or Article
1 in conjunction with Article 3 of the SCM Agreement.
4.1.1.3
Export subsidies not prohibited under
the Illustrative List
The previous discussion has shown that the Illustrative List elaborates a
non-exhaustive list of export subsidies prohibited per se under the SCM
Agreement. Nonetheless, fiercely disputed before different panels is the
extent to which the Illustrative List could be used, not to demonstrate
that a subsidy is prohibited, but, a contrario, to justify that a subsidy is
not prohibited pursuant to Article 3.1(a) of the SCM Agreement.
Footnote 5 of the SCM Agreement only partly solves this question,
since it is open to different interpretations: ‘[m]easures referred to in
Annex I as not constituting export subsidies shall not be prohibited
under this or any other provision of this Agreement.’ On the one hand,
one could adopt an expansive interpretation of ‘referring’ because the
Illustrative List deals with certain specific types of subsidy (e.g., direct

export credits), and therefore sets ‘a dispositive legal standard’ to deter-

mine whether these are prohibited.118 This expansive reading, endorsed
by both Brazil (Brazil – Aircraft case) and South Korea (Korea –
Commercial Vessels case), would allow for an a contrario reading of all
aspects of the Illustrative List. For example, if the standard of item (k),
paragraph 1, would not be met (e.g., no cost to the government), a direct
export credit would not be prohibited even if it would qualify as an
export subsidy within the meaning of Article 1 in conjunction with
Article 3.1(a). However, panels have systematically rejected such an a
contrario defence for all items on the Illustrative List. In their reading,
the text of footnote 5 requires an affirmative statement that a measure is
not subject to the Article 3.1(a) prohibition, that it is not prohibited, or
that it is allowed.119 Only if such an affirmative statement is formulated
under an item of the Illustrative List could it be used to show, a contrario,
that a subsidy is not prohibited.
Because of the lack of such an affirmative statement, panels have, on
the one hand, decided that item (j), dealing with pure cover support, and
the first paragraph of item (k), dealing with official financing support,
could not be used a contrario:120 if export credit support does not
meet all the criteria set by item (j) or (k), paragraph 1, it could still
constitute a prohibited export subsidy in the case where it is demon-
strated to be a subsidy contingent on export performance (Article 1 in
conjunction with Article 3 of the SCM Agreement). For example, an
export credit guarantee under a programme that runs at break-even, as
prescribed under item (j), is prohibited pursuant to Article 3.1(a) of the
SCM Agreement if it is a subsidy in the meaning of Article 1 and
contingent on exportation. As of the time of writing the Appellate
Body has not revealed whether it would agree that an a contrario reading
of items (j) and (k), paragraph 1 should indeed be rejected.121 Through
118
This was argued by Brazil and the United States. See Panel Report, Brazil – Aircraft
(Article 21.5 – Canada), para. 6.38. The United States indicated (para. 6.39) that the
negotiating history supports a broad interpretation, since a draft version (Cartland III)
read ‘[m]easures expressly referred . . .’ and the word ‘expressly’ was dropped in
Cartland IV.
119
Panel Report, Brazil – Aircraft (Article 21.5 – Canada), paras. 6.36 and 6.38.
120
Ibid., paras. 6.36–6.37; Panel Report, Brazil – Aircraft (Article 21.5 – Canada II), paras.
5.269–5.275; Panel Report, Korea – Commercial Vessels, para. 7.198.
121
The Appellate Body stated that if Brazil had discharged its burden to show that its
financing programme did not secure a material advantage, it would ‘have been prepared
to find that the payments made under the [financing programme] are justified under

analysing other arguments, it will be explained in Part II why the panels’

rejection of such an a contrario claim is indeed correct.122
On the other hand, panels have pointed to some items containing affir-
mative statements that a measure is not an export subsidy or that it is allowed.
Hence these items could be relied on to formulate an a contrario reasoning.
First of all, the safe haven for export credits in conformity with the interest
rate provisions of the OECD Arrangement included in the second paragraph
of item (k) is such an affirmative statement.123 Thus if export credit support
is covered under this safe haven, it shall not be prohibited under the SCM
Agreement. Next, the panel in Brazil – Aircraft (Article 21.5 – Canada)
detected an affirmative statement in item (h), dealing with BTAs on prior-
stage cumulative indirect taxes, as well as in item (i), in that part dealing with
substitution drawback systems.124 Again, such BTAs or substitution draw-
back systems are therefore not prohibited. Lastly, the same panel referred to
affirmative statements in footnote 59, which offer exceptions to the prohib-
ition on BTAs on direct taxes and social welfare charges.125 The first sentence
indicates that the deferral specially related to exports of direct taxes or social
welfare charges is not an export subsidy where, for example, appropriate
interest charges are collected, and the last sentence sets out the exception of
measures imposed to avoid double taxation of foreign-source income.
Although the panel listed these provisions as examples, it is hard to detect
other types of affirmative statement in the Illustrative List.
In sum, this limited group of provisions could be used to demonstrate
that a measure is not prohibited under the SCM Agreement. However,
are measures falling under one of these affirmative statements also
shielded from a challenge on the basis of the actionable subsidy provi-
sions (multilateral) and/or a countervailing duty response (unilateral)?
Here, a distinction has to be made. On the one hand, the affirmative
statements in items (h) and (i) should be read together with note 1 of the
SCM Agreement. Indeed, BTAs and duty drawback systems in
item (k) [paragraph 1] of the Illustrative List’ (emphasis added), and thus seems to
accept an a contrario defence. Yet the Appellate Body continued by stating that ‘in
making this observation, we wish to emphasize that we are not interpreting footnote 5
of the SCM Agreement, and we do not opine on the scope of footnote 5, or on the
meaning of any other items in the Illustrative List’. Appellate Body Report, Brazil –
Aircraft (Article 21.5 – Canada), paras. 80–81; see also Appellate Body Report, US –
Upland Cotton, para. 731; Panel Report, Brazil – Aircraft (Article 21.5 – Canada II),
n. 214; Panel Report, Korea – Commercial Vessels, paras. 7.193–7.207.
122
123
Panel Report, Brazil – Aircraft (Article 21.5 – Canada), para. 6.36.
124
Ibid., nn. 35 and 36. 125 Ibid., n. 34.

conformity with the Illustrative List are simply exempted from the
subsidy definition and, therefore, can be neither challenged under the
actionable subsidy provisions nor countervailed.126 On the other hand,
with regard to the other affirmative statements, in particular the double
income tax exemption (note 59) and the safe haven for export credit
support (item (k), para. 2), such multilateral and unilateral actions seem
not ipso facto foreclosed. Whereas this conclusion needs some further
analysis (as will be undertaken in Part II),127 notice for now that this
seems to be implied in the Appellate Body’s statement in US – FSC:
Under footnote 5 of the SCM Agreement, where the Illustrative List

indicates that a measure is not a prohibited export subsidy, that measure
is not deemed, for that reason alone, not to be a ‘subsidy’. Rather, the
measure is simply not prohibited under the Agreement. Other provisions
of the SCM Agreement may, however, still apply to such a ‘subsidy’.128
These ‘other provisions’ referred to by the Appellate Body include the

provisions on actionable subsidies (Part III of the Agreement) as well as on
the imposition of countervailing measures (Part V).129 Hence the Appellate
Body left open the possibility that a double income tax exemption in
conformity with footnote 59 could still be a ‘subsidy’ under Article 1 of
the SCM Agreement and thus be actionable or countervailable if causing
adverse effects or injury to the domestic industry. Next, as discussed more in
depth in Part II of this study, export credit support justified under the safe
haven might equally be actionable and countervailable.
To conclude, the affirmative statements singled out above could be
used a contrario, namely to show that a measure is not prohibited under
the SCM Agreement. BTAs and duty drawback systems are simply
excluded from the subsidy definition.130 Other affirmative statements
shield measures from the prohibited subsidy category, but do not
exclude that they might be actionable and/or countervailable.131
126
See also Appellate Body Report, US – FSC, para. 93; Panel Report, Brazil – Aircraft
(Article 21.5 – Canada), para. 6.36, n. 37.
127
See below Chapter 10, section 10.3.
128
Appellate Body Report, US – FSC, para. 93 (emphasis in original). See also Decision by
the Arbitrators, Brazil – Aircraft (Article 22.6 – Brazil), para. 3.39.
129
Appellate Body Report, US – FSC, para. 93.
130
Obviously, this conclusion also holds on the basis of footnote 1 of the SCM Agreement
for border tax adjustments and duty drawback systems containing no affirmative
statement (e.g., item (g)).
131
The same conclusion holds for agricultural export subsidies not prohibited under the
Agreement on Agriculture (see below Chapter 6, section 6.2.3.1).

It is noteworthy that the Draft Consolidated Chair Text would amend

the current footnote 5 of the SCM Agreement according to how it is
consistently interpreted by different panels.132 The new text of this
footnote would confirm that the Illustrative List could not in principle
be used a contrario. Only measures explicitly referred to in the
Illustrative List as not constituting prohibited export subsidies could be
used a contrario. Moreover, it would be clarified that this note is without
prejudice to the operation of note 1, which confirms that BTAs and duty
drawback systems in conformity with the Illustrative List are simply
excluded from the scope of the SCM Agreement. However, a number
of delegates have opposed this amendment precisely because it would
render inoperative the cost-to-government standard of items (j) and (k),
paragraph 1.133 These members arguably still hope that the Appellate
Body would accept an a contrario reading of both items.
4.1.2 Local content subsidies

Local content subsidies (also called import-substitution subsidies) are the
second group of prohibited subsidies.134 They are defined as subsidies
contingent on the use of domestic over imported goods (Article 3.1(b)
of the SCM Agreement).135 In fact, this is the only form of domestic
subsidy that is prohibited.136 The United States was one of those during
the Uruguay Round to propose including them in the category of
prohibited subsidies, arguing that they ‘are as effective as any tariff in
protecting domestic input supplying industries and distorting the flow of
resources internationally’.137 The citation reveals that this type of
132
Draft Consolidated Chair Text, above Chapter 2, n. 113.
133
TN/RL/W/254, 21 April 2011, para. 19.
134
For S&D treatment see below Chapter 6, section 6.1.1.2. For the normative discussion
on the disciplines on local content subsidies, see below Chapter 16, section
16.1.1.2.2.2.1 (developed countries) and Chapter 17, section 17.2.1 (developing
countries).
135
For a discussion on the relationship between the TRIMs Agreement and the SCM
Agreement, see P. Sauvé, Trade Rules behind Borders: Essays on Services, Investment
and the New Trade Agenda (London: Cameron May, 2003), at 313–18.
136
Local content subsidies might in addition be contingent on export performance and
thus constitute export subsidies (Article 3.1(a) of the SCM Agreement). As mentioned
above, ‘contingency’ does not require that import substitution should be a sufficient
condition to receive the subsidy. So, export performance can be an additional require-
ment. See, e.g., the Canada – Autos case.
137
MTN.GNG/NG10/W/29, 22 November 1989; Stewart, above Chapter 2 n. 49, at 889;
Anderson and Husisian, above Chapter 2 n. 80, at 309.

subsidy is conceptually different from other kinds of subsidies. It focuses

on the trade distortion in the input industry market and not in the
market of the industry receiving the beneficial financial contribution
by the government.138
Similarly to export subsidies under Article 3.1(a), two elements have
be demonstrated to fall within this prohibition. First, a subsidy within
the meaning of Article 1 of the SCM Agreement has to be identified.
Second, this subsidy must be ‘contingent . . . upon the use of domestic
over imported goods’. The term ‘use’, as well as the negotiating history
set out above, seems to suggest that only subsidies that are effectively
processed in the production process by the recipient (and not merely, for
instance, resold) are targeted. The Appellate Body defined ‘goods’ as
‘tangible items’, which are ‘often contrasted against “services”, which are
intangible’.139 Thus Article 3.1(b) prohibits subsidies contingent on the
use of domestic over imported tangible items in the production process.
Contrary to export subsidies, the type of conditionality is not explicitly
prescribed, but the Appellate Body decided that it likewise covers not only
de jure but also de facto contingency.140 Moreover, the legal standard of
contingency is considered similar to that under the export contingency
standard.141 Hence the well-developed case law on the required strength of
export contingency and its determination is equally relevant to assess the
contingency element under paragraph (b) (e.g., the ‘geared to induce’ test
developed by the Appellate Body to determine de facto contingency).
In Canada – Autos, the challenged subsidy (i.e., the import duty exemp-
tion) was not only contingent on exportation but was equally linked to
138
For instance, such a subsidy would no longer be prohibited if the local content
condition were to be deleted. This subsidy could still be challenged as an actionable
or export subsidy (see also above n. 136).
139
Appellate Body Report, US – Large Civil Aircraft, n. 1295
140
Appellate Body Report, Canada – Autos, paras. 135–43. The Appellate Body’s principal
reason seemed to be the circumvention argument.
141
The Appellate Body in Canada – Autos (para. 123) explained that ‘this legal standard
applies not only to ‘contingency’ under Article 3.1(a), but also to ‘contingency’ under
Article 3.1(b) of the SCM Agreement’. Although it dealt with a de jure contingency
claim, there is no reason why it would not adopt similar reasoning with regard to de
facto contingency claims, certainly given the fact that the Appellate Body has equally
confirmed that the legal standard is similar for de jure and de facto contingency claims
(Appellate Body Report, EC – Large Civil Aircraft, para. 1037; Appellate Body Report,
Canada – Aircraft, para. 167). By logical implication, the legal standard for de facto
contingency should therefore be similar under items (a) and (b). If the legal standard
for contingency is similar under both items, the test to detect whether this standard is
met in fact (e.g., the ‘geared to induce’ test) should also be similar.

certain Canadian value added (CVA) requirements. The definition of CVA

included parts and materials of Canadian origin (i.e., local content) among
other elements such as labour costs and manufacturing overheads. The
panel considered the subsidy as not contingent on local content because a
manufacturer was able to satisfy the CVA requirements without using any
domestic good whatsoever. However, the Appellate Body rejected the
panel’s conclusion because it had overlooked the level of CVA require-
ments. As an example, the Appellate Body indicated that ‘if the level of the
CVA requirements is very high, we can see that the use of domestic goods
may well be a necessity and thus be, in practice, required as a condition for
eligibility for the import duty exemption’.142 The panel should have exam-
ined the actual level of CVA required in order to decide whether the subsidy
was de jure dependent on the use of domestic goods.143 The Appellate Body
agreed that the value-added requirement would be outside the scope of the
de jure local content subsidy prohibition if it could be satisfied without
using domestic goods.
Local content subsidies within the meaning of Article 3.1(b) clearly
violate the GATT’s national treatment provision (Article III:4) because
the regulations at hand discriminate between the domestic and foreign
input supplying industry. The carve-out stipulated in Article III:8(b) of
the GATT is not applicable because the discrimination exists between
the domestic and foreign input industries and not between the subsi-
dized domestic industry and foreign industry.144 Moreover, local content
subsidies are covered by the Illustrative List of the Agreement on Trade-
Related Investment Measures (TRIMs Agreement) that are inconsistent
with Article III:4 of the GATT (as local content requirements).145,146
Demonstrating a violation of the TRIMs Agreement is easier than
demonstrating the presence of a local content subsidy under the SCM
Agreement, because the latter needs to pass the subsidy definition.147
142
Canada – Aircraft, para. 130 (first emphasis added).
143
The Appellate Body did not complete the analysis and also did not decide on de facto
contingency. Ibid., paras. 130–131.
144
See below Chapter 4, section 4.4.1.
145
See Article 2 in conjunction with paragraph 1(a) of the Illustrative List of the TRIMs
Agreement. The panel in Indonesia – Autos concluded that measures challenged under
both the TRIMs Agreement and the SCM Agreement must be reviewed under both
(paras. 14.49–14.55).
146
This has to be nuanced insofar as S&D treatment under the TRIMs Agreement is still in
place (see below Chapter 6, section 6.1.1.2).
147
The local content element is also formulated more broadly under the TRIMs Agreement. It
refers not only to ‘the use’ of domestic products but also to their ‘purchase’.

The claim under the TRIMs does, obviously, not need to characterize the
challenged measure as a financial contribution. It only needs to dem-
onstrate an ‘advantage’ contingent on local content, which is much
easier than demonstrating that it is a ‘benefit’ under the SCM
Agreement. In Canada – Renewable Energy / Feed-in Tariff Program,
‘mere participation’ in [the] FIT Program was viewed as obtaining an
‘advantage’, because the progamme guaranteed a fixed price to be paid
over a period of twenty years.148 At the same time, as explained above,
the complainants failed to demonstrate that the same FIT Program
conferred a ‘benefit’ under the SCM Agreement. The Appellate Body
confirmed that, as a matter of law, a ‘benefit’ within the meaning of the
SCM Agreement has a more specific meaning than an ‘advantage’ within
the meaning of the TRIMs Agreement.149 Thus a violation of the TRIMs
Agreement (and Article III:4 of the GATT) is much easier than demon-
strating the presence of a local content subsidy under the SCM
Agreement.
Nonetheless, an additional claim under the SCM Agreement might be
valuable because, in contrast to the GATT and the TRIMs Agreement,
this agreement does not provide any ground for justification for local
content subsidies, and has specific and stricter dispute settlement pro-
visions in place.150 Panels may, therefore, not exercise judicial economy
regarding a claim under Article 3 of the SCM Agreement when they find
a violation of Article III:4 of the GATT.151
4.2 Actionable subsidies

Actionable subsidies are defined by default in Part III of the SCM
Agreement: any specific subsidy within the meaning of Articles 1 and 2
148
Panel Report, Canada – Renewable Energy / Feed-in Tariff Program, para. 7.165
(emphasis in original); Appellate Body Report, para. 5.206.
149
Hence a benefit within the meaning of the SCM Agreement de facto always amounts to
an advantage within the meaning of the TRIMs Agreement. Appellate Body Report,
paras. 5.207–5.209.
150
See below Chapter 5. A violation of Article III:4 of the GATT or Article 2 of the TRIMs
can still be justified on the basis of the general exceptions (Article XX of the GATT and
Article 3 of the TRIMs) or balance-of-payments exception (Articles XII and XVIII:B of
the GATT and Article 4 of the TRIMs). On the applicability of Article XX of the GATT
to violations of the SCM Agreement, see below Chapter 4, section 4.4.2.
151
Appellate Body Report, EC – Export Subsidies on Sugar, para. 335. This is not, however,
always obeyed. Panel Report, China – Auto Parts, para. 7.635.

of the SCM Agreement is potentially actionable.152 This means that such

amber light subsidies can be subject to multilateral action if they cause
‘adverse effects’ to the interests of other members.153 As one panel
explained, ‘the focus the trade effects of subsidization in Article XVI:1 of
the GATT 1994 on any other member is carried into Part III of the SCM
Agreement’.154 The main onus on the complaining member is indeed to
demonstrate such ‘adverse effects’ (Article 5 of the SCM Agreement). This
covers three situations: (a) injury to the domestic industry; (b) nullification
or impairment of benefits accruing to another member under GATT 1994;
and (c) serious prejudice to the interests of another member.155
4.2.1 Injury to the domestic industry

The term ‘injury to the domestic industry’ is used in the same way as in
the context of CVD procedures.156 It will be explained in the discussion
on CVD action that ‘injury’ covers material injury as well as ‘a threat’ of
material injury to a domestic industry.157 Instead of undertaking a CVD
procedure (unilateral remedy), members can thus opt for the dispute-
settlement system (multilateral remedy) and demonstrate injury to their
domestic industry.
The panel in EC – Large Civil Aircraft considered it appropriate – but
not mandatory – to adopt the two-step approach used in CVD inves-
tigations: (i) asserting the presence of material injury; and, if this is
found,158 (ii) establishing that subsidized imports are causing that mate-
rial injury.159
First, regarding injury, it has to be established that the domestic
industry was materially injured during the reference period.160 The
152
Even prohibited subsidies could in theory still be challenged as actionable subsidies. See
Panel Report, Korea – Commercial Vessels, para. 7.334. For S&D treatment on action-
able subsidies see below Chapter 6, section 6.1.2. For the normative discussion on the
disciplines on domestic subsidies see below Chapter 16, section 16.1 (developed
countries) and Chapter 17, section 17.2.2 (developing countries).
153
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 238.
154
Panel Report, US – Upland Cotton, para. 7.1405, n. 1503. See also Panel Report,
Indonesia – Autos, para. 14.200.
155
Article 5 of the SCM Agreement.
156
Article 5(a), footnote 11 of the SCM Agreement. See Part V and Article 15 of the SCM
Agreement.
157
Article 15, footnote 45 of the SCM Agreement.
158
Panel Report, EC – Large Civil Aircraft, para. 7.2082. 159 Ibid., para. 7.2058.
160
On the determination of the reference period, see ibid., paras. 7.2045–7.2053.

question whether a degree of injury is ‘material’ is ‘fact-specific in

each case, depending on the nature of the product and the industry
in question’.161 The panel’s analysis implies that material injury is only
present if the domestic industry’s overall performance (or at least
some relevant factors)162 deteriorated over the reference period.163
This claim is unsuccessful in situations where the domestic industry’s
performance does not reflect material injury, but would have been
materially better without the subsidized imports.164 This differs from a
‘serious prejudice’ claim, which is available regardless of the state of the
(domestic) industry during the reference period. Turning to the facts of
the case, the panel refuted the claim that Boeing was materially injured,
because by the end of the reference period (2006) it had recovered from
the fall in demand for aircraft following the 9/11 terrorist attacks on the
United States.165
Second, regarding the causation element, the panel imported the flexible
standard applicable in CVD investigations: the relevant causal factor is
subsidized imports and not the subsidies.166 It only has to be demonstrated
that the injury is caused by the subsidized imports, in view of their volume
and price effects and their consequent impact on the domestic industry. As
further explained below, demonstrating causation in a serious prejudice
analysis is more demanding because it has to be shown that the effects are
caused by the challenged subsidies. The implications of this difference in
causation standard are further explored below.167
In sum, the ‘injury to the domestic industry’ form of adverse effects is
only available in case of a materially injured domestic industry, but
demonstrating causation is less demanding than under the serious prej-
udice analysis. An ‘injury to the domestic industry’ claim requires
demonstration of injury caused by subsidized imports, whereas a ‘serious
161
Ibid., para. 7.2083. See also below Chapter 5, section 5.2.2.2.
162
Ibid., para. 7.2084. The relevant factors are mentioned in Article 15.4 of the SCM
Agreement. See below Chapter 5, section 5.2.2.2.3.
163
Ibid., paras. 7.2084, 7.2110–7.2112, 7.2157.
164
Ibid., para. 7.2157. The panel seemed to consider this to be a necessary implication of
the bifurcated approach. In a two-step analysis, material ‘injury’ could – and should – be
established without consideration of the subsidized imports. The determination of
material ‘injury’ under step 1 is made in the light of the industry’s performance over
time (i.e., it is ‘injured’ because performance criteria deteriorated over the reference
period), and not in the light of the subsidized imports (i.e., it is ‘injured’ because it
would have performed better but for the subsidized imports).
165
Ibid., paras. 7.2111–7.2112. 166 Ibid., paras. 7.2059–7.2071.
167

prejudice’ claim requires demonstration of trade effects caused by sub-

sidies. The Appellate Body has not yet confirmed this interpretation, but
has only explained that in assessing injury to the domestic industry, the
focus is on the impact of subsidized imports ‘on the domestic industry of
another Member’, whereas the focus of the serious prejudice analysis is
on the effects in a market.168
4.2.2 Nullification or impairment of benefits

The concept of nullification or impairment of benefits is formulated in
the same sense as in the GATT.169 This concept is applied in non-
violation complaints (NVCs) (Article XXIII:1(b) of the GATT) where
three elements are established: (i) the use of a subsidy,170 (ii) the exis-
tence of a benefit accruing under the GATT, in particular the benefit of
tariff concessions,171 and (iii) the nullification or impairment of a benefit
(e.g., tariff concessions) as a result of the use of a subsidy (i.e., causation
element).172 Because the non-violation remedy should remain of an
‘exceptional nature’,173 the standard for determining the element of
causation (iii) is set high by the adjudicating bodies: ‘non-violation
nullification or impairment would arise when the effect of a tariff con-
cession is systematically offset or counteracted by a subsidy program’.174
There is a double motivation behind NVCs. First, members have rea-
sonable expectations that they can benefit from tariff concessions.175
Second, these members have ‘paid’ for the binding by making tariff
concessions themselves. If trade partners offered tariff bindings for a
product and were subsequently to be allowed to subsidize so as to reduce
market access for imports of the same product, the dynamism of recip-
rocal tariff concessions would be undermined. As explained, the NVC
under the original GATT 1947 was deemed useful precisely in prevent-
ing domestic subsidies from eroding tariff concessions.176
168
Appellate Body Report, EC – Large Civil Aircraft, n. 2461. See also Panel Report,
Korea – Commercial Vessels, paras. 7.578–7.579.
169
Footnote 12 of the SCM Agreement.
170
Panel Report, US – Offset Act (Byrd Amendment), paras. 7.121–7.123.
171
Ibid., para. 7.124.
172
Ibid., paras. 7.125–7.131. All non-violation complaints in the GATT period concerned
subsidies that were claimed to nullify a prior negotiated concession. P. C. Mavroidis,
Trade in Goods (Oxford University Press, 2007), at 413.
173
Panel Report, US – Offset Act (Byrd Amendment), para. 7.127.
174
Ibid., para. 7.127 (emphasis added); GATT panel report, EEC – Oilseeds I, para. 148.
175
GATT panel report, EEC – Oilseeds I, para. 148. 176 See above Chapter 2, section 2.1.

Whereas the first type of adverse effect (i.e., injury to the domestic
industry) occurs in the domestic market of other members, this second
type tackles adverse effects on the export industry of other members in
the market of the subsidizing country. Yet the latter is confined to the
situation where benefits accruing under the GATT are nullified. The
‘serious prejudice’ type of adverse effect expands the scope to effects
occurring in third countries as well as in the subsidizing country, even
when no other benefits are nullified.
4.2.3 Serious prejudice

Originally, the SCM Agreement provided for a rebuttable presumption
that ‘serious prejudice’ exists in four cases: (a) the total ad valorem
subsidization of a product exceeds 5 per cent; (b) the subsidy covers
operating losses sustained by an industry; (c) the subsidy covers operat-
ing losses sustained by an enterprise, other than one-time measures; or
(d) direct debt forgiveness.177 This provision facilitated the difficult
demonstration of serious prejudice. The complaining member only
had to demonstrate the existence of such a specific subsidy and could
place the burden on the defending party to refute the claim that this
subsidy caused serious prejudice (in the meaning of Article 6.3). Yet this
so-called ‘dark amber’ category set out in Article 6.1 expired at the end of
1999 because there was no consensus among members in favour of
continuing its existence. From the perspective of WTO law, it is rather
exceptional that a stricter discipline is subject to extinction. However, the
drafters during the Uruguay Round considered this stricter discipline
(Article 6.1), which reflected in a diluted form the quantitative approach
suggested by the United States, as a quid quo pro for the inclusion of
non-actionable subsidies (Article 8), insisted on by other negotiating
countries.178 Their future was thus bound together, and both expired
after a period of five years.179
At present, it is thus always up to the complaining party to demon-
strate that a specific subsidy causes serious prejudice or a threat of
serious prejudice.180 Such serious prejudice is present pursuant to
177
Articles 6.1, 6.2 of the SCM Agreement. 178 Article 31 of the SCM Agreement.
179
See also Panel Report, Korea – Commercial Vessels, para. 7.583. See A. Hoda and R. Ahuja,
‘Agreement on Subsidies and Countervailing Measures: Need for Clarification and
Improvement’, 39:6 Journal of World Trade (2005), 1009–969, at 1061.
180
The subsidizing country had to prove that the subsidization did not result in any of the
effects described in Article 6.3 of the SCM Agreement (Article 6.2 of the SCM Agreement).

Article 6.3 of the SCM Agreement if one or more of the following market
effects has been demonstrated:181
(a) the effect of the subsidy is to displace or impede the imports of a like
product of another Member into the market of the subsidizing Member;
(b) the effect of the subsidy is to displace or impede the exports of a like
product of another Member from a third-country market;182
(c) the effect of the subsidy is a significant price undercutting by the
subsidized product as compared with the price of a like product of
another Member in the same market or significant price suppression,
price depression, or lost sales in the same market;183
(d) the effect of the subsidy is an increase in the world market share of
the subsidizing Member in a particular subsidized primary product
or commodity as compared with the average share it had during the
previous period of three years, and this increase follows a consistent
trend over a period when subsidies have been granted.184
Serious prejudice is present if one of these four types is demonstrated,

without the need of establishing any additional element.185 Their focus is on
the trade effects of subsidization.186 These could be effects on volume (items
(a), (b), lost sales under (c), and (d)) and/or on prices (item (c)) that are
adverse from the perspective of other members’ trade interests. To formu-
late a successful claim, the presence of one of the listed volume or price
market phenomena should be demonstrated, and these should be shown to
be ‘the effect of the subsidy’ (i.e., the causation element). While the magni-
tude of subsidies might be relevant in this respect, there is no strict require-
ment to quantify the precise subsidy amount.187 The requirement
effectively to demonstrate trade effects forms the essential difference from
the prohibited subsidy category and explains the higher hurdle, in legal and
financial terms, to formulate an actionable subsidy claim.188
Instead of actual ‘serious prejudice’, a complaining member could show a
‘threat of serious prejudice’.189 Such a threat of serious prejudice suffices to
181
The list is considered non-exhaustive (Panel Report, Korea – Commercial Vessels, paras.
7.601–7.603; Panel Report, US – Upland Cotton, para. 7.1388). In the light of the variety
of listed adverse effects, the list seems de facto closed once a threat of serious of
prejudice is taken on board.
182
See also Article 6.4 of the SCM Agreement.
183
See also Article 6.5 of the SCM Agreement.
184
Emphasis added, original footnote omitted.
185
Panel Report, Korea – Commercial Vessels, paras. 7.552–7.603.
186
Panel Report, US – Upland Cotton, n. 1503.
187
Appellate Body Report, US – Large Civil Aircraft, para. 697. 188 Ibid., n. 2428.
189
Footnote 13 of the SCM Agreement.

trigger the remedies available under Article 7 of the SCM Agreement.190 Yet
such a claim relates to a different situation from a claim of actual serious
prejudice: ‘a claim of present serious prejudice relates to the existence of
prejudice in the past, and present, and that may continue in the future. By
contrast, a claim of threat of serious prejudice relates to prejudice that does
not yet exist, but is imminent such that it will materialize in the near
future’.191 Therefore the Appellate Body concluded that ‘a threat of serious
prejudice claim does not necessarily capture and provide a remedy with
respect to the same scenario as a claim of present serious prejudice’.192 The
determination of a threat of serious prejudice should ‘be based on facts and
not merely on allegation, conjecture, or remote possibility’ and ‘the change
in circumstances’ that would trigger actual prejudice must be ‘clearly
foreseen and imminent’.193
Before discussing the specific types of serious prejudice (section
4.2.3.4), we explain the need under all types to demonstrate (i) present
adverse effect (section 4.2.3.1); (ii) often on like products (section
4.2.3.2); (iii) competing in the same relevant product market (section
4.2.3.3).
4.2.3.1 The need to demonstrate present adverse effects

The complainant has to demonstrate that the challenged subsidies are
presently causing adverse effects. Given the lack of immediate data on the
effect of the subsidy at the moment of the complaint, historical data could,
however, be relied on to draw a conclusion about present adverse effects.194
The complainant has to show that adverse effects occurred during this
reference period. The most recent available data should be part of this
reference period, even if it dates from after the panel’s establishment.195
The Appellate Body explicitly rejected that subsidies, and in particular
the conferred benefit, should be present during this reference period.
Article 6.3 only requires present adverse effects during the reference
period, and not present subsidization. Given that there could exist a
time lag between the use of the subsidy and the adverse effects, ‘under
190
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 7.1494–7.1497.
191
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 244 (final
emphasis added).
192
Ibid., para. 244. 193 Appellate Body Report, EC – Large Civil Aircraft, para. 1171.
194
Panel Report, EC – Large Civil Aircraft, para. 7.194; Panel Report, US – Upland Cotton
(Article 21.5 – Brazil), paras. 10.18, 10.265.
195
Appellate Body Report, EC – Large Civil Aircraft, para. 1167; Panel Report, EC – Large
Civil Aircraft, para. 7.1714.

certain circumstances, a past subsidy that no longer exists may be found

to cause or have caused adverse effects that continue to be present during
the reference period’.196 A subsidy has a ‘finite life’ that comes to an end
either through ‘the removal of the financial contribution and/or the
expiration of the benefit’.197 Although adverse effects also accrue and
diminish over time, they might thus end after the point when the subsidy
has expired.198 Hence, in the Appellate Body’s view, a subsidy could still
cause adverse effects even though it no longer confers a benefit (no
‘continuing benefit’) in the meaning of Article 1 of the SCM Agreement.
In conducting the adverse effect analysis, panels must take into account
how the subsidy materialized over time; this is affected by two compo-
nents.199 First, the subsidy is affected by the depreciation that was projected
for it ex ante, which depends on its nature, amount, and projected use.
Elements that could be considered as part of this ex ante analysis are:
whether the subsidy is allocated to purchase inputs or fixed assets; the
useful life of these inputs or assets; whether the subsidy is large or small;
and the period of time over which the subsidy is expected to be used for
future production.200
Second, the projected value resulting from this ex ante analysis might
be affected by ‘intervening events’ occurring after the grant of the sub-
sidy that ‘diminish it or bring it to an end’.201 Such intervening events
may attenuate the causal link between the subsidy and its adverse effects.
Two of the three ‘intervening events’ advanced by the EU in EC – Large
Civil Aircraft have been discussed above. The Appellate Body did not
solve the question of whether subsidies could become extinct as a result
of partial privatization or private-to-private sales.202 Next, regarding the
EU’s argument that the restructuring and legal reorganization of a
previously subsidized company implied that subsidies did not pass-
through to the new company, the need for a ‘pass-through’ analysis
was rejected in this case, but the Appellate Body seemed to accept that
restructuring of companies could be an intervening event.203 Finally, the
196
198
200
Allocation of a subsidy over the anticipated marketing life ‘may be one way to assess the
duration of a subsidy over time’. Ibid. paras. 707, 1241.
201
Ibid., paras. 709, 716.
202
Ibid., para. 735. See above Chapter 3, section 3.2.2.2; below Chapter 15, section
15.2.2.1.
203
Appellate Body Report, EC – Large Civil Aircraft, paras. 774–777; above Chapter 3,
section 3.2.2.1.2.

Appellate Body accepted that the subsidies could be removed in full or

part by the removal of cash or cash equivalents. Two conditions have to
be met to accept such extraction of subsidies. First, a relationship must
be present between the subsidy and cash extracted. Hence the Appellate
Body rejected the EU’s argument that, ‘because a subsidy would be
reflected in a company’s balance sheet, and cash is fungible, once cash
is removed there is an adequate link established between the subsidies
provided and the cash extracted’.204 At a minimum it has to be shown
how the specific subsidies were reflected in the balance sheets, and how
the cash extracted represented the remaining value of these subsidies.
Second, the cash must be truly extracted and thus have permanently left
the company. Because the EU failed to meet the first condition, cash
removed by the parent company from wholly owned subsidiaries was not
an intervening event affecting the causal relationship.
4.2.3.2
The origin and likeness of products under a
serious prejudice claim
To which products should serious prejudice be caused? First, regarding
the origin of products, the panel in Indonesia – Autos stressed that a
member cannot bring a claim that another member has suffered serious
prejudice. Hence products not originating in a complaining member
country cannot be the object of this claim. For instance, products made
by a US company at its foreign plant cannot give rise to a US claim of
serious prejudice. Instead, serious prejudice must occur in relation to
products made within the territory of the complaining party.205
Similarly, the panel in US – Upland Cotton found that allegations of
serious prejudice affecting other members may be taken into account ‘to
the extent these constitute evidentiary support of the effect of the subsidy
borne by [the complaining Member]’, but its decision could not be based
‘on any alleged serious prejudice caused to’ the other members.206 A
complaining member can only challenge serious prejudice caused to
products originating in its territory.207 If this claim is successful, the
204
Likewise, cash would not be considered extracted if it were in the form of dividends
representing the profits of a company. Appellate Body Report, EC – Large Civil Aircraft,
paras. 744–746.
205
Panel Report, Indonesia – Autos, paras. 14.201–14.202.
206
Panel Report, US – Upland Cotton, paras. 7.1414–7.1415.
207
Obviously, as Article 6.3 plainly shows, such serious prejudice to products originating in the
complaining member’s country could very well occur in the third-country markets.

implementation standard could imply that adverse effects to other

members will be removed, too.208
Second, regarding the scope of products in a serious prejudice claim, it is
explicitly stipulated for some types of serious prejudice that their effect
should be caused to ‘a like product’ (Articles 6.3(a), 6.3(b), and 6.3(c)).209
The term ‘like product’ should throughout the SCM Agreement be under-
stood narrowly as ‘identical, i.e., alike in all respects’ or, at least, as having
‘characteristics closely resembling’ the product under consideration.210 The
panel in Indonesia – Autos held that the term ‘characteristics closely resem-
bling’ is ‘on its face . . . quite narrow’ and ‘includes but is not limited to
physical characteristics’.211 Nonetheless, physical characteristics were con-
sidered an important element especially because other criteria (e.g., end-use,
consumer perception, substitutability, price, tariff classification) were
viewed as primarily based on physical characteristics. Therefore the panel
in Korea – Commercial Vessels summarized that this previous panel had
found that the term ‘characteristics closely resembling’ should ‘be based
principally on physical characteristics of the product’,212 but this seems to
overstate the panel’s emphasis on physical characteristics.213
Not all types of serious prejudice should explicitly pass this likeness
test, however. The panel in Korea – Commercial Vessels concluded that
this is only required regarding those types of serious prejudice in which
an explicit reference is made to ‘a like product’ (i.e., Article 6.3(a), Article
6.3(b), and price undercutting under Article 6.3(c) of the SCM
Agreement). No like product test should therefore be conducted when,
for instance, price suppression or depression under Article 6.3(c) is
invoked.214 Here, the panel viewed ‘the product issue ultimately as
pertaining to the demonstration of causation, on the basis of such facts
as may be relevant to the particular case’.215
208
See Decision by the Arbitrator, US – Upland Cotton (Article 22.6 – United States)
(actionable subsidies), paras. 4.80–4.92; below Chapter 5, section 5.1.3.2.
209
Emphasis added. 210 Footnote 46 of the SCM Agreement.
211
See Panel Report, Indonesia – Autos, paras. 14.172–14.173. The Appellate Body in EC –
Large Civil Aircraft (n. 2464) referred to the legal criteria set out by this panel.
212
Panel Report, Korea – Commercial Vessels, n. 289.
213
The panel considered it relevant that the prefix ‘physical’ characteristics in one of the
draft texts was deleted in the final Kennedy Round Anti-Dumping Agreement. See
Panel Report, Indonesia – Autos, paras. 14.173 (n. 730), 14.191.
214
Panel Report, Korea – Commercial Vessels, para. 7.553. The Appellate Body has not yet
revealed whether it would share the same view. Appellate Body Report, US – Upland
Cotton, n. 453.
215
Panel Report, Korea – Commercial Vessels, paras. 7.559, 7.600.

4.2.3.3 The delineation of the relevant product market under

a serious prejudice claim
The Appellate Body in EC – Large Civil Aircraft decided that the word
‘market’ in the various sub-paragraphs of Article 6.3 calls for the delinea-
tion of the relevant market for a serious prejudice analysis.216 The Appellate
Body’s decision was confined to paragraphs (a) and (b) of Article 6.3, but
its reasoning seems equally applicable to paragraph (c), as this also refers
to the market and is based on a similar economic mechanism.
The Appellate Body reasoned that specific mechanisms by which the
effects in Article 6.3 operate are ‘essentially economic mechanism[s], the
existence of which is to be assessed by reference to events that occur in
the relevant product market.’217 The effects (such as displacement) can
only occur if products compete in the same market, implying that ‘An
examination of the competitive relationship between products is there-
fore required so as to determine whether such products form part of the
same market.’218 Only if two products are substitutable in the market
could they exercise competitive constraint on each other and cause
serious prejudice by subsidization of each other. The Appellate Body there-
fore defined a ‘market’ as ‘a set of products in a particular geographical area
that are in actual or potential competition with each other’.219
Hence the relevant market analysis aims at detecting which products
are substitutable. To define this relevant market, demand-side and
supply-side considerations should be taken into account:
Demand-side substitutability – that is, when two products are considered

substitutable by consumers – is an indispensable, but not the only
relevant, criterion to consider when assessing whether two products are
in a single market. Rather, a consideration of substitutability on the
supply-side may also be required. For example, evidence on whether a
supplier can switch its production at limited or prohibitive cost from one
product to another in a short period of time may also inform the question
of whether two products are in a single market.220
This relevant market analysis is related to, but still distinct from, the
above-discussed likeness analysis.221 According to the Appellate Body,
216
Appellate Body, EC – Large Civil Aircraft, para. 1119. 217 Ibid., para. 1119.
218
220
Ibid., para. 1121. See also Appellate Body report, Canada – Renewable Energy / Feed-in
Tariff Program, para. 5.171.
221
The Appellate Body explained that, ‘while a complaining Member may identify a
subsidized product and the like product by reference to footnote 46, the products

the word ‘market’ in Article 6.3 must be read together with the concept of
‘like product’, implying that ‘identity or close resemblance of characteristics
are one factor to consider in assessing whether products are in the same
market’.222 But the Appellate Body further emphasized that ‘Although
physical characteristics, end-uses, and consumer preferences may assist
in deciding whether two products are in the same market, they should
not be treated as the exclusive factors to consider in deciding whether
those products are sufficiently substitutable so as to create competitive
constraints on each other.’223 For instance, it may also be relevant to
consider whether customers demand a range of products or whether they
are interested only in a particular product type.224 Further, in addition to
demand-side substitutability, supply-side substitutability should be present
in order for two products to be part of the same relevant market.
In the following sections, the specific types of adverse effects covered
under the serious prejudice category are scrutinized. Two steps are
distinguished. First, the different market phenomena (price and volume
effects) are identified. Second, the causation element is explored, as these
market phenomena have to be caused by the subsidies. It will be
explained that the Appellate Body has a preference for analysing both
steps simultaneously (i.e., the so-called unitary analysis), but equally
allows a separate analysis thereof (i.e., a bifurcated analysis). Finally,
the relevant case law is discussed briefly in chronological order.
4.2.3.4 Types of market phenomena

4.2.3.4.1 Volume effects
4.2.3.4.1.1 Displacement or impedance of exports to the market of the
subsidizing member or a third country Paragraphs (a) and (b) of Article
6.3 deal with volume effects in the markets of the subsidizing country
and third markets respectively. First, Article 6.3(a) implies that, even in
the absence of any prior tariff commitment (Article 5(b)), members
thereby identified must be analyzed under the discipline of the product market so as to
be able to determine whether displacement is occurring. Ordinarily, the subsidized
product and the like product will form part of a larger product market. But it may be the
case that a complainant chooses to define the subsidized and like products so broadly
that it is necessary to analyze these products in different product markets. This will be
necessary so as to analyze further the real competitive interactions that are taking place,
and thereby determine whether displacement is occurring.’ Appellate Body report, EC –
Large Civil Aircraft, para. 1119.
222
Ibid., para. 1118. 223 Ibid., para. 1120 (emphasis in original).
224
‘In the former case, when customers procure a range of products to satisfy their needs,
this may give an indication that all such products could be competing in the same
market.’ Ibid., para. 1120.

could cause adverse effects through subsidization by displacing or

impeding ‘the imports of another Member into the market of the subsidiz-
ing Member’.225 This further opens the door to challenge adverse effects
in the subsidizing country’s domestic market. Second, Article 6.3(b)
addresses a similar situation in which the effect of subsidization results in
displacement or impeding of exports from a third-country market, which
is a market other than the market of the complaining (and subsidizing)
country.226 Depending on the conditions of competition, the geographical
dimension of the market could cover the entire territory or a sub-territory,
such as a region. Equally, the geographical dimension could transcend
national boundaries and could even be the world market, although the
focus of the analysis should be confined to the territory of either the
subsidizing country (Article 6.3(a)) or a third country (Article 6.3(b)).227
The Appellate Body explained that in case of displacement, there is ‘a
substitution effect between the subsidized product and the like product of
the complaining Member’.228 Put otherwise, it refers to a situation ‘where
imports or exports of a like product are replaced by the sales of the
subsidized product’, in the market of the subsidizing country or in a
third-country market.229 The concept of impediment connotes a broader
array of situations, as it refers to situations where the trade of the like
product of the complainant ‘would have expanded had they not been
“obstructed” or “hindered” by the subsidized product’, or ‘did not materi-
alize at all because production was held back by the subsidized product’.230
Recognizing that it may be difficult to demarcate both concepts, the
Appellate Body suggested that ‘evidence that actual sales have declined
would be relevant for a determination of displacement, whereas evidence
that sales would have increased more than they did, or would have declined
less than they did, would be relevant to a claim of impedance’.231
Displacement is only present if changes in relative market share can be
demonstrated over a sufficiently representative period so as to show ‘clear
trends’ in the development of the market.232 Two elements have to be
shown: (i) ‘at least a portion of the market share of the exports of the like
product of the complaining Member must have been taken over or sub-
stituted by the subsidized product’; and (ii), ‘it must be possible to discern
225
Emphasis added.
226
228
Ibid., para. 1160. 229 Ibid., para. 1119. 230 Ibid., para. 1161.
231
Ibid., para. 1162.
232
The Appellate Body found contextual support in Article 6.4, but erroneously consid-
ered that this applied to both Article 6.3(a) and (b) (see below n. 240).

trends in volume and market share’.233 Regarding impedance, the Appellate

Body held that a clear trend should in principle also be discerned, although
it recognized that ‘unlike with displacement, however, impedance may not
be a visible phenomenon, evidence of trends may not be dispositive, or may
hold less probative value, for a finding of impedance’.234
Given that these volume effects cannot occur in the absence of com-
petition in the same product market, the Appellate Body reasoned that
their presence has to be demonstrated by reference to events that
occurred in the relevant product market. Therefore the relevant product
market has to be analysed so as to determine whether products are part
of a single-product or several-products market. This mandates not only
consideration of demand-side substitutability (are products deemed
substitutable by consumers?)235 but might also involve supply-side
substitutability (can suppliers switch production at limited or prohib-
itive costs from one product to another?).236 As explained above, the
identification of a subsidized product and like product is considered
insufficient to make this determination. On this basis, the Appellate Body
in EC – Large Civil Aircraft concluded that the panel should not simply
have accepted the US assertion that all large civil aircraft (LCA) models
are a single product, because this simply assumed that all LCAs competed
in the same market.237 The panel should have analysed the EU’s claim
that the LCA models, rather, compete in five distinct product markets.238 If
they compete in different markets, the displacement analysis has to be
conducted separately for the different product markets.
Finally, is there a minimum threshold requirement for the establishment
of displacement or impedance? The Appellate Body correctly observed that,
in contrast to some other types of serious prejudice, displacement or
impedance need not be ‘significant’ according to the text of Article 6.3.
Still, both should at least be ‘discernable’, because otherwise they could not
be clearly identified and amount to serious prejudice.239
4.2.3.4.1.1.1 Contextual support to show displacement or impedance

to third-country markets Article 6.4 of the SCM Agreement provides
contextual support to show impedance or displacement of trade to third-
country markets (Article 6.3(b)), but not to the subsidizing country
233
Appellate Body Report, US – Large Civil Aircraft, para. 1182. 234 Ibid., para. 1162.
235
Cross-price elasticity could be a useful tool for understanding whether products com-
pete in the same market. Appellate Body Report, EC – Large Civil Aircraft, para. 1134.
236
239
Ibid., paras. 1169, 1170, 1179.

market (Article 6.3(a)240).241 According to this provision, ‘displacement

or impeding exports shall include any case in which . . . it has been
demonstrated that there has been a change in relative shares of the
market to the disadvantage of the non-subsidized like product. . .’,
which is assessed ‘over an appropriately representative period sufficient
to demonstrate clear trends in the development of the market for the
product concerned, which, in normal circumstances, shall be at least
one year’.242
This ‘change in relative shares of the market’ includes any of the three
following situations: ‘(a) there is an increase in the market share of the
subsidized product; (b) the market share of the subsidized product
remains constant in circumstances in which, in the absence of the
subsidy, it would have declined; (c) the market share of the subsidized
product declines, but at a slower rate than would have been the case in the
absence of the subsidy’.243 Hence not only subsidies causing the expan-
sion of market share but also those keeping market shares constant or
slowing their decline could be challenged.
Under situations (b) and (c), the question whether the volume effects
are caused by the subsidy is part of the test, as both items call for a
comparison of the evolution in actual market share of the subsidized
product with the counterfactual situation ‘but for’ the subsidy.
On the other hand, and somewhat counter-intuitively, an increase in
the market share of the subsidized product (situation (a)) seems suffi-
cient to establish displacement or impedance, regardless of whether such
increase is caused by the subsidy. Indeed, according to the panel in
Indonesia – Autos, a prima facie case of displacement or impedance
could ‘arguably’ be made ‘simply by demonstrating that the market
share of a subsidized product has increased over an appropriately rep-
resentative period’.244 The panel in EC – Large Civil Aircraft agreed, and
even concluded that a relative increase in the subsidizing country’s
market share ‘will suffice to demonstrate that “the effect of the subsidy”
240
Although the text of Article 6.4 is unambiguous (‘[f]or the purpose of paragraph 3(b)’),
the Appellate Body in US – Large Civil Aircraft (paras. 1081, 1086) erroneously
considered, without further reflection, that Article 6.4 refers to both Article 6.3(a)
and 6.3(b).
241
Panel Report, Indonesia – Autos, para. 14.210. 242 Article 6.4.
243
Emphasis added.
244
Panel Report, Indonesia – Autos, paras. 14.209, 14.215. See also R. H. Steinberg and
T. E. Josling, ‘When the Peace Ends: The Vulnerability of EC and US Agricultural
Subsidies to WTO Legal Challenge’, 6:2 Journal of International Economic Law (2003),
369–417, at 387.

is to displace or impede exports’.245 Hence an assessment whether this

shift in market share is caused by the subsidy seems neither necessary
nor relevant. Contrary to all other items of Article 6.3 (and 6.4246),247 a
complainant could thus make a successful actionable subsidy claim
without demonstrating that its relative loss in market share in third
countries is effectively caused by the challenged subsidies.
Importantly, the panel in EC – Large Civil Aircraft held that complai-
nants can invoke this lower threshold only if their own like products are
not subsidized. The panel reached this interpretation on the basis of a
literal reading of the text of Article 6.4, which refers to a change in
market share to the disadvantage of the ‘non-subsidized like product’.248
It rejected the argument that this term simply refers to a like product
which does not benefit from the challenged subsidies. As a result, a sort
of ‘clean hand’ requirement is read into Article 6.4: complainants can
benefit from the lower threshold if their competing product is unsubsi-
dized. Even a minimal or non-specific subsidy would preclude recourse
to this lower threshold.249 In those circumstances, a complainant will
have to rely on the general serious prejudice analysis under Article 6.3(b)
and show that the market phenomena of displacement or impedance are
caused by the subsidy.250 By limiting the scope of Article 6.4, the panel
confined the potentially wide loophole resulting from its interpretation
of the substance of this provision.
In sum, the causation element could only be circumvented by the
complainant if three cumulative conditions are fulfilled. First, there
should be an increase in relative market share of the subsidized product
‘over an appropriately representative period sufficient to demonstrate
clear trends in the development of the market for the product con-
cerned’.251 Second, this market effect has to occur in third markets,
and not in the subsidizing market or in the complainant’s domestic
245
Panel Report, EC – Large Civil Aircraft, para. 7.1767 (emphasis added).
246
In its statement, the panel in Indonesia – Autos (para. 14.209) refers to Article 6.4 of the
SCM Agreement as such and not to Article 6.4(a) in particular, but its example refers to
the situation elaborated in Article 6.4(a) (i.e., increase in market share). Likewise, the
panel in EC – Large Civil Aircraft refers to an increase in market share (paras. 7.1767,
7.1769). As explained above, the text of items (b) and (c) seems to mandate a causation
analysis.
247
Except for price undercutting (see below). 248 Emphasis added.
249
Panel Report, EC – Large Civil Aircraft, para. 7.1770. 250 Ibid., para. 7.1771.
251
Article 6.4; see also Appellate Body Report, EC – Large Civil Aircraft, para. 1166;
Appellate Body Report, US – Upland Cotton, para. 487.

market.252 Third, the complainant’s like domestic product should be

fully unsubsidized.
The panel in EC – Large Civil Aircraft suggested that its reading of
Article 6.4 would fit the overall architecture of Article 6 as originally
negotiated, with different burdens for establishing serious prejudice:
ranging from a presumption of serious prejudice in Article 6.1, to a lesser
burden for complaints with ‘clean hands’ (i.e., an unsubsidized product),
and, finally, to the general requirement to demonstrate that the market
phenomena are caused by the subsidy. Yet it seems highly unlikely that this
sliding scale was in the negotiators’ mind. There seems no convincing
reason why negotiators would have aimed at making it easier to establish
serious prejudice in third markets than in the subsidizing country’s market.
If anything, one would rather expect the opposite in the light of the
historical context, since trade effects in third markets were for a long
time unchallengeable, whereas at least a non-violation claim could be
formulated against subsidies causing harm in the subsidizing market.
Moreover, one would have expected at least some discussion in the nego-
tiating history of such a so-called ‘clean hand’ requirement, since it presents
a fundamental departure from the rest of the SCM Agreement (and its
predecessors). Indeed, the fact that the complaining member equally sub-
sidizes its products is simply irrelevant under all other aspects of the SCM
Agreement (as amply shown by the parallel Boeing and Airbus disputes).253
The Appellate Body has not yet confirmed this reading of Article 6.4,
in which ‘the core of the analysis’254 (i.e., causation element) under
Article 6.3 is simply neglected. The Appellate Body’s emphasis on the
‘fundamental precept’255 of demonstrating causation suggests that it
would not easily agree with the reading advanced by these panels.256
In my view, Article 6.4 (item (a)) could be relied upon as contextual
support to show displacement or impedance in a two-step approach (i.e.,
252
Panel Report, EC – Large Civil Aircraft, 7.1760–7.1761. Recall, however, that the
Appellate Body erroneously held that Article 6.4 also applies to paragraph (a) (see
above n. 240).
253
The panel’s approach is difficult to reconcile with Article 6.7 of the SCM Agreement,
which explicitly states that serious prejudice shall not be present if the effect is caused by
certain factors unrelated to subsidization. Both provisions cannot be read in a harmo-
nious way if serious prejudice could arise under Article 6.4, even if the shift in market
share is explained by factors listed in Article 6.7.
254
Appellate Body Report, EC – Large Civil Aircraft, para. 1109. See also Appellate Body
Report, China – GOES, para. 142.
255
Appellate Body Report, US – Large Civil Aircraft, n. 1863.
256
See, e.g., Appellate Body Report, EC – Large Civil Aircraft, paras. 1229–1235.

a bifurcated approach), in which displacement or impedance is demon-

strated on the basis of Article 6.3(b) in conjunction with Article 6.4,
before the question of causality is tackled. However, as the Appellate
Body emphasized, the latter part of the analysis seems unavoidable:
‘Where a two-step approach is used under Article 6.3(a) and (b), and
displacement has been shown on a preliminary basis, the complaining
Member will have to establish, in addition, that such displacement is the
effect of the challenged subsidies’.257
4.2.3.4.1.1.2 Circumstances in which no displacement or impedance of

trade would arise A number of situations are listed in Article 6.7 of the
SCM Agreement in which serious prejudice in the form of displacement
or impedance within Article 6.3(a) and (b) shall not be considered to
arise.258 The panel in Korea – Commercial Vessels explained that this
provision is ‘concerned with alternative reasons (including, for example,
prohibition or restriction on exports, and situations of force majeure
affecting production, qualities, quantities or prices of the product avail-
able for export) for declines in the overall volume and/or market share of
the complaining Member’.259 Hence this provision somewhat resembles
the non-attribution requirement in CVD procedures.260 A decline in
volume or market share should not be attributed to subsidization if it is
caused by one of the six listed alternative reasons. But, contrary to the
CVD context, the complaining member seems not to be required affir-
matively to demonstrate that these alternatives are not present. Rather,
according to the case law, this provision ‘allows a subsidizing Member to
raise a defence to a displacement/impedance claim . . .’.261
In EC – Large Civil Aircraft, the panel rejected the EU’s argument that
the 9/11 terrorist attacks on the United States could qualify as force
majeure within the meaning of Article 6.7. In principle, these force
majeure events are related to supply-side shocks which explain the
downturn in the complaining member’s production, and not demand-
side shocks such as 9/11.262 Even if demand-side shocks were covered,
257
The Appellate Body did not qualify this statement by reference to Article 6.4. Ibid.,
para. 1170.
258
See also Panel Report, EC – Large Civil Aircraft, para. 7.1697.
259
Panel Report, Korea – Commercial Vessels, para. 7.586.
260
261
Panel Report, US – Upland Cotton, n. 1503 (emphasis added). See also Steinberg and
Josling, above n. 244, at 410.
262
Panel Report, EC – Large Civil Aircraft, para. 7.1699.

the panel correctly reasoned that it would only be accepted if it does not
affect the production of the defending member in a similar way.263
Indeed, if the event affects production in a similar manner, it can never
be the cause of any relative change in market share between both
members’ level of production.
4.2.3.4.1.2 Significant lost sales in the same market One of the elements
listed under Article 6.3(c) of the SCM Agreement refers to a volume
effect. Serious prejudice would arise if the subsidy leads to significant
‘lost sales in the same market’. The Appellate Body affirmed that this is a
relation concept: the subsidized firm(s) must have won sales that firm(s)
of the complaining member failed to obtain. It therefore requires con-
sideration of the behaviour of the firms of both countries.264 Depending
on the market characteristics, the assessment could be made either at the
level of aggregate sales (by supplier or customer, or on a country-wide or
global basis) or at the level of specific sales campaigns.265
The Appellate Body rightly observed that there could be some overlap
with the concepts of displacement or impedance under paragraphs (a)
and (b) of Article 6.3 if an analysis is made from a market-wide per-
spective. Nonetheless, two distinctions were highlighted.266 First, in
contrast to displacement or impedance, ‘lost sales’ does not refer to a
well-defined geographical market but can occur in any market, even the
world market. Second, the loss in sales should be ‘significant’ and not
simply discernable,267 which implies that the assessment can have quan-
titative as well as qualitative dimensions. In contrast, demonstration of
displacement or impedance calls for an assessment that is primarily
quantitative in nature.268 Another important difference is that a finding
of displacement is only present when a clear trend in the market can be
shown, whereas such a trend is not needed to show lost sales.269
4.2.3.4.1.3 Increase in the world market share of primary product or

commodity Article 6.3(d) of the SCM Agreement addresses a situation
263
Ibid., para. 7.1700.
264
265
It was left undecided whether lost sales could refer to one single unit of product. Ibid.,
n. 2627.
266
267
See, e.g., Panel Report, US – Large Civil Aircraft, 7.1788.
268
269
See, e.g., ibid., paras. 1051–1090.

in which subsidies to a primary product or commodity have the effect of

‘an increase in the world market share of the subsidizing Member’. Its
scope is thus explicitly confined to subsidies to primary products or
commodities.270
Regarding the increase in the world market share, two elements have to
be established.271 The subsidizing member’s share (i) should have increased
compared with its average share ‘during the previous period of three years’,
and (ii) this increase should follow ‘a consistent trend over a period when
subsidies have been granted’. The increase has to be observed in ‘the world
market’, which, according to the panel in US – Upland Cotton, refers to the
subsidizing member’s share in world production and not to its share in
world exports. This implies that the portion of subsidized products not
exported is also taken into consideration.272 In sum, Article 6.3(d) ‘calls for
an examination of the portion of the world’s supply that is satisfied by the
subsidizing Member’s producers’.273 In the compliance procedure, the
panel rejected Brazil’s claim under Article 6.3(d) because the data did not
reveal an increase in the US share of world production (financial year 2006)
as compared with the average US share in the previous three years (financial
years 2002–5).274
These demanding conditions to show an increase in the subsidizing
member’s world market share clearly limit the usefulness of this type of
serious prejudice. For instance, a subsidy programme boosting the sub-
sidizing member’s world market share at the time of introduction and
sustaining this market share in subsequent years would not seem to
violate Article 6.3(d).275 Generally speaking, only subsidy programmes
increasing a member’s share in world production year by year would be
captured under Article 6.3(d). In challenging subsidy programmes to
270
Footnote 17 indicates that Article 6.3(d) applies ‘[u]nless other multilaterally agreed
specific rules apply to the trade in the product or commodity in question’. Both parties
in US – Upland Cotton seemed to agree that this certainly does not refer to the
Agreement on Agriculture. See also Panel Report, US – Upland Cotton, n. 1508.
271
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), para. 10.261.
272
The panel defined ‘world market share’ as ‘the share of the world market supplied by the
subsidizing Member’. Panel Report, US – Upland Cotton, para. 7.1450.
273
274
275
An increase relative to the average share in three previous years should be demonstrated,
and this increase should follow a consistent trend over the time period of the subsidy
programme. Hence the comparison is not made with regard to the pre-subsidy period. For
example, in US – Upland Cotton the US market share in FY2006 was compared with its
average market share in previous subsidized years (FY2002–FY2005).

primary products or commodities, complainants might therefore more

easily be successful in claims under other types of serious prejudice.276
4.2.3.4.2 Price effects

4.2.3.4.2.1 Significant price undercutting The first type of ‘serious prej-
udice’ related to price effects refers to significant price undercutting
caused by the subsidized product as compared with the price of a like
product of another member in the same market (Article 6.3(c)). Such
price undercutting could thus also occur in the domestic market of the
subsidizing country.277
Similarly to other items of Article 6.3(c), serious prejudice is only
present if such price undercutting is ‘significant’. This qualifier was,
according to the panel in Indonesia – Autos, ‘presumably . . . intended
to ensure that margins of undercutting so small that they could not
meaningfully affect suppliers of the imported product whose price was
being undercut are not considered to give rise to serious prejudice’.278 In
the panel’s reading in Korea – Commercial Vessels, ‘the term “significant”
as a de minimis concept intended to screen out very small, unimportant
price effects that might be caused by subsidies but that would have no
real impact in the market’.279 Arguably, the standard set by the Appellate
Body that the effect should be ‘important, notable or consequential’
requires more than that it is simply large enough to be discernable.280
Regarding the demonstration of price undercutting, Article 6.5 clari-
fies that this includes any case in which it has been demonstrated
‘through a comparison of prices of the subsidized product with prices
of a non-subsidized like product supplied to the same market’. Such
comparison ‘shall be made at the same level of trade and at comparable
times, due account being taken of any other factor affecting price com-
parability’.281 Alternatively, and only ‘if such a direct comparison is not
possible, the existence of price undercutting may be demonstrated on the
276
Whereas Brazil failed to challenge the US subsidies for upland cotton under Article 6.3
(d), its claim under price suppression/depression (Article 6.3(c)) was successful (see
below Chapter 4, section 4.2.4.2).
277
Panel Report, Indonesia – Autos, paras. 14.241–14.256. 278 Ibid., para. 14.254.
279
This low-threshold interpretation was considered similar to the interpretation given in
the US – Upland Cotton case, in which the panel referred to ‘important, notable or
consequential’. Panel Report, Korea – Commercial Vessels, paras. 7.570–7.571.
280
See below n. 294; see also Appellate Body Report, EC – Large Civil Aircraft, paras. 1169,
1170, 1218.
281
Article 6.5 of the SCM Agreement. See, e.g., Panel Report, Indonesia – Autos, para.
14.244.

basis of export unit values’.282 According to the panel in EC – Large Civil

Aircraft, this provision performs a similar function to Article 6.4 and
thus provides for a substantively lower threshold to show price under-
cutting. If its own like products are unsubsidized (i.e., clean hands
requirement), a complainant could demonstrate price undercutting
simply on the basis of Article 6.5 without the need to show that this
has been caused by subsidization.283 As explained, it seems doubtful that
negotiators had this interpretation in mind.
4.2.3.4.2.2 Significant price suppression or depression The case law has

revealed that price suppression occurs when prices284 ‘either are pre-
vented or inhibited from rising (i.e., they do not increase when they
otherwise would have) or they do actually increase, but the increase is
less than it otherwise would have been’, whereas price depression covers
‘the situation where prices are pressed down, or reduced’.285 The com-
parison has to be made with the price of products ‘in the same market’,
which could be defined either narrowly or broadly (e.g., world market),
depending on the facts of the case.286
The Appellate Body acknowledged that situations of price suppression
and depression could overlap, but stressed that both concepts are dis-
tinct, also on a conceptual level. It correctly reasoned in US – Upland
Cotton (Article 21.5 – Brazil) that price suppression ‘presupposes a
comparison of an observable factual situation (prices) with a counter-
factual situation (what prices would have been) where one has to deter-
mine whether, in the absence of the subsidies (or some other controlling
phenomenon), prices would have increased or would have increased
more than they actually did’.287 Price depression, by contrast, does not
presuppose such comparison, as it could be ‘directly observed, in that
falling prices are observable’.288
282
Article 6.5 of the SCM Agreement.
283
Panel Report, EC – Large Civil Aircraft, para. 7.1771.
284
The term ‘price’ covers ‘the amount of money set for sale’ of the product in question or
the ‘value or worth’ of that product. Appellate Body Report, US – Upland Cotton, para.
423; Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 350.
285
Panel Report, US – Upland Cotton, para. 7.1277 and n. 1388; Appellate Body Report,
US – Upland Cotton, para. 423; Appellate Body Report, US – Upland Cotton
(Article 21.5 – Brazil), para. 350.
286
See, e.g., Panel Report, Korea – Commercial Vessels, paras. 7.565, 7.566.
287
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 351 (emphasis
added).
288
Ibid., para. 351.

Article 6.3(c) requires that the degree of price suppression or depression

should be significant, which the Appellate Body has understood as ‘impor-
tant, notable or consequential’.289 Obviously, ‘what needs to be significant is
the degree of price suppression [or depression], not necessarily the degree of
each factor used as an indicator for establishing its existence’.290 Moreover,
the panel in US – Upland Cotton (Brazil – Article 21.5) agreed with the
original panel that this degree of significance may ‘vary from case to case,
depending upon the factual circumstances’.291 With regard to the upland
cotton market, for instance, the required degree of significance was not set
too high because:
for a basic and widely traded commodity, such as upland cotton, a
relatively small decrease or suppression of prices could be significant
because, for example, profit margins may ordinarily be narrow, product
homogeneity means that sales are price sensitive or because of the sheer
size of the market in terms of the amount of revenue involved in large
volumes traded on the markets experiencing the price suppression.292
4.2.3.5 Causation
4.2.3.5.1 Causation standard What is the required strength of the
causal link between the challenged subsidies and the price or volume
effects spelt out under Article 6.3? The Appellate Body imported the
general (and vague) causation standard which it had articulated in the
context of the Safeguards Agreement: there should be a ‘genuine and
substantial relationship of cause and effect’ between the subsidies and
the alleged market phenomena.293 The term ‘genuine’ simply requires
that the nexus is ‘real’ or ‘true’, whereas ‘substantial’ concerns the relative
importance of subsidies in bringing about the market phenomena.294
The challenged subsidies should not be a sufficient cause of the price or
289
Appellate Body Report, US – Upland Cotton, para. 426; Appellate Body Report, US –
Upland Cotton (Article 21.5 – Brazil), para. 416; Panel Report, US – Upland Cotton,
paras. 7.1326–7.1328.
290
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 416 (emphasis
added).
291
292
Panel Report, US – Upland Cotton, paras. 7.1328–7.1329 (emphasis added); Panel
Report, US – Upland Cotton (Article 21.5 – Brazil), para. 10.50.
293
Appellate Body Report, EC – Large Civil Aircraft, paras. 1232, 1233, 1376; Appellate Body
Report, US – Large Civil Aircraft, para. 913; Appellate Body Report, US – Upland Cotton
(Article 21.5 – Brazil), para. 374 (see also ibid., para. 372); Appellate Body Report, US –
Upland Cotton, para. 438; Appellate Body Report, US – Wheat Gluten, para. 69.
294
The Appellate Body referred to the dictionary meaning of ‘substantial’: ‘[h]aving solid
worth or value, of real significance; solid; weighty; important, worthwhile’ and ‘[o]f

volume effects, but only a substantial cause thereof.295 Hence a sufficient

condition seems not necessary, but a necessary condition (sine qua non)
seems also not sufficient in itself.296 Panels should ensure that the effects
of other factors on the market phenomena are not improperly attributed
to the challenged subsidies and do not dilute the ‘genuine and substantial
link’ (i.e., a non-attribution analysis).297
4.2.3.5.2 Methodology: unitary or bifurcated analysis The SCM

Agreement does not indicate whether the adverse volume or price effects
(i.e., the market phenomena spelt out under Article 6.3) should or could
be analysed separately from the causation element. Panels therefore have
‘a certain degree of discretion in selecting an appropriate methodol-
ogy’298 for tackling the causation element, including the non-attribution
analysis.299 In principle, panels are free to opt for a bifurcated approach
(i.e., a two-step approach), in which they analyse both steps separately
(as the panels did in US – Upland Cotton and EC – Large Civil Aircraft),
or to adopt a unitary approach and analyse both steps together (as the
panels did in US – Upland Cotton (Article 21.5 – Brazil) and US – Large
Civil Aircraft).300 If a bifurcated approach is adopted, the panel first seeks
to identify one of the market phenomena identified in Article 6.3 and, as
a second step, will assess whether these are caused by the subsidies. If a
unitary approach is adopted, the panel simply assesses whether subsidies
cause one of the identified market phenomena.301
Although a bifurcated approach is allowed, the Appellate Body has
several times expressed its preference for the unitary approach.302 After
all, ‘consideration of the effects of the challenged subsidies is intrinsic to
the identification of those market phenomena’.303 First, any attempt to
ample or considerable amount or size; sizeable, fairly large’. Appellate Body Report,
US – Large Civil Aircraft, nn. 1865, 1866.
295
296
Ibid., para. 1233; Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil),
para. 374.
297
See, e.g., Appellate Body Report, US – Large Civil Aircraft, para. 914.
298
Appellate Body Report, US – Upland Cotton, paras. 431, 436; Appellate Body Report,
US – Upland Cotton (Article 21.5 – Brazil), para. 368.
299
See, e.g., Appellate Body Report, US – Large Civil Aircraft, para. 914.
300
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 354; Appellate
Body Report, EC – Large Civil Aircraft, paras. 1107–1111, 1163.
301
302
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 354; Appellate
Body Report, EC – Large Civil Aircraft, paras. 1107–1111, 1163.
303
Appellate Body Report, EC – Large Civil Aircraft, para. 1109 (emphasis added).

identify one of the market phenomena in Article 6.3 in isolation can only
be preliminary in nature, because it should still be demonstrated that
these are the effect of the subsidy (causation element). For instance, price
depression could, at first sight, easily be identified in data but compar-
ison with a counterfactual situation (what prices would have been with-
out subsidies) will anyway be required to establish the causational
requirement.304 Second, the opposite problem might also occur: ‘by
artificially leaving aside the question of whether the market phenom-
enon is the effect of the subsidy, one could overlook market phenomena
that are in fact occurring’.305 Here, it seems that the Appellate Body had
in mind situations in which there would, for example, be no price
depression observable in the data but in which subsidies still caused
price suppression.306
If a unitary approach is adopted, the causation assessment could be
conducted by a ‘but for’ test (also called a counterfactual test), which
seeks to identify what would have occurred ‘but for’ the subsidies.307 To
be sure, in the Appellate Body’s view, a ‘but for’ test, strictly speaking,
only ensures that the challenged subsidies are a necessary condition
explaining the market phenomena: would these market phenomena
have existed but for the subsidies? Hence the ‘but for’ test might suffice
in those cases where it shows that the subsidy is not only a necessary
condition but also a substantial cause. On the other hand, it would not
suffice in those circumstances where it merely detects subsidies as a
necessary cause that is ‘too remote and other intervening causes sub-
stantially account for the market phenomenon’.308 This illustrates,
according to the Appellate Body, that a proper non-attribution analysis
should be conducted, in which the effect of those other intervening
causes is filtered out.
Arguably, the Appellate Body’s causation standard and ‘but for’ test
set out above reflect a bifurcated rather than a unitary approach, even
though it purports to be the opposite. The question whether the market
phenomenon (i.e., price and/or volume effect) would have taken place
‘but for’ the subsidies can only be answered if that market phenomenon
304
305
306
See, for instance, Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil),
para. 351.
307
Appellate Body Report, EC – Large Civil Aircraft, para. 1233, n. 2655.
308
Ibid., para. 1233.

is first detected.309 The fact that its ‘but for’ test could reveal subsidies
that are only a remote cause of the market phenomenon amply illustrates
that the market phenomenon in the Appellate Body’s reasoning is not
fully explained by the subsidies. Thus it can also not be simply derived
from an assessment of the effect of the subsidies. The Appellate Body
seems to assume that it can detect an overall actual market phenomenon,
which only needs to be partly explained by the subsidies. The Appellate
Body’s causation standard and ‘but for’ test could be understood on the
basis of a hypothetical example. Suppose that data shows that prices over
the reference period were depressed by 10 per cent and that it could be
demonstrated that price depression would only have amounted to 3 per
cent ‘but for’ the subsidies. Here, subsidies are indeed not a sufficient
condition to explain the full market phenomenon, but they are arguably
a ‘substantial’ cause thereof. Yet this is clearly a bifurcated test: it
requires the full market phenomenon to be identified (i.e., 10 per cent
price depression) and a subsequent causation analysis showing how
much subsidies contributed to this phenomenon (i.e., 7 per cent). A
proper unitary test, which is for the reasons spelt above favoured by the
Appellate Body, formulates the ‘but for’ test in a different way: what
would have happened to price and/or volume (instead of price and/or
volume effect) ‘but for’ the subsidies? In several instances, the Appellate
Body in EC – Large Civil Aircraft correctly formulated this ‘but for’ or
counterfactual approach: ‘the actual market situation [has to be com-
pared] with the market situation that would have existed in the absence
of the challenged subsidies’.310 Such a ‘but for’ test, if properly conducted
(other factors are held constant), only reveals the effect of the subsidies
on prices and/or volume. Turning back to our hypothetical example, it
would only show that subsidies depressed prices by 7 per cent and would
not reveal that the overall price depression was 10 per cent. Hence the
causation standard set out by the Appellate Body is in fact not mean-
ingful in a proper unitary test, given that the effect of the subsidies
cannot be assessed in the light of an overall market phenomenon, as
the latter is not separately identified. The unitary approach is preferred
by the Appellate Body precisely because it seems to recognize that the
overall market phenomenon is simply not detectable. The unitary test
only reveals the contribution made by the subsidies on their own, as
sufficient condition, to prices or volume. This should be assessed to
309
Indeed, the counterfactual situation is compared to the factual situation.
310
Appellate Body Report, EC – Large Civil Aircraft, paras. 1110, 1163.

determine whether it amounts to the required level of price or volume

effects (e.g., significant price suppression or depression).
The discrepancy between a proper unitary approach by the Appellate
Body and its causation standard seems to explain other frictions in its
reasoning. First, the Appellate Body remains unclear on the need for a
separate non-attribution test if a unitary approach is adopted. On the
one hand, as stated above, the fact that its ‘but for’ test could reveal that
subsidies are only a remote cause implies that a proper non-attribution
analysis should be conducted in which the effect of other intervening
causes are filtered out.311 On the other hand, the same Appellate Body
report rightly stated that a proper unitary approach does not need a
separate non-attribution test, given that the economic models relied on
to conduct such a test single out only the effect of the subsidy on price or
volume, while keeping other factors constant.312 Second, the Appellate
Body remains equally unclear on whether the subsidies should on their
own create the required level of price or volume effects (i.e., ‘significant’).
As explained above, the price effects (and some volume effects) have to
be significant according to Article 6.3. On the one hand, the Appellate
Body’s causation standard (‘genuine and substantial link’) might suggest
that subsidies only need to be a substantial cause, and not a sufficient
cause, of, for instance, significant price suppression.313 On the other
hand, the Appellate Body in US – Upland Cotton (Brazil – Article 21.5)
correctly seemed to hold that the subsidies should on their own be
sufficient to create significant price suppression.314 Indeed, the text of
Article 6.3 plainly requires that the effect of the subsidy is significant
price suppression.
4.2.3.5.3 The nature of the causation analysis: quantitative and/or

qualitative? Demonstrating that subsidies are a ‘genuine and substan-
tial cause’ of serious prejudice is a ‘fact-intensive exercise, and one that
inevitably involves extensive, case-specific evidence’.315 The Appellate
311
Ibid., para. 1233.
312
Ibid., para. 1265; Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil),
para. 375.
313
In the context of the Safeguards Agreement, in which the ‘genuine and substantial link’
standard was developed, the Appellate Body decided that the causation standard does
not require that ‘increased imports on their own must be capable of causing serious
injury’. Appellate Body Report, US – Wheat Gluten, paras. 67–70 (emphasis in
original).
314
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 360–366.
315
Appellate Body Report, US – Large Civil Aircraft, para. 915.

Body accepts that such an analysis ‘may be highly quantitative, or

predominantly qualitative in nature, or it may involve both quantitative
and qualitative elements’.316
The following qualitative elements are particularly relevant.317 First,
the nature and duration of the challenged subsidies are clearly impor-
tant. Subsidies tied to production, and certainly those tied to exports, are
more likely to cause adverse effects.318 It has equally been accepted that a
relatively small amount of R&D subsidies could cause adverse effects
because of its multiplier effect. Showing serious prejudice resulting from
untied subsidies, which simply increase non-operating cash flow, will be
more difficult, though not impossible.319 Second, the magnitude of
subsidies in absolute terms and relative terms (i.e., compared with, e.g.,
prices, production costs, market size) plays a role, although relatively
small subsidies could, depending on the nature of the subsidies and
market, generate significant effects.320 Third, the competitive conditions
in the market at issue are relevant. Serious prejudice is more likely to be
caused by subsidization in markets where market players exercise market
power (e.g., a duopoly) rather than in perfectly competitive markets.
These qualitative elements in and of themselves do not suffice, how-
ever, to establish the requisite causal connection. In US – Large Civil
Aircraft, the Appellate Body found serious prejudice (in the form of lost
sales) for those sales campaigns where evidence of fierce price competi-
tion was present and where they could not be explained by non-price
factors.321 The fact that serious prejudice was demonstrated without
resort to quantitative analysis appears, however, highly dependent on
the facts of this case (i.e., duopoly).
In most cases, and certainly in those dealing with price or volume
effects in competitive markets, the qualitative elements will have to be
supplemented with a quantitative analysis.322 Trend analysis, showing a
correlation between price or volume and subsidy levels, is considered
relevant, although it is neither necessary nor sufficient. Indeed, the
316
Ibid., para. 1019. 317 Ibid., paras. 1250 – 1261.
318
See also Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil).
319
Appellate Body Report, US – Large Civil Aircraft, paras. 1336, 1348. As explained in the
next section, even if they are insufficient to cause serious prejudice, their effect might be
cumulated to subsidies causing such prejudice if they have a genuine link to the same
market phenomena.
320
Ibid., paras. 1254–1255. 321 Ibid., para. 1261.
322
Unless, arguably, for demonstrating lost sales, as this could be based on specific
evidence of lost sales.

Appellate Body seems to recognize that economic analysis should, in

principle, be part of the serious prejudice determination, since other
relevant factors cannot establish a causal relationship. Although the
Appellate Body in the original US – Upland Cotton was still rather
flexible on the need for economic modelling, it emphasized in the
compliance procedure that an analysis of price suppression ‘would
normally include a quantitative component’, and the complexity of
such models could not be an excuse for panels ‘to remain agnostic
about them’.323 A panel should assess the probative value it accords to
economic simulations or models presented to it.324,325
To conduct this quantitative analysis, econometric models (e.g.,
regression analysis) establishing statistically the relationship between
the subsidy and price or volume effects cannot be used, because there
is a lack of sufficient observations of the independent variable (i.e., the
subsidy).326,327 Instead, complainants have advanced a particular type
of simulation model, which is known as a ‘calibrated model’.328 In
creating such models, assumptions are made of (i) the presence of a
relationship between subsidies and price or volume; (ii) the nature of this
relationship, which is reflected in the parameters.329 As explained by the
panel in US – Large Civil Aircraft, ‘[c]alibration of the model involves
choosing values for the parameters that ensure that the model assump-
tions and the observed data are consistent.’330 In contrast to econometric
323
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 356, 357, 358,
434. See also W. J. Davey and A. Sapir, ‘United States – Subsidies on Upland Cotton
Recourse to Article 21.5 by Brazil, WT/DS267/AB/RW (2 June 2008)’, 9:1 World Trade
Review (2010), 1–19.
324
325
In EC – Large Civil Aircraft, the Appellate Body even recognized that any unitary
analysis ‘requires the adjudicator to undertake a modelling exercise as to what the
market would like in the absence of subsidies’ (para. 1110).
326
See A. Sapir and J. Trachtman, ‘Subsidization, Price Suppression, and Expertise: Causation
and Precision in Upland Cotton’, 7:1 World Trade Review (2008), 183–209, at 201–5.
Econometric modelling is used in trade remedy cases (see below Chapter 5, section 5.2.2.3.).
327
Steinberg and Josling do, however, not exclude the use of regression analysis. They
tackled the problem of the lack of sufficient observations by pooling time series and
cross-section data (i.e., aggregating ten potential claims). See Steinberg and Josling,
above n. 244, at 397–400. As Sapir explains, pooling of data is certainly not a precise tool
to be used in a specific claim; its appropriateness depends on how far the specific claim
is from the average of the other claims.
328
Such a model was advanced by Brazil in US – Upland Cotton and by the EU in US –
Large Civil Aircraft.
329
Panel Report, US – Large Civil Aircraft, Appendix VII.F.2, para. 61.
330
Ibid., Appendix VII.F.2, paras. 61, 62.

modelling, which use statistical techniques to test validity, simulating

models simply assume that there is a relationship between subsidies and
lower prices or volumes but do not test it.331,332 It is, however, possible to
test the robustness of the model’s prediction.333 The parameter values
(e.g., elasticities) that are included in the model are based on estimations,
which could also differ widely between the parties.334 In the case where a
panel accepts both the model’s structure and parameter values, simulat-
ing models can be useful to quantify the effect of the challenged subsidy
on price or volume.335
4.2.3.5.4 Collective assessment of subsidies and their effect The seri-

ous prejudice test requires demonstration that a subsidy has a genuine
and substantial link (i.e., a causation standard) to one of the market
phenomena identified in Article 6.3. What if a subsidy (‘subsidy B’) is, on
its own, insufficient to cause such serious prejudice, but together with
other subsidies (‘subsidy A’) reaches this threshold? The Appellate Body
has explained that, at least in two circumstances, such subsidies could be
collectively assessed to show serious prejudice.
First, aggregation of subsidies (subsidy A + subsidy B) is allowed in the
case where they share a similar design, structure, and operation, so that
there is a reasonable likelihood that the causal mechanism and market
phenomena are the same.336 The relevance of such an aggregated anal-
ysis is that it may establish the required causal nexus with the market
phenomena (price or volume effect), where no such link would have
been established on the basis of an isolated analysis of each of the
331
Ibid.
332
For instance, the model used in the US – Upland Cotton case assumed a competitive market
structure and that upland cotton is a homogeneous product. H. Vandenbussche,
‘Comment – Upland Cotton Case’, 7:1 World Trade Review (2008), 211–17, at 212–15.
333
See Sapir and Trachtman, above n. 326, at 183–209; see also Steinberg and Josling,
above n. 244, at 393.
334
See, for instance, the difference in parameter values advanced by the parties in US –
Upland Cotton.
335
The panel in US – Large Civil Aircraft (Appendix VII.F.2, paras. 68–76) rejected the
EU’s simulation model, because its predictions were not considered consistent with
reality.
336
Relevant are the nature and design of the subsidies, the implications of that nature and
design for the operation of the subsidies, their relationship to the subsidized product,
and the structure of the market in which that product competes. Appellate Body Report,
US – Large Civil Aircraft, para. 1291.

individual subsidies.337 Under this first scenario, subsidies A and B share

the same nature and are insufficient on their own to cause serious
prejudice, but, when aggregated, do cause serious prejudice. Such an
aggregated analysis was conducted by the panel in US – Upland Cotton. It
aggregated the effect of all price-contingent subsidies for its price sup-
pression analysis, but declined also to add non-price subsidies, because
they had a more attenuated nexus with the world price of cotton.338
Second, cumulation of the effects of subsidies (effect of subsidy A + effect
of subsidy B) occurs when one subsidy (or group) (subsidy A), by itself,
establishes the required causal nexus (i.e., a genuine and substantial link) to
the market phenomena, but could still be cumulated with another subsidy
(or group) (subsidy B) that would not by itself establish this nexus (i.e., a
genuine but not substantial link). The relevance of such a cumulation
exercise, which is conducted after it has been shown that subsidy A has
caused serious prejudice, is that it allows adding the adverse effect of
dissimilar subsidies (i.e., subsidy B) that by themselves are insufficient to
cause adverse effects. Subsidy B does not have to be similar to subsidy A or
to operate along the same causal pathway, but merely needs to have a
‘genuine causal connection’ to the same market phenomena.339 Under
this second scenario, cumulation of the effect of subsidies of a different
nature takes place, in the case where subsidy A suffices by itself to cause
adverse effects and subsidy B has a ‘non-trivial’ or ‘meaningful’ effect on the
same market phenomena.340 The Appellate Body considered the scope for
such cumulation to be larger in markets where the recipient exercises
market power (e.g., a duopoly), because the recipient ‘may be more likely
to be able to take advantage of potential interaction between different
subsidies, and to exploit these effects to the disadvantage of its competitors,
than would be the case in a perfectly competitive market’.341 Such cumu-
lation was, therefore, conducted in both EC – Large Civil Aircraft and US –
Large Civil Aircraft with regard to those subsidies for which a genuine
connection to the same market phenomena had been established.342
337
Ibid., para. 1291.
338
Panel Report, US – Upland Cotton, paras. 7.1289, 7.1291, 7.1299, 7.1303; Appellate
Body Report, US – Large Civil Aircraft, para. 1286.
339
The magnitude of subsidy B is still relevant, but somewhat less than when it has to be
shown that this subsidy as such causes serious prejudice. Even the effect of relatively
small subsidies could thus be cumulated. Appellate Body Report, US – Large Civil
Aircraft, paras. 1339, 1348.
340
342
Ibid., paras. 1347–1348; Appellate Body Report, EC – Large Civil Aircraft, paras. 1391,
1400, 1408.

The Appellate Body left open whether a collective assessment could

also be conducted in a third scenario, in which subsidies A and B are of a
different nature and are on their own insufficient to cause adverse effects,
but collectively meet the threshold. The Appellate Body’s wording (that
there are ‘at least’ two approaches343) suggests that it left open the option
of such a third scenario. Its lesson to future panels also suggests that the
Appellate Body would not accept that a member could circumvent its
actionable subsidy obligations by providing many small subsidies of a
different nature, which collectively cause adverse effects. Future panels
are mandated to
employ an approach that will enable it to take due account of all of the
subsidies that provide a relevant and identifiable competitive advantage
to the recipient and its products in the market and that relate to alleged
adverse effects phenomena. Only by doing so can a panel ensure a full
appreciation of all of the challenged subsidies that may be contributing,
or conducing, to the serious prejudice.344
At the same time, the Appellate Body warned that panels must ensure
that complainants show that each of the challenged subsidies is a ‘gen-
uine cause’ for the market phenomena, and that the challenged subsidies
taken together are ‘a genuine and substantial cause’ of such adverse
effects.345 This guidance suggests that the Appellate Body would, at
least in theory, accept collectively assessing small subsidies of a different
nature, as long as the complainant can show that each of them is a
‘genuine’ cause and that they collectively amount to a ‘substantial’
cause of adverse effects. In practice, it might, however, be difficult for
the complainant to meet this burden of proof. Recalling the Appellate
Body’s jurisprudence on cumulation, it might be more readily accepted
in markets where recipients exercise market power than in competitive
markets.
4.2.4 Analysis of relevant case law

4.2.4.1 Indonesia – Autos
The first substantive serious prejudice claims were advanced by the EU
and the United States in Indonesia – Autos. They argued that subsidies
343
Appellate Body Report, US – Large Civil Aircraft, paras. 1284, 1290.
344

granted to the Indonesian automotive industry undercut their prices and

impeded or displaced their imports into the Indonesian market.346
First, in examining price undercutting, the panel found that the price
difference between the EU’s imported cars and domestic like cars was
significant even when differences in physical characteristics were taken
into account.347 Because Indonesia conceded that its subsidies were
responsible for the significant price undercutting, the panel had no
difficulty in concluding that Indonesia’s car subsidy programme caused
serious prejudice.
Second, regarding the alleged volume effects, the panel analysed
whether sales of like products of the complainants would have been
higher in the subsidizing country ‘but for’ the subsidy.348 The panel
concluded that both complainants failed to advance sufficient evidence
to underpin their claim.349 A declining market share was considered
highly relevant but not decisive to show that ‘but for’ the subsidies sales
would have been higher in absolute terms.350 Because Indonesia made a
convincing argument that the subsidy itself could have caused a rapid
expansion of the market, leading to a higher market share for the
subsidized domestic product, even a dramatic drop in the EU’s market
share was considered insufficient to demonstrate that it also lost sales in
absolute terms.351 In sum, according to this panel, impedance or dis-
placement in the subsidizing market is only present under Article 6.3(a)
if complainants’ sales in absolute terms would have been higher in the
absence of the subsidies, even though a relative loss might be relevant to
showing an absolute loss. Hence the fact that domestic firms capture the
extra domestic demand created by subsidization does not amount to
serious prejudice to foreign producers. In contrast, by virtue of Article
6.3(b) in conjunction with Article 6.4, a prima facie case of displacement
or impedance in third-country markets would be established if complai-
nants would have lost sales in relative terms (i.e., market share) as a
result of the subsidy to domestic producers.352 This panel thus drew an
important distinction between Article 6.3(a) and Article 6.3(b). It is,
however, unclear whether the Appellate Body would agree with the need
to demonstrate lost sales in absolute terms under Article 6.3(a), as
demonstration of displacement appears to focus on changes in market
346
Panel Report, Indonesia – Autos, paras. 14.209, 14.215.
347
Ibid., paras. 14.246, 14.253. 348 Ibid., para. 14.218.
349
Ibid., paras. 14.222, 14.229, 14.235, 14.236. 350 Ibid., para. 14.211.
351
Ibid., para. 14.222. 352 Ibid., para. 14.215.

share. To this end, it (erroneously) applied Article 6.4 not only to Article
6.3(b), but also to Article 6.3(a).353
4.2.4.2 US – Upland Cotton

In US – Upland Cotton, Brazil successfully claimed that US cotton
subsidies caused serious prejudice to its cotton producers. This case
law offers important guidance on how price suppression has to be
demonstrated in competitive markets. The analysis focuses on the com-
pliance proceedings.
Adopting a unitary approach, the panel in US – Upland Cotton
(Brazil – Article 21.5) assessed whether, but for the cotton subsidies,
the world market price for upland cotton would have increased signifi-
cantly, or would have increased by significantly more than was in fact the
case.354 As the Appellate Body explained, such a price-suppressing effect
would result from additional, ‘marginal’, production induced by subsi-
dization.355 To conduct the unitary ‘but for’ analysis, a number of differ-
ent relevant factors were considered collectively.
First, a group of factors revealed whether it was likely that the subsidy
(i.e., marketing loan and countercyclical payments), through creating
marginal production, could have a significant price-suppressing effect.
The large US share in world production and exports implied that the
United States exerted ‘a substantial proportionate influence’ on the
world market for upland cotton.356 Further, the nature (i.e., ‘structure,
design, and operation’) as well as magnitude of the subsidy were relevant.
Marketing and countercyclical payments affect US production as a result
of their mandatory and price-contingent nature and their revenue-
stabilizing effect.357 Obviously, ‘[a] large subsidy that is closely linked
to prices of the relevant product is likely to have a greater impact on
prices than a small subsidy that is less closely linked to prices’.358 Next,
total production costs (compared with market revenue) suggested that
353
Appellate Body Report, US – Large Civil Aircraft, para. 1162 (see above n. 240). See also
the Appellate Body Report in EC – Large Civil Aircraft, as discussed in Chapter 4,
section 4.2.4.3.
354
355
356
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 10.54–10.58; Panel
Report, US – Upland Cotton, para. 7.1348.
357
Report, US – Upland Cotton, paras. 7.1289, 7.1349; Appellate Body Report, US – Upland
Cotton (Article 21.5 – Brazil), para. 362.
358

the subsidies were an important factor affecting the economic viability of

US cotton production.359 Total production costs included opportunity
costs, whose calculation was based on total costs (and not merely on
variable costs) because the time period of the subsidy programme called
for a medium- to long-term analysis.360,361
Second, trend analysis was used to detect whether there was a correlation
between subsidy levels and price evolutions. The Appellate Body under-
stood that mere correlation is not sufficient to demonstrate causation.
Indeed, the price trend might very well be explained (i.e., caused) in full
or in part by other factors. Equating correlation to causation would lead to
so-called Type II errors: accepting the hypothesis of causality when it is
actually false. Nonetheless, because ‘one would normally expect a discern-
ible correlation between significantly suppressed prices and the challenged
subsidies’, such trend analysis is deemed an ‘important factor in any
analysis’ of price suppression.362 At the same time, the compliance panel
rightly concluded that the absence of a general decline in prices over the
subsidization period does not exclude price suppression.363 Indeed, prices
could still have been higher ‘but for’ the subsidies.364
Third, and finally, economic modelling was used to estimate effec-
tively whether prices were suppressed as a result of subsidization. Brazil
introduced an economic simulation model in the compliance procedure.
This two-country (United States and rest of the world) demand and
supply log-linear displacement model simulated a counterfactual situa-
tion: how much would world prices, production, and imports or exports
in the United States and the rest of the world change if the subsidies were
to be removed?365 The United States questioned the very structure of the
model itself (without, however, suggesting its own alternative model) as
well as the parameter values (e.g., demand and supply elasticities) and
proposed its own set of parameter values. The panel noted ‘the advantage
of the modeling approach that Brazil has chosen to adopt’, but remained
359
Report, US – Upland Cotton, paras. 7.1353–7.1354.
360
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 10.167–10.176;
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 423, 427–
428; Appellate Body Report, US – Upland Cotton, para. 453.
361
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 423, 427–428.
362
363
364
365
Ibid., para. 10.199. The Arbitrator relied on this model to quantify the effect of a
withdrawal of US subsidies. See below Chapter 5, section 5.1.3.2.2.

‘mindful of the criticism by the United States that [it] “has no foundation
within economic circles”’.366 Regarding the parameters, it explicitly
refrained from taking a position on which parameter values were appro-
priate. The panel simply concluded that under all simulations run by the
parties (thus also with the US parameter values) price suppression was
the outcome.367 To be sure, the panel should have reached this conclu-
sion only if it were confident about the economic underpinning of the
model and that price suppression was significant also with US parameter
values (which resulted in price suppression of 1 per cent). The Appellate
Body simply concluded that such price suppression would have been
considered significant by the panel.368 On the basis of all three discussed
elements collectively, the panel in the compliance procedure reached the
conclusion that US cotton subsidies caused significant price suppression.369
The Appellate Body, finally, read a non-attribution assessment into
the panel’s analysis, although the latter had rather correctly dismissed
the need for such an analysis given its unitary approach.370 The United
States argued that China’s significant role in the market for upland
cotton explained the price suppression. The Appellate Body dismissed
this argument, because the increase in imports and consumption of
upland cotton in China would rather have put upward pressure on the
price.371 The finding of price suppression caused by the cotton subsidies
was, therefore, not altered by the non-attribution analysis.
4.2.4.3 EC – Large Civil Aircraft

In EC – Large Civil Aircraft, the United States successfully demonstrated
that certain types of subsidy to Airbus caused displacement and lost sales
to Boeing.
The panel opted for a bifurcated approach to analysing the US
claims.372 The presence of the market phenomena (i.e., displacement
and lost sales) was first assessed, and only in a second step was it analysed
whether these were the effect of the subsidy (causation element). In
essence, the subject for its displacement analysis was not the challenged
366
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 10.219, 10.220.
367
368
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 365, 435.
369
Panel Report, US – Upland Cotton (Article 21.5 – Brazil), paras. 10.244–10.257.
370
Appellate Body Report, US – Upland Cotton (Article 21.5 – Brazil), para. 375; Panel
Report, US – Upland Cotton (Article 21.5 – Brazil), para. 10.243.
371
372
Appellate Body Report, US – Upland Cotton, para. 433, n. 521.

subsidy itself but the subsidized product (i.e., Airbus LCAs).373 Hence the
panel first assessed whether Airbus caused displacement and/or lost sales
to Boeing. In a second step, the panel analysed whether these market
phenomena were caused by the challenged subsidies. In essence, the
panel here analysed the causal link between the subsidy and the presence
of the subsidized product. Because the subsidized product (i.e., Airbus
LCAs) caused displacement and lost sales (i.e., to Boeing LCAs) (step 1)
and the subsidized product (i.e., Airbus LCAs) would not have been
similarly present but for the challenged subsidies (step 2), the panel
concluded that the challenged subsidies caused displacement and lost
sales to Boeing.
Although the Appellate Body reiterated its preference for a unitary
approach, it considered the panel’s two-step analysis to be permissi-
ble.374 In the following sections, both steps of the analysis as reviewed by
the Appellate Body are further explored.
4.2.4.3.1 Displacement and significant lost sales Displacement is

present if there is ‘a substitution effect between the subsidized product
and the like product of the complaining Member’ during the reference
period.375 To establish such a substitution effect, a focus on market
shares has to complement an examination of trade volumes, as the latter
is insufficient to detect substitution effects when total consumption in
the market increases or decreases. Hence a gain in market share to the
subsidized product at the expense of the complaining member has to be
established.
Obviously, the presence of a duopoly clearly facilitated this demon-
stration, because ‘a decrease in the market share of one competitor will
be reflected in an increase in the market share of the other competi-
tor’.376 As the Appellate Body acknowledged, the analysis would become
more difficult when more than two competitors are present in the
market, as it will be necessary to analyse ‘more fully the market share
data of various competitors in order to determine whether the substitu-
tion effect is between the subsidized product and the like product in the
complaining Member’.377 The substitution effect has to be reflected in a
373
Observe that the Appellate Body’s definition of displacement allows such bifurcation, as
it refers to displacement by the subsidized product (see above n. 228).
374
Appellate Body Report, EC – Large Civil Aircraft, paras. 1163–1164.
375

clearly discernable trend during the reference period and could not
simply be based on an end point-to-end point analysis.378
Examining the presence of a decline in Boeing’s market share,379 the
Appellate Body found displacement in the single-aisle and twin-aisle
LCA product markets in the subsidizing member (i.e., the EU) and some
third markets, but rejected such displacement in a number of other
markets. With regard to the Indian market, a claim of threat of displace-
ment was relevant because the data referred to orders instead of effective
deliveries. Yet data on orders did not support a finding of displacement
in any of the product markets.380
Turning to significant lost sales, it had to be established that the sub-
sidized firm (i.e., Airbus) had won significant sales that the complaining
member’s firm (i.e., Boeing) failed to obtain. The focus on specific sales
campaigns (each involving the sale of a number of LCAs) was appropriate,
although an aggregated approach would also have been permissible. The EU
did not appeal the panel’s finding that Boeing lost sales to Airbus involving
purchases by a number of airlines and that these lost sales were ‘significant’
(i) in the light of the number of aircraft and the dollar amounts involved in
the sales, (ii) their strategic importance, (iii) the learning effects and econo-
mies of scale they generated, and (iv) the advantages of the incumbent
supplier provided by the sales.381 The reason why Boeing lost these sales
to Airbus (e.g., price or other factors) was rightly considered irrelevant
under this first step of the analysis.
4.2.4.3.2 Causation In this second step of the bifurcated test, it had to

be determined whether the identified market phenomena (i.e., displace-
ment and significant lost sales) were the effect of subsidies. In the
Appellate Body’s understanding, the panel appeared to have adopted a
‘but for’ test to analyse causation, which it considered difficult to under-
stand because such a test is normally part of a unitary approach.382 The
Appellate Body failed to understand that a ‘but for’ test could very well be
part of a bifurcated test, depending on its exact formulation. In fact, as
explained above, the Appellate Body’s own analysis of the ‘but for’ test
reveals a bifurcated rather than a unitary approach. In a unitary
approach, the ‘but for’ test examines what the price and volume would
378
Ibid., paras. 1166, 1179. 379 Ibid., para. 1179. 380 Ibid., para. 1204.
381
See Panel Report, EC – Large Civil Aircraft, para. 7.1845; see also Appellate Body
Report, EC – Large Civil Aircraft, paras. 1212, 1219.
382
Appellate Body Report, EC – Large Civil Aircraft, para. 1234, n. 2655.

be but for the subsidies. Hence the market phenomena and causation
element are simultaneously identified. In contrast, as part of a bifurcated
approach adopted by the panel, the relevant question was whether the
previously identified market phenomena (i.e., displacement and signifi-
cant lost sales) would have been present ‘but for’ the subsidies. The
Appellate Body agreed with the panel’s substantive analysis set out next.
Adopting the so-called ‘product’ theory of causation, the panel exam-
ined whether, but for the subsidies, Airbus would have been able to
develop and launch the LCA ‘it did when it did’.383 If not, Airbus could
not without subsidization have displaced Boeing or won sales as identi-
fied under the first step of the analysis. Next to qualitative evidence, the
panel examined the Dorman Report introduced by the United States.
This report simulated the impact of the launch aid/member state financ-
ing (LA/MSF) on the decision to develop and launch a particular LCA by
calculating its net present value in the absence of LA/MSF. The panel
found that this report demonstrated that the provision of LA/MSF was
‘likely to change the behaviour of the recipient with respect to a decision
to launch a LCA by increasing the likelihood of an affirmative decision to
go forward with the launch’.384 Although it recognized that public state-
ments on the need for subsidies to launch a LCA might involve some
self-interest, the panel concluded that such statements also supported
the inference that, but for the LA/MSF, Airbus would not have been able
to launch any of its existing range of LCA.385
The panel found that this ‘product’ effect of LA/MSF to launch a LCA
was complemented by other subsidies it had found to exist (e.g., equity
infusions, infrastructure measures, R&TD subsidies). Insofar as the
panel had previously found that LA/MSF by itself was a ‘substantial
cause’ of displacement or lost sales, in the Appellate Body’s view it was
not required to analyse (i) whether each of these other forms, taken
individually, were themselves a substantial cause, or (ii) whether all
subsidies aggregated constituted such a cause. Insofar as there was a
‘genuine causal link’ between these subsidies and the identified market
phenomena (i.e., displacement and significant lost sales), their effect
could be considered to complement the effect of LA/MSF (cumulation
of subsidy effects).386 The Appellate Body agreed that equity infusions
and infrastructure measures had a genuine connection to the ability of
Airbus to launch an LCA and therefore complemented and supplemented
383
Ibid., para. 7.1880. 384 Ibid., paras. 7.1911, 1.1912 385 Ibid., para. 7.1920.
386

the effect of LA/MSF. In contrast, R&TD subsidies did not perform such a
complementary role because it was not demonstrated that they provided a
competitive advantage to Airbus by improving the technologies incorpo-
rated in LCA models or their production process.387
As a result, the panel concluded that Airbus would not have launched
the aircrafts ‘as originally designed and at the time that it did’ but for the
challenged subsidies.388 However, the panel understood that this did not
necessarily show that no displacement or lost sales would have occurred
to Boeing but for the subsidies to Airbus. In this counterfactual situation,
another competitor could have entered the market and developed the
LCA to compete with Boeing, causing displacement and lost sales to the
US producers. After analysing the characteristics of the LCA market,
the panel distinguished two plausible scenarios of what would have
happened if no subsidies had been given to Airbus.389 In the first
scenario, neither Airbus nor any other competitor would have entered
the LCA market and Boeing would have been in a monopoly position,
holding 100 per cent of the market. In this scenario, the link between the
subsidies, enabling Airbus to enter the market, and the displacement and
lost sales is ‘self-evident’.390 In a second plausible scenario, Airbus would
not have entered the market but another competitor would be present,
which would most likely have been McDonnell Douglas (which later
merged with Boeing). Because this was also a US manufacturer, the
nexus between the subsidies enabling a non-US manufacturer, Airbus,
to enter the market and serious prejudice to the United States’ interest
would still be present. As a result, displacement and lost sales to the
United States would not have occurred but for the subsidies.
4.2.4.4 US – Large Civil Aircraft

The EU was successful in its counterclaim that the United States caused
serious prejudice by offering a variety of subsidies to Boeing. The
Appellate Body agreed with the panel’s unitary approach to causation,
387
Ibid., para. 1407. 388 Panel Report, EC – Large Civil Aircraft, para. 7.1976.
389
In two ‘unlikely’ scenarios, Airbus would have entered the market without the subsi-
dies, either in competition with Boeing (scenario 3) or in competition with two US
companies (scenario 4). The panel had previously concluded that Airbus would have
been a significant different LCA manufacturer in those scenarios and that serious
prejudice would have resulted from the subsidization. Panel Report, EC – Large Civil
Aircraft, para. 7.1984.
390

under which the market phenomena were assessed not in isolation but as
part of an integrated causation analysis.391 A counterfactual (‘but for’)
analysis was adopted in two steps: (i) the effect of the subsidies on
Boeing’s commercial behaviour (i.e., products and prices) was assessed;
and (ii) the effect of the subsidies, through their effect on Boeing’s
commercial behaviour, on Airbus’ prices and sales (i.e., price and volume
effects) was assessed.392 Hence the analysis was conducted the other way
round from that in EC – Large Civil Aircraft.
The EU had structured its claims on the basis of the different causal
mechanism through which the challenged subsidies adversely affected
Airbus. R&D subsidies were claimed to have ‘product effects’ (or ‘tech-
nology’ effects) (i.e., they affected the development of the 787) on
Boeing, whereas tied tax subsidies (which were tied either to exports or
to production) were claimed to have ‘price effects’ on Boeing. The EU
claimed that each group of subsidies caused different kinds of adverse
price and volume effects (step (ii)). Further, the EU requested the panel
collectively to assess the effect of the R&D subsidies and (some of the)
tied tax subsidies, as well as to assess collectively the effect of the tied tax
subsidies and other remaining subsidies. We briefly analyse the panel’s
and Appellate Body’s findings on these claims.
4.2.4.4.1 Serious prejudice resulting from product effects caused by

R&D subsidies The Appellate Body agreed with the panel’s two-step
counterfactual analysis of the effect of R&D subsidies. As a first step, the
panel employed a counterfactual analysis to determine whether a genu-
ine and substantial link existed between the R&D subsidies and the
‘product effect’ (i.e., the development of the 787) (step (i).393 The panel
concluded that, but for the R&D subsidies, Boeing would not have
developed the 787 by 2004. To reach this conclusion, the panel attached
importance to the nature of the subsidies at issue – that is, R&D
subsidies.394 In response to the US argument that the subsidy amount
was small compared to Boeing’s own R&D spending, the panel
responded that R&D subsidies could not simply be reduced to their
cash value because they are intended to have a multiplier effect, and
391
392
Panel Report, US – Large Civil Aircraft, para. 7.1659.
393
Ibid., para. 7.1775; Appellate Body Report, US – Large Civil Aircraft, para. 1030.
394
The panel considered four elements: (i) the objectives of the R&D subsidies; (ii) the
structure and design of the R&D subsidies; (iii) the operation of the R&D subsidies; and
(iv) the conditions of competition in the LCA industry.

they allowed Boeing to overcome significant disincentives in investing in

long-term, high-risk aeronautics R&D.395 The Appellate Body agreed
that subsidies of a relatively small magnitude could have substantial
effects.396
As a second step (step ii), the panel analysed whether, but for the effect
of the R&D subsidies (i.e., development of the 787 by 2004), Airbus
would not have suffered negative volume and/or price effects.397,398 The
Appellate Body agreed with the panel that Airbus had incurred signifi-
cant lost sales in a number of sales campaigns as well as significant price
suppression.399 The Appellate Body, however, disagreed that a threat of
displacement or impedance was present in a number of third-country
markets, because the panel did not identify clear trends showing such
effects.400
4.2.4.4.2 Serious prejudice resulting from price effects caused by tied

tax subsidies The EU had successfully argued before the panel that two
types of tied subsidy (i.e., FSC/ETI subsidies and B&O tax rate reduc-
tions) resulted in ‘price effects’ on Boeing’s sales, which caused adverse
volume and price effects on Airbus. The Appellate Body reversed the
panel’s analysis for various reasons, but decided that it could complete
this analysis and, in the end, upheld most of its findings. The Appellate
Body’s analysis was grounded in four aspects.
First, the nature of the subsidies was considered relevant. Both types of
subsidy were tied to production, and therefore could have a significant
effect on prices and output; and the FSC/ETI subsidies were particularly
likely to produce these effects, given their duration and nature as export
subsidies.
395
The panel rejected the argument that research funded by the NASA R&D subsidies was
publicly disseminated, and would therefore not confer a competitive advantage to
Boeing vis-à-vis Airbus.
396
Appellate Body Report, US – Large Civil Aircraft, paras. 1006, 1035.
397
See also ibid., para. 1025.
398
The Appellate Body agreed that the panel did not have to pursue a counterfactual (i.e.,
market without the subsidy), involving the development of a 767-plus aircraft by
Boeing. Ibid., para. 1040.
399
The price suppression argument was not based on the effect of R&D subsidies alone, but
rather of all subsidies. The EU adduced evidence relating to the nature and magnitude
of these subsidies, the estimated price effect of the subsidies (i.e., showing counter-
factual prices that would have prevailed absent the subsidy), the conditions of com-
petition in the relevant market, and results of some specific sales campaigns.
400

Second, the absolute and relative magnitude of these subsidies was

also relevant.401 They were considered significant in absolute terms, and
there was also some support for evidence of their relative importance, as,
when deployed strategically, they were of a sufficiently significant mag-
nitude to influence Boeing’s pricing and sales.
Third, the particular conditions of competition in this duopolistic market
were relevant. In such a market, both players possess market power, and
equally, through their supply and pricing decisions, influence the pricing of
others. Hence ‘the effects of one firm’s commercial behaviour – including
any price effects resulting from subsidies – will necessarily impact the rival
firm’.402 Such price competition between Boeing and Airbus was considered
more intense in some campaigns than in others.
Based on these three elements, the Appellate Body concluded that
Boeing had both the ‘ability and incentive’ to use the tied subsidies to
lower prices, and that there was a ‘substantial likelihood’ that this
occurred in those sales campaigns where price competition was
intense.403 Therefore the Appellate Body accepted that the subsidies
‘contributed in a genuine and substantial way’ to lowering Boeing’s
prices, but only for those sales campaigns for which (i) it was established
that Boeing was under particular pressure to reduce its price in order to
secure the sales; and (ii) where no other non-price factors explained
Boeing’s success in obtaining the sale or suppressing Airbus’ price.404
Hence, as a fourth and final element, evidence on this pricing dynamic
had to be present and a non-attribution analysis had to be conducted in
order to show the requisite causal connection.405 Regarding two sales
campaigns where this dynamic was present, the Appellate Body con-
cluded that Airbus incurred significant lost sales as a result of subsidi-
zation (volume effect), because (i) the tied subsidies (sufficiently)
explained Boeing’s price reductions; and (ii) the price reductions
(instead of non-price factors) explained why Boeing sales campaigns
were more successful.406
401
Ibid., paras. 1193, 1254. 402 Ibid., para. 1257. 403 Ibid., para. 1260.
404
Ibid., para. 1260.
405
The Appellate Body did not exclude that serious prejudice could have been established
on other evidence than circumstances of specific sales campaigns. Ibid., n. 2550.
406
The Appellate Body could not complete that panel’s analysis on whether the subsidies
also caused price suppression.

4.2.4.4.3 Collective assessment of subsidies and their effects The

Appellate Body analysed the need for a collective assessment of two
different sets of subsidies.
First, it criticized the panel for failing to assess whether the effects of
the R&D subsidies and tied tax subsidies (in particular, the B&O sub-
sidies) could be cumulated.407 The fact that both operated through
distinct causal mechanisms (through product and price effect, respec-
tively) did not a priori preclude cumulation. The panel should have
analysed whether the B&O subsidies complemented and supplemented
the effects of the R&D subsidies so that both could be cumulated.
Second, it analysed whether tied tax subsidies and the remaining sub-
sidies could be collectively assessed. These subsidies could not be aggregated
because of their different nature and effect: whereas tied tax subsidies
directly reduced Boeing’s marginal costs, the remaining subsidies only
increased Boeing’s non-operating cash flow. However, this did not a priori
preclude an analysis of whether the effect of the remaining subsidies could
be cumulated with the effect of the tied tax subsidies. The Appellate Body
agreed with the parties that, depending on the facts, subsidies not contin-
gent on sales or production could still affect the recipient’s behaviour in a
manner that causes serious prejudice.408 Hence they could have a ‘genuine’
causal link to Boeing’s pricing behaviour in the two sales campaigns where
the tied tax subsidies caused serious prejudice. The Appellate Body only
accepted cumulating the effect of those subsidies for which it was demon-
strated that they were somehow linked to the production of the relevant
aircraft in these two sales campaigns. For instance, the effect of subsidies
received for relocating Boeing’s headquarters was not cumulated to the
effect of the tied tax subsidies. The argument that such a general subsidy
could reasonably be deemed to benefit all the company’s products was
considered insufficient to show the required link.409 On the other hand,
IRBs received by the city of Wichita were specially aimed at, and were used
for the purpose of, enhancing Boeing’s manufacturing facilities. The fact
that relevant parts of the relevant aircraft were produced in those facilities,
and the fierce price competition in those sales campaigns, suggested that the
IRBs were used by Boeing to reduce its price to the extent necessary to
secure these sales. The IRBs could also not be considered non-trivial in
terms of their overall magnitude and duration. For these reasons, their effect
407
Appellate Body Report, US – Large Civil Aircraft, paras. 1309–1321.
408
Ibid., paras. 1336, 1348. 409 Ibid., para. 1342.

complemented and supplemented the price effect of the tied tax subsidies,
thereby causing serious prejudice in the form of lost sales.410
4.3 Non-actionable subsidies

Under the original SCM Agreement, an actionable subsidy claim could not
be undertaken against a limited group of subsidies. This group of green light
subsidies also could not be countervailed. During the Uruguay Round the
United States had taken the most defensive position, as their purported
legitimate objective could easily be circumvented by firms and govern-
ments.411 Because of ‘the fungible nature of money’, firms receiving, for
example, R&D subsidies could simply reallocate resources normally used
for R&D to another purpose.412 Moreover, ‘identifying non-actionable
subsidies is inherently a dangerous task’ because governments could turn
actionable subsidies into green light subsidies.413 This negative stance
altered with regard to R&D subsidies when the Clinton administration
came to power (1993).414 Other negotiating countries were generally
more open to the inclusion of green light subsidies, because they wanted
to place such subsidies outside the scope of CVD action. Ultimately, three
subsidy categories were, under strict conditions, considered non-actionable
but their green light status was extinguished at the end of 1999.415,416
410
Ibid., para.1348.
411
According to Anderson and Husisian, US negotiators viewed green light subsidies as
‘necessary evils’ which they had to accept in exchange for stricter disciplines on other
subsidies. Anderson and Husisian, above Chapter 2 n. 80, at 316.
412
MTN.GNG/NG10/W/29, 22 November 1989.
413
Despite these reservations, the US had proposed a list of subsidies that could be
considered non-actionable (MTN.GNG/NG10/W/29, 22 November 1989).
414
The US even successfully pushed for widening the scope of R&D subsidies as foreseen in the
Dunkel Draft. Strategic trade theory (see above Chapter 1) provided new arguments for
such subsidies. See Kleinfeld and Kaye, above Chapter 2 n. 79, at 51–52; J. Odell and
B. Eichengreen, ‘The United States, the ITO, and the WTO: Exit Options, Agent Slack, and
Presidential Leadership’, in A. O. Krueger (ed.), The WTO as an International Organization
(University of Chicago Press, 1998), 181–209, at 203; L. W. Nowicky, ‘Alternative
Approaches to International Competitiveness: Does the Clinton Administration Need
French Lessons?’, in M. E. Kreinnin (ed.), Contemporary Issues in Commercial Policy
(Oxford: Pergamon, 1995), 191–202, at 193; H. A. Hazard, ‘Microeconomic Initiatives to
Promote Technology and Industry in the United States’, ibid., at 185–90.
415
Previous drafts had listed more green light subsidies, but the EU finally agreed to the US
demand to reduce the list in exchange for less demanding language under Article 6.1.
See Stewart, above Chapter 2 n. 49, at 911.
416
These were mainly created on the demand of the EU (focusing on R&D, environment,
and regional subsidies), Canada (focusing on regional subsidies), and Mexico (which
successfully pushed for environmental subsidies in the last days of negotiations).

First, certain subsidies for R&D conducted by firms or by higher

education or research establishments on a contract basis with firms
were non-actionable, although this was limited to a certain level and to
certain types.417 Second, assistance to disadvantaged regions was like-
wise non-actionable if the geographical region was clearly defined
and the determination was based on neutral and objective criteria.418
The third category covered a limited type of environmental subsidy,
namely assistance to promote adaptation of existing facilities to new
environmental requirements.419 In principle, covered subsidies
could not be challenged or countervailed, regardless of their trade
effects.420
The SCM Agreement provided for a notification requirement for such
non-actionable subsidies, but no single formal notification was made
over the entire period that this category was in place.421 In fact, members
had to notify the SCM Committee of all subsidy programmes they
wanted to classify as non-actionable in advance of implementation.422
When another member disputed the non-actionable nature of a subsidy,
it could start a review procedure that could end in binding arbitration.423
If a subsidy was not notified, in principle it could not benefit from non-
actionability and became thus actionable and countervailable if it caused
adverse effects or injury respectively.424 However – and this was the weak
spot of the notification procedure – such subsidies were still non-
actionable if it was found during a countervailing or multilateral pro-
ceeding that they conformed to the standards of one of the three
categories.425
The category of green light subsidies was thus extinguished at the end
of 1999, because there was no consensus in favour of continuing its
application. As mentioned, the discussion was linked to the extension of
Article 6.1. Several developing countries, such as Brazil and India, were
Stewart, above Chapter 2 n. 49, at 907; Anderson and Husisian, above Chapter 2 n. 80, at
319. Regarding the shift in US stance on R&D subsidies, see above n. 414.
417
Article 8.2(a) of the SCM Agreement. Excluded from the SCM Agreement were sub-
sidies for ‘fundamental research’ independently conducted by higher education or
research establishments (footnote 26 of the SCM Agreement).
418
Article 8.2(b) of the SCM Agreement. 419 Article 8.2(c) of the SCM Agreement.
420
See below (Chapter 5, section 5.1.3.3) on the multilateral remedy to respond to non-
actionable subsidies.
421
G/SCM/W/546/Rev.2, 26 April 2011, para. 17.
422
Article 8.3 of the SCM Agreement. 423 Articles 8.3–8.5 of the SCM Agreement.
424
Footnote 35 of the SCM Agreement. 425 Ibid.

also not in favour of its extension in the existing form because the
categories overly reflected the interest of developed countries. The EU
and Canada, in contrast, favoured a continuation of the category of non-
actionable subsidies, whereas the United States articulated mixed views.
Hence there seemed not even a consensus present among developed or
developing countries.426 Therefore it seems unlikely that members will
reinstall this category in its present form. Part III of this study will
explore whether its re-inclusion seems warranted.427
4.4 The applicability of the GATT with regard to subsidies

Before assessing the remedies with regard to subsidization, two rele-
vant – if somewhat opposite – questions are explored on the applic-
ability of the GATT with regard to subsidies. First, whether Article III of
the GATT stipulates a relevant obligation with regard to subsidies
beyond the disciplines of the SCM Agreement is examined. Second,
whether Article XX of the GATT could be invoked as an exception to
justify violations of the SCM Agreement disciplines set out above is
examined.
4.4.1 The national treatment obligation stipulated

in Article III of the GATT
The national treatment obligation outlaws internal fiscal and non-fiscal
measures that discriminate between imported and domestic like prod-
ucts. This discipline, elaborated in Article III of the GATT, is applicable
to subsidies, unless they are covered under the carve-out provided for in
paragraph 8(b) of Article III, which provides in the relevant parts that
the ‘[t]he provisions of this Article shall not prevent the payment of
subsidies exclusively to domestic producers. . .’. According to the case
law,428 this carve-out has a limited scope and only covers (i) payments
426
G/SCM/M/24, 26 April 2000; G/SCM/M/22, 17 February 2000.
427
See below Chapter 16, section 16.1.1.
428
The case law seems to be divided on whether this provision is an exception to Articles
III:2 and III:4 of the GATT, or simply confirms that certain subsidies to producers (and
not to products) fall outside the scope of paragraphs 2 and 4 of Article III. The latter
position was adhered to by the panel in Indonesia – Autos (para. 14.43 and n. 709) and
the GATT panel in US – Malt Beverages (para. 5.8). However, the majority of the case
law, including the Appellate Body report in Canada – Periodicals (p. 34), seems to
qualify Article III:8(b) as an exception: see Panel Report, Canada – Milk, para. 7.97;
Panel Report, EC – Bananas III, n. 448; GATT panel report, Italian Tractors, para. 13.

(i.e., the expenditure of revenue by a government429,430) of subsidies, (ii)

exclusively to domestic producers (and thus not to purchasers431 or
processors432), and (iii) that do not discriminate ‘between imported
and domestic products . . . when using such subsidies’.433
To grasp the scope of Article III, it is useful to distinguish between
potential discrimination in the market of the subsidized product and the
input market.
First, regarding distortions related to the market of the subsidized
product, Article III:8(b) only carves out direct subsidies to producers
(second condition), and thus not so-called product subsidies – that is,
subsidies given to purchasers or final consumers of these products (e.g.,
tax reduction when buying domestic products).434 The GATT panel in
US – Tobacco recognized that ‘a remission of a tax on a product and the
payment of a producer subsidy out of the proceeds of such a tax could
have the same economic effects [but] noted that the distinction in Article
III:8(b) is a formal one, not one related to the economic impact of a
measure’.435 In other words, a remission of a tax to purchasers or
consumers (i.e., product subsidy) is disciplined by Article III, whereas
a subsidy to producers is exempted by virtue of Article III:8(b). The
panel in Indonesia – Autos referred to the ‘structural relationship
between the rules on national treatment on products [Article III] and
the rules on subsidies to producers [Article XVI of the GATT; SCM
Agreement]’.436 In its view, ‘Article III prohibits discrimination between
domestic and imported products while the SCM Agreement regulates the
429
Appellate Body Report, Canada – Periodicals, p. 34. This means that subsidization by
negative action (i.e., forgoing revenue) is not covered by the carve-out.
430
In US – Tobacco, the GATT panel found that the tax remission to domestic producers
qualified as a payment in the meaning of Article III:8(b) of the GATT. In United States –
Malt Beverages, the panel found that a reduction of taxes on a good did not qualify as a
‘payment of subsidies’ for the purposes of Article III:8(b) of the GATT.
431
GATT panel report, Italian Agricultural Machinery, para. 14.
432
GATT panel report, EEC – Oilseeds.
433
Panel Report, Indonesia – Autos, para. 14.30.
434
Support has been found in the negotiating history to exclude consumption subsidies as
the drafters of Article III explicitly rejected a proposal by Cuba at the Havana
Conference to amend the Article to read: ‘the provisions of . . . Article [III] shall not
preclude the exemption of domestic products from internal taxes as a means of indirect
subsidization in the cases covered under Article [XVI]’ (E/CONF.2/C.3/6, 8 December
1947, at 17).
435
GATT panel report, US – Tobacco, para. 109.
436
Panel Report, Indonesia – Autos, para. 14.33 (emphasis added).

provision of subsidies to enterprises’.437 Hence subsidies given to pur-

chasers or final consumers that favour domestic over imported products
could be challenged under Article III:4 of the GATT. In contrast to
actionable subsidy claims under the SCM Agreement, there is no need
to demonstrate effectively that such subsidies cause adverse trade
effects,438 and Article XX of the GATT could still be available to justify
a violation.
Second, regarding distortions in the input market, subsidies to down-
stream domestic producers are inconsistent with Article III:4 of the
GATT where they discriminate between imported and domestic inputs
(fail the third condition).439 Hence a distortion in the input market
resulting from an advantage (e.g., a subsidy) given to domestic producers
when they use domestic inputs instead of imported inputs is inconsistent
with Article III:4 of the GATT. Similarly to Article 3 of the SCM
Agreement, the distortion challenged by the complainant relates to the
input market and not to the market of the subsidized producers. As
explained above, a complainant might opt to challenge local content
subsidies under the explicit prohibition of the SCM Agreement, rather
than under Article III.4 of the GATT.440 Still, an (additional) claim
under the GATT could certainly be valuable because it does not require
formal (and demanding) demonstration of (i) a subsidy (as defined in
Article 1 of the SCM Agreement) (ii) contingent upon (iii) local use.
Regarding the third element, a measure could be challenged under
Article III even if it does not affect the internal use, but rather the
‘internal sale [or] purchase’.441 The US – FSC (Article 21.5 – EC) case
has shed light on the two other elements. The Appellate Body concluded
that eligibility requirements (for a tax break) mandating that a certain
percentage of the value of a good stem from domestic inputs or domestic
labour violated Article III.4 of the GATT.442 Such eligibility require-
ments ‘affect’ the internal use of imported products, because they
437
Ibid., para. 14.33; GATT panel report, US – Malt Beverages, para. 5.10.
438
The analysis of whether imported goods are treated less favourably is designed to
ascertain whether the measure at issue ‘modifies the conditions of competition in the
relevant market to the detriment of imported products’. Appellate Body Report,
Korea – Various Measures on Beef, para. 137; see also Appellate Body Report,
Thailand – Cigarettes, para. 128.
439
Article III:5 of the GATT could also be relevant, but has not been used much in
practice.
440
See above Chapter 4, section 4.1.2. 441 Article III:4 of the GATT.
442
Appellate Body Report, US – FSC (Article 21.5 – EC), para. 222.

influence the manufacturer’s choice between imported and domestic

input products if it wished to obtain the financial advantage in question.443
This illustrates, first of all, that it is not necessary to demonstrate that this
advantage amounts to a ‘benefit’ within the meaning of Article 1.1(b) of the
SCM Agreement,444 which found confirmation in the Canada – Renewable
Energy / Feed-in Tariff Program case.445 Further, the test articulated by the
Appellate Body might be less demanding than the contingency test under
the SCM Agreement (e.g., the ‘geared to induce’ test in case of de facto
contingency). Where the use of domestic goods counts towards the local
content requirement to receive the advantage, whereas the use of imported
goods does not, this amounts to ‘less favourable treatment’ within the
meaning of Article III:4.446 In sum, a claim under Article III:4 of the
GATT might be articulated in order to (partly) circumvent the difficult
demonstration of a local content subsidy within the meaning of Articles 1
and 3 of the SCM Agreement.
4.4.2 The general exceptions stipulated in Article XX

of the GATT
An unresolved question is whether Article XX of the GATT could be
invoked to justify subsidies inconsistent with the SCM Agreement.447 Even
though no panel or the Appellate Body has explicitly expressed its view, the
majority of authors considering such a defence unavailable seems to be
correct.448 Howse, on the other hand, suggests that Article XX is applicable
443
Ibid., para. 212.
444
The Appellate Body’s reference to ‘a real and substantive advantage’ (ibid., para. 218)
was a factual determination and was not advanced as the legal standard for demon-
strating ‘less favourable treatment’ under Article III:4. In China – Auto Parts, the
Appellate Body referred to the ‘incentive’ (para. 195) created by the challenged measure
to limit the use of imported auto parts over domestic auto parts.
445
See above Chapter 4, section 4.1.2. The Illustrative List of the TRIMs Agreement lists
examples of TRIMs inconsistent with Article III:4 of the GATT and the Appellate Body
concluded that ‘advantage’ within the meaning of the Illustrative List has a different
meaning from a ‘benefit’ within the meaning of the SCM Agreement. Appellate Body
Report, Canada – Renewable Energy / Feed-in Tariff Program, paras. 5.207–5.209.
446
Appellate Body Report, US – FSC (Article 21.5 – EC), paras. 214–222.
447
For the normative discussion see below Chapter 16, section 16.1.1.2.2.2.
448
See B. J. Condon, ‘Climate Change and Unresolved Issues in WTO Law’, 12:4 Journal of
International Economic Law (2009), 895–926; A. Green, ‘Trade Rules and Climate Change
Subsidies’, 5:3 World Trade Review (2006), 377–414, at 407–10; G. N. Horlick, ‘The WTO
and Climate Change “Incentives”’, in T. Cottier, O. Nartova, and S. Z. Bigdeli (eds.),
International Trade Regulation and the Mitigation of Climate Change (Cambridge
University Press, 2009), 193–6, at 194; C. Tran, ‘Using GATT, Art. XX to justify climate

in as much as the SCM Agreement is a lex specialis, or specialized legal

regime, derived from the GATT itself.449 Yet this qualification would of itself
offer no legal basis to transpose the general defence of Article XX to measures
inconsistent with the SCM Agreement. Somewhat to the contrary, the fact
that the SCM Agreement clearly goes ‘well beyond merely applying and
interpreting’450 the GATT disciplines would rather suggest that the general
exceptions under Article XX are not applicable unless the specific applicable
regime (i.e., the SCM Agreement) points to the opposite conclusion.451
The Appellate Body in China – Raw Materials has indeed confirmed
that Article XX of the GATT could not be invoked outside the GATT
framework unless textual support could be found within the specific legal
regime.452 The lack of such a textual hook in the SCM Agreement,
therefore, suggests that no recourse could be taken to Article XX in the
case where a member violates Article 3 (prohibited subsidies) and/or
Article 5 (actionable subsidies) of the SCM Agreement. Indeed, no
language in the text of Article 3 and Article 5 suggests that Article XX
of the GATT could be invoked to justify a violation of these provisions.
Rather, the fact that both provisions stipulate an explicit exception for
some agricultural subsidies supports the view that recourse to Article XX
is, a contrario, not available.453 The (temporary) exception for certain
change measures in claims under the WTO Agreements’, 27 Environmental and Planning
Law Journal (2010), 346–59.
449
An interpretative understanding would be sufficient to clarify that Article XX is indeed
applicable, implying that a treaty amendment would not be needed. R. Howse, ‘Climate
Mitigation Subsidies and the WTO Legal Framework: A Policy Analysis’, International
Institute for Sustainable Development (May 2010), at 18 and 25.
450
Appellate Body Report, Brazil – Desiccated Coconut, at p. 17.
451
The lex specialis maxim of interpretation precisely means that the more specific regime
supersedes the lex generalis. Legal disciplines elaborated under a lex specialis could very
well go beyond or deviate from the general legal regime. Annex 1A of the Marrakesh
Agreement Establishing the World Trade Organization confirms that the SCM
Agreement takes precedence over the GATT in case of a conflict.
452
The Appellate Body confirmed that it had allowed China (in China – Publications and
Audiovisual Products) to have recourse to Article XIV of the GATS to justify a violation of
Article 5.1 of its Accession Protocol because of the explicit textual support in this provision
(‘[w]ithout prejudice to China’s right to regulate trade in a manner consistent with the
WTO Agreement’). Because Paragraph 11.3 of China’s Accession Protocol lacks such a
textual hook and stipulates an explicit commitment, no recourse was available to Article XX
of the GATT to justify a violation of this commitment. Appellate Body Report, China – Raw
Materials, paras. 303–306. The panel in this case was even more explicit that it would reject
recourse to Article XX for any violation of the covered agreements that does not explicitly
incorporate Article XX (see below n. 455).
453
See below Chapter 6, section 6.2.3. Footnote 5 also offers a limited exception regarding
the prohibition of export subsidies (see below, Chapter 4, section 4.1.1.3).

green light subsidies offers further contextual support to the view that
members intended to construct the SCM Agreement as a specific regime
with its own exceptions and flexibilities.454 The specific taxonomy into
prohibited/actionable/non-actionable types of subsidy indeed supports
the interpretation that recourse to Article XX of the GATT was not
intended. Further, a WTO panel will attach significance to the fact that
members have explicitly incorporated Article XX into some covered
agreements (in particular, the TRIMs Agreement), but not into the
SCM Agreement.455 Next, the availability of Article XX would have the
odd legal consequence that multilateral action against certain subsidies
would be precluded (e.g., against those subsidies justified under Article
XX of the GATT), whereas unilateral CVD action against these subsidies
would still be possible.456,457,458 In addition, given the particular excep-
tions and flexibilities inscribed in the SCM Agreement and the double
track to respond to foreign subsidization, some discussions in the nego-
tiation history of the applicability of Article XX would have been
expected in case members intended to make this exception available.
The absence of any discussion in the Uruguay Round provides
454
See also Panel Report, China – Raw Materials, para. 7.153.
455
Appellate Body Report, China – Raw Materials, para. 303. According to the panel in
China – Raw Materials (para. 7.153), the reference in Article XX to ‘this Agreement’ a
priori suggests that these exceptions only relate to the GATT and are, therefore, only
available outside the GATT if they are incorporated into the specific legal regime at
stake (e.g., the TRIMs Agreement). In that case, the legal basis for this justification is
not Article XX but the provision in the specific legal regime incorporating this provi-
sion. Hence this reasoning clearly implies that Article XX is not available in the context
of the SCM Agreement because no provision of this agreement incorporates Article XX.
456
Nothing in either the GATT or the SCM Agreement suggests that CVD action could not
be undertaken against subsidies that are justified under Article XX of the GATT.
457
The relevance of this argument should be tempered because this situation was already
present under the GATT as it allowed for CVD action against foreign subsidization
(Articles II:2(b) and VI of the GATT) and, at the same time, imposed disciplines on
subsidies (Article XVI of the GATT) that could potentially be justified under Article XX
of the GATT. This option was, however, largely theoretical in nature because those
subsidies to producers that were effectively disciplined (e.g., export subsidies) could
arguably not meet the criteria of the necessity test in Article XX of the GATT.
458
Another systemic challenge considered by Bigdeli is the application of the chapeau test
of Article XX of the GATT (e.g., non-discriminatory application of a measure), because
this suits the context of border measures rather than that of domestic regulations and
subsidies. See S. Z. Bigdeli, ‘Resurrecting the Dead? The Expired Non-Actionable
Subsidies and the Lingering Question of “Green Space”’, 8:2 Manchester Journal of
International Economic Law (2011), 2–37, at 33. More precisely, this challenge seems to
result from the application of the chapeau test to the trade expanding effect of subsidies
abroad (trade is expanded, rather than restricted).

supplementary support to exclude recourse to Article XX of the GATT to

justify violations of the SCM Agreement. It finally bears noting that the
Appellate Body in Canada – Renewable Energy / Feed-in Tariff Program
felt the need to inject policy space for certain support measures at the
level of the subsidy definition (i.e., benefit analysis), which also seems to
suggest that it does not consider that an Article XX GATT defence is
available to justify subsidies violating the SCM Agreement.

16.3 PP 115 195 Disciplines On Subsidies

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

16.3 PP 115 195 Disciplines On Subsidies

Uploaded by

Copyright:

Available Formats

4

Articles 1 and 2 merely deﬁne the concepts of subsidy and speciﬁcity,

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

actionable subsidies (amber light), meaning that they could be challenged if

4.1 Prohibited subsidies

4.1.1 Export subsidies

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

The opening part of Article 3.1 installs an exception on agricultural export

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

Agreement regarding prohibited subsidies in Part II . . . ’.17 This suggests

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

4.1.1.1 Export subsidies in the meaning of Article 1

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

exportation or export earnings.29 The mere fact that a subsidy is granted

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

the subsidy geared to induce the promotion of future export perform-

If relevant evidence is present, it could involve a comparison between

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

assessment must be based on the information available to the granting

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

some previous case law. For instance, the panel in Australia –

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

4.1.1.2.1 The provision of goods or services favourable to

retention scheme could form an important incentive to export because, depending on

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

directly or indirectly through government mandated schemes; and

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

4.1.1.2.2 Border tax adjustments and duty drawback systems Five

4.1.1.2.2.1 Border tax adjustments

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

4.1.1.2.2.1.3 Indirect taxes Regarding indirect taxes, the conditions

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

Under a system of BTAs on exportation, no VAT is thus charged on

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

generating the risk that rebates on exports would be in excess of taxes

4.1.1.2.2.2 Duty drawback systems In addition to indirect taxes, the

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

charges should be rebated on exports. After all, import charges are in

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

paid on capital inputs.108 Observing only administrative difﬁculties in

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

4.1.1.2.3 The (potential) direct transfer of funds: export credit

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

or guarantee programmes against increases in the cost of exported

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

export credits), and therefore sets ‘a dispositive legal standard’ to deter-

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

analysing other arguments, it will be explained in Part II why the panels’

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

Under footnote 5 of the SCM Agreement, where the Illustrative List

These ‘other provisions’ referred to by the Appellate Body include the

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

It is noteworthy that the Draft Consolidated Chair Text would amend

4.1.2 Local content subsidies

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

subsidy is conceptually different from other kinds of subsidies. It focuses

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press

certain Canadian value added (CVA) requirements. The deﬁnition of CVA

https://doi.org/10.1017/CBO9781139046589.008 Published online by Cambridge University Press