3.

The GATT
The original GATT contract underwent a substantial transformation during the Uruguay
round: WTO Members agreed to add to the original text of the GATT a series of
Understandings adopted during the Uruguay Round, as well as the so-called GATT
acquis: the decisions adopted by the GATT CONTRACTING PARTIES since 1947,
including protocols of tariff concessions, accessions and waivers still in force. In what
follows we discuss the meaning and substance of the main legal disciplines of the GATT
— the key Articles are summarized in Box 3.1. Space constraints preclude any in-depth
discussion of the underlying economic or policy issues. We will return to some of these in
Chapter 7, but in order to understand the substance of the GATT a more “legal” treatment
is required.”
Box 3.1 Summary of the main GATT Articles
I MFN requirement.
II Tariff schedules (bindings).
III National treatment.
V Freedom of transit of goods
VI Allows antidumping and countervailing duties. Superseded by the GATT 1994
Agreement on Antidumping, and the Agreement on Subsidies and Countervailing
Measures (SCM).
VII Valuation of goods for customs purposes to be based on actual value. Superseded by
the GATT 1994 Agreement on the Implementation of Article VII
VIII Fees connected with import and export formalities to be cost-based.
XII Obligation to publish trade laws and regulations.
Prohibition on quantitative restrictions
XII Permits trade restrictions to safeguard the balance of Payments
XIII Non-discriminatory administration of quantitative restrictions
XVI Prohibits export subsidies. Complemented by the WTO
Agreement on Subsidies and Countervailing Measures
XVII State trading enterprises to abide by MFN
XVIII Allows developing countries to restrict trade to promote infant industries and to
protect the balance-of-payments (weaker conditionality than Article XII) XIX Allows for
emergency action to restrict imports of particular products if these cause serious injury to
the domestic industry. Complemented by the WTO Agreement on Safeguards
XX General exceptions provision—allows trade restrictions if necessary to attain non-
economic objectives (health, safety).
XXI National security exception.
XXII Requires consultations between parties involved in trade disputes.
XXIII GATT’s main dispute settlement provision, providing for violation and non-
violation complaints. Complemented by the WTO Understanding on Rules and
Procedures Governing the Settlement of Disputes.
XXIV Lays out criteria for the formation of free trade areas and customs unions.
XXVIII Allows for renegotiation of tariff concessions.
XXVIII bis Calls for periodic trade rounds to lower trade barriers.
XXXIII Allows for accession of new members.
P IV Calls for more favorable and differential treatment of developing countries.
Border measures*
There are two types of border measures: quantitative restrictions (QRs) and tariffs. WTO
membership brings with it the obligation to abolish QRs (Art. XI GATT) and to
participate in tariff negotiations (Art. XXVIII bis) aimed at establishing tariff ceilings for
products. Such negotiated ceilings, generally called tariff “bindings,” establish maximum
levels that applied tariffs may not exceed.
The cornerstone of the GATT is the principle of non-discrimination. It takes the form of a
legal obligation that binds the actions of WTO
Members with respect to both border (trade) policies and internal regulatory or tax
policies, 1.e., measures destined to apply to both domestic products and imports alike. A
comprehensive reading of Arts. I, II, IH, XI and XVI GATT and the associated case law
implies that since (import and export) quantitative restrictions are prohibited (Art. XI
GATT), and subsidies are regulated (Art. XVI GATT and the WTO Agreement on
Subsidies and Countervailing Measures SCM), the only permissible form of protection of
domestic products is through customs duties (Art. II GATT).
Customs duties, other border measures and non-discrimination (MFN)
Irrespective of whether duties have been bound or not, they must be applied on a non-
discriminatory basis (Art. I). Since customs duties concern only imported products, the
legal obligation amounts to a prohibition on discrimination between two imported
products. For tariff commitments to be meaningful, a common language to describe
goods is required: otherwise it is difficult to make any claim that country A discriminates
against cars originating in B by subjecting them to an import regime worse than that
afforded to cars originating in C, There is a need to have a common understanding as to
what a car, in this example, amounts to. The Harmonized Commodity Description and
Coding System (HS) supplies the common language. The HS is a classification system
for goods that has been elaborated in the World Customs Organization (WCO), an
international organization with headquarters in Brussels, Belgium. The function of the HS
is to describe goods in a multilaterally agreed manner. Product descriptions are expressed
in digits: the fewer the number of digits, the more generic the product category (for
example, at the 2-digit level one might find the term “motor vehicles”); the higher the
number of digits, the more specific the product category (for example, at the 8-digit level,
one might find something like “passenger cars weighing less than 2 tons and with an
engine not exceeding 1.5 liters, with a catalytic converter”).
The HS plays a key role in the operation of the WTO as it defines the scheduling of tariff
commitments (concessions) at the 6-digit level. While a number of WTO Members have
formally not acceded to the international treaty establishing the HS, de facto all WTO
Members follow the HS classification up to the 6-digit level. Beyond that, WTO
Members are free to “shape” their tariff schedule, and thus their WTO concessions (tariff
bindings) to their liking. Art. 3.3 of the HS Con- vention reads in this respect: Nothing In
this Article shall prevent a Contracting Party from establishing, in its Customs tariff or
statistical nomenclatures, subdivisions classifying goods beyond the level of the
Harmonized system, provided that any such subdivision is added and coded at a level
beyond that of the six-digit numerical code set out in the Annex to this Convention.
By adding digits (disaggregating product categories), WTO Members cannot circumvent
higher-level tariff bindings. Thus, a tariff at a more disaggregated level cannot be higher
than the bound rate at the 6-digit or more aggregated level. An example from the US
Tariff Schedule can help explain this point: Chapter 87 of the HS is entitled “Vehicles
Other Than Railway Rolling-Stock, and Parts and Accessories Thereof.” Heading 8708 is
entitled “Parts and Accessories of the Motor Vehicles of Headings 8701 to 8705” (the two
categories corresponding to tractors, motor vehicles for the transport of ten or more
persons and motor cars principally designed for the transport of persons). Heading
8708.10 reads “Bumpers and Parts Thereof.” Heading 8708.10.60 reads “Bumpers” (i.e.,
stampings). The United States bound their tariffs in Chapter 87 at the 8-digit level at 2.7
percent. This means that no tariff can be imposed on any 10- or 12-digit level that
exceeds the maximum duty of 2.7 percent.
Periodically, the HS Committee of the WCO undertakes a review of the HS to take
account of changes in technology and patterns in international trade and recommends
certain amendments to the HS. The first of such changes came into force on | January
1992 (referred to as HS92). Another, more substantial change entered into force on 1
January 1996 (HS96). The WTO Committee on Tariff Concessions established simplified
procedures to implement these changes and any future changes in the HS relating to
GATT concessions (GATT Doc. BISD 395/300). This was needed to reduce the
administrative costs associated with remapping all the concessions made in the past.*
Non-discrimination with respect to customs duties amounts to the obligation (Art. I
GATT) to treat goods of various origins (from any WTO Member) in an identical manner,
that is, the customs treatmen afforded by country A to, say, a car from country B should
be equal to that afforded to a car originating in country C. To respond to the question
whether A is obliged to treat the two cars from countries B and C mentioned in our
example above in an identical manner, the logic of the GATT requires that comparability
between car B and car C is first established. This is where the notion of “likeness” kicks
in. Likeness 1s far from being a self-interpreting term. GATT/WTO jurisprudence has
had recourse to various criteria to establish likeness, but tariff classification is the
dominant criterion. Thus, Members need to be sure that the tariff classification is
sufficiently detailed. As a rule of thumb, tariff classifications at the 6-digit level seem to
fit the bill in this respect, as WTO Members have the discretion to make tariff distinctions
across substitutable products, to the extent that they come under different tariff lines.
There is long line of case law in this respect.
The 1978 panel report on EEC — Animal Feed Proteins concluded that “animal, marine
and synthetic proteins were not products like those vegetable proteins covered by the
[contested] measures” (§ 4.20).7 Subsequently, a GATT panel on Spain — Un-roasted
Coffee set aside the relevance of process-based distinctions when it comes todefining
likeness. The case involved a complaint by Brazil regarding Spanish classification of un-
roasted coffee, which distinguished between Colombian mild, other mild, unwashed
Arabica, Robusta and other. The first two categories had duty-free treatment, while the
last three were subject to a seven percent import duty. The duty for roasted coffee was
unbound. Brazil argued this was inconsistent with Art. I GATT. The panel noted that:
all arguments that had been advanced during the proceedings for the justification
of a different tariff treatment for various groups and types of un-roasted coffee ... mainly
related to organoleptic differences resulting from geographical factors, cultivation
methods, the processing of the beans, and the genetic factor. The Panel did not consider
that such differences were sufficient reason to allow for a different treatment.
The Panel furthermore found relevant to its examination of the matter that un-
roasted coffee was mainly, if not exclusively, sold in the form of blends, combining
various types of coffee, and that coffee in its end-use, was universally regarded as a well-
defined and single product intended for drinking.
The Panel noted that no other contracting party applied its tariff regime in respect of un-
roasted, non-decaffeinated coffee in such a way that different types of coffee were subject
to different tariff rates.
In light of the foregoing, the Panel concluded that un-roasted, non-decaffeinated coffee
beans listed in the Spanish Customs Tariff ... should be considered as like products within
the meaning of Article I.1.
The panel report on Japan — SPF Dimension Lumber provides the most eloquent
acknowledgment of the relevance of tariff classification as the dominant criterion to
establish likeness:
if a claim of likeness was raised by a contracting party in relation to the tariff
treatment of its goods on importation by some other contracting party, such a claim
should be based on the classification of the latter, i.e., the importing country’s tariff.
The Panel noted in this respect that “dimension lumber” as defined by Canada was a
concept extraneous to the Japanese Tariff ... nor did it belong to any internationally
accepted customs classification. The Panel concluded therefore that reliance by Canada
on the concept of dimension lumber was not an appropriate basis for establishing
“likeness” of products under Article I.1 of the General Agreement.
Art. I GATT does not explicitly refer to either de jure or de facto discrimination. The
Appellate Body clarified this issue in its report on Canada — Autos where in § 78 it
notes:
In approaching this question, we observe first that the words of Article I.1 do not
restrict its scope only to cases in which the failure to accord an “advantage” to like
products of all other Members appears on the face of the measure, or can be demon-
strated on the basis of the words of the measure. Neither the words “de jure’ nor “de
facto” appear in Article II. Nevertheless, we observe that Article I.1 does not cover only
“in law,” or de jure, discrimination. As several GATT panel reports confirmed, Article I.1
covers also “in fact,” or de facto, discrimination. Like the Panel, we cannot accept
Canada’s argument that Article I.1 does not apply to measures which, on their face, are
“origin-neutral.”
To establish a violation of Art. I GATT, a complainant need not show actual trade effects.
It suffices that a WTO Member creates more favorable competitive opportunities for
some WTO Members only, for violation of Art. I GATT to be established.* Furthermore,
the wording of Art. I GATT makes it clear that a WTO Member cannot treat imports from
a non-WTO Member better than those of a WTO Member: in other words, imports from
WTO Members will, in principle, receive the best possible treatment, hence the term
most-favored-nation (MFN) which reflects the non-discrimination principle when applied
to customs duties and all other measures in connection with importation.
Of course, states may apply at the border measures other than customs duties. The
Understanding on Art. If GATT distinguishes between “ordinary customs duties”? and
“other duties and charges” and provides for the possibility that both categories of duties
be bound (since the extent of tariff bindings is not a legal obligation but a matter of
negotiation).'° Second, other border practices and measures might affect imports: the time
goods spend at the border, health inspections, etc. The MFN obligation covers any
advantage in relation with importation and exportation of goods. The term “advantage”
has been given a very wide coverage in GATT/WTO case law.
The MFN obligation moreover obliges WTO Members to extend any advantage (as
understood above) immediately and unconditionally to all WTO Members. Whereas the
first term seems to suggest the simple passage of time, the interpretation of the latter is an
issue. The panel report on Canada — Autos (§§ 10.22 and 10.24) concluded that
“unconditionally” means that a WTO Member cannot impose conditions beyond those
imposed on the original beneficiary. It does not imply that it cannot impose any
conditions at all. This report is particularly important for 8-digit classifications: in this
line of thinking, a country which conditions benefits on, say, exports of textile goods
upon satisfaction of specific labor standards, will not be violating MEN if it applies the
mentioned condition on an MEN basis. Note that the legal value of this report remains
uncertain because the AB vacated, albeit on other grounds, the panel report.
The panel report on EC — Tariff Preferences took a different route. It adopts a blanket
prohibition on attaching any conditions as the proper way to understand the term
“unconditionally”:
In the Panel’s view, moreover, the term “unconditionally” in Article I.I has a
broader meaning than simply that of not requiring compensation. While the Panel
acknowledges the European Communities’ argument that conditionality in the context of
traditional MFN clauses in bilateral treaties may relate to conditions of trade
compensation for receiving MEN treatment, the Panel does not consider this to be the full
meaning of “unconditionally” under Article I.1. Rather, the Panel sees no reason not to
give that term its ordinary meaning under Article I.1, that is, “not limited by or subject to
any conditions.” Because the tariff preferences under the Drug Arrangements are
accorded only on the condition that the receiving countries are experiencing a certain
gravity of drug problems, these tariff preferences are not accorded “unconditionally” to
the like products originating in all other WTO Members, as required by Article I:1. The
Panel therefore finds that the tariff advantages under the Drug Arrangements are not
consistent with Article I:1 of GATT 1994.
This makes sense when dealing with tariff lines up to the 6-digit level, where description
of goods is harmonized. It is still an open question, however, whether this line of thinking
is reconcilable with the contractual freedom that WTO Members in principle have to draft
their own 8-digit-level classifications. As things stand, there is no definitive answer to
this question.!!
The treatment of quantitative restrictions
Import and export quantitative restrictions (QRs) are illegal per se.Even non-
discriminatory QRs violate Art. XI GATT, since there is no discrimination-test embedded
in Art. XI GATT. The term QR is not self-interpreting. Duties and taxes cannot be
considered QRs, nor can domestic measures that are enforced at the border (Interpretative
Note ad Art. III). For example, a ban on sales of asbestos containing materials that is
enforced at the border, if challenged, will be adjudicated under Art. III GATT, not Art. XI
GATT.
WTO case law has opted for defining the term QR to cover not only cases where a
numerical target has been set, but any QR-equivalent measure. The GATT panel report on
Japan — Semiconductors established the wide coverage of Art. XI GATT. It held for the
proposition that when a government provides incentives to private parties to act in a
manner inconsistent with Art. XI GATT, such behavior is GATT inconsistent. In the
semiconductor case, Japanese firms raised prices (which led to reduced exports) as a
result of incentives (including administrative guidance and monitoring of costs and
prices) by the Japanese government. [his report 1s important in two distinct ways: first,
showing that measures which are not strictly speaking quantitative restrictions but result
in reduction of trade, are covered by Art. XI GATT; and, second, that government actions
that do not require (mandate) a specific behavior of private parties, but do provide
incentives to do so, can be caught by Art. XI GATT.
This panel report has been cited by all subsequent reports dealing with the interpretation
of Art. XI GATT. Thus, it is now settled case law that the term QR covers not only de jure
quantitative restrictions (numerical limits or targets) but also de facto QRs (anything
which might operate as a QR), irrespective whether the subject of the challenged activity
is the state or the private sector. However, for private actions to be captured by the
prohibition enshrined in Art. XI GATT, they must be attributed to a government. The
degree of intervention for attribution to the state is quite low: incentives suffice.
Internal measures: national treatment
The national treatment obligation (Art. II GATT) appears right after Art. II GATT
(consolidation of customs duties), indicating that Art. III GATT becomes legally relevant
only once imports have paid the “entry ticket” into a particular market (in the form of
customs duties). The national treatment obligation is needed because virtually all
domestic policy instruments are left unconstrained in the WTO, potentially leaving the
parties with plenty of scope to undo tariff bindings. In order to limit the scope for this,
tariff bindings are complemented with an obligation not to discriminate between domestic
and foreign products once the latter have entered a given market. The purpose of Art. III
GATT is thus to prevent “concession erosion” through the discriminatory application of
domestic policies.
National treatment is a cost-effective defense against opportunistic (protectionist) use of
domestic policy instruments to circumvent tariff bindings. The essential function of the
national treatment instrument is to make domestic measures b/unter instruments of
protection. In the case of taxation, the more “fine tuned” tax policy instruments
governments have at their disposal, the more tempting it will be for them to pursue
beggar-thy-neighbor policies. If all (like) domestic products have to be burdened with the
higher taxes imposed on imported products, taxes become a less attractive instrument of
protection.
As noted in the AB report on Japan — Alcoholic Beverages II (p. 16):
The broad and fundamental purpose of Article III is to avoid protectionism in the
application of internal tax and regulatory measures. More specifically, the purpose of
Article III “is to ensure that internal measures ‘not be applied to imported or domestic
products so as to afford protection to domestic production.”” Toward this end, Article III
obliges Members of the WTO to provide equality of competitive conditions for imported
products in relation to domestic products.
The coverage of the national treatment obligation in positive law
The national treatment obligation extends to both bound and unbound tariff lines. The
wording of Art. III GATT does not explicitly address this point. The AB in its report on
Japan — Alcoholic Beverages IT eliminated any remaining doubt as to the coverage of
Art. III GATT in the following terms:
The Article III national treatment obligation is a general prohibition on the use of
internal taxes and other internal regulatory measures so as to afford protection to
domestic production. This obligation clearly extends also to products not bound under
Article II.
It follows that the obligation not to discriminate in favor of domestic products is legalese
for the obligation not to protect domestic pro- duction. The term protection, however,
lends itself to different interpretations and, as things stand, we still lack an operational
definition of this term.
Art. Ill GATT extends to regulatory interventions of both a fiscal (Art. IIL2 GATT), and
non-fiscal nature (Art. HI.4 GATT). A fiscal imposition on goods affects trade by
modifying the price of the good concerned. This explains why Art. II.2 GATT (which
covers fiscal measures) does not include the words “affecting trade,” whereas these words
are found in Art. IH.4 GATT (covering non-fiscal measures). In contrast to European
Community disciplines on free movement of goods — where over time the coverage of
analogous disciplines has been reduced to avoid punishing behavior that has minimal or
unintended/tangential impacts on the free movement of goods, WTO adjudicating bodies
interpret the term “affecting” very widely. To date, there has never been a case which has
failed the “affecting” threshold in the context of Art. LH GATT litigation.
Two policy measures are explicitly exempted from the national treatment obligation by
virtue of Art. IIL§ GATT: subsidies and government procurement. The treatment of the
former is regulated in a separate multilateral agreement, the WTO Agreement on
Subsidies and Countervailing Measures (SCM). This Agreement defines subsidies as
government schemes that grant benefits to specific entities. It further divides subsidies
into actionable and prohibited. The latter comprise explicit export subsidies (Art. 4
SCM). The former are defined by default: any scheme which is attributed to government
and confers a benefit to a specific entity and which is not a prohibited subsidy is an
actionable subsidy.
Exclusion from the national treatment obligation for procurement essentially means that
WTO Members do not have to abide by the national treatment obligation when
government entities purchase goods without the intention to re-sell. In light of the
importance of the government procurement market but also the unwillingness of many
WTO Members to limit their discretion in this respect, a sub-set of WTO Members
entered into a plurilateral agreement (¡.e., bindingonly the signatories) whereby they
essentially re-introduced the national treatment obligation for all purchases of entities
listed in their respective schedules of concessions.
The “new generation” WTO agreements (such as the Agreement on Technical Barriers to
Trade (TBT), and the Agreement on Sanitary and Phytosanitary Measures (SPS)) are, in
this context, less troublesome in the sense that they reflect more elaborate understandings
of the nondiscrimination obligation. A domestic regulation can simultaneously fall under
the Art. III GATT, the TBT and/or the SPS agreements. If this is the case, then the SPS
prevails by virtue of Art. 1.5 TBT (which establishes that the SPS takes precedence over
the TBT) and because the EC — Asbestos jurisprudence establishes that the TBT takes
precedence over the GATT.
According to the TBT agreement, a WTO Member enacting a technical regulation or
standard (as defined in the TBT) must respect the national treatment obligation (assuming
no relevant international standard exists, which, in principle, must be followed according
to Art. 2.4 TBT) and will further have to ensure that its legislation is necessary for it to
achieve its unilaterally set regulatory objective. Necessity in this context means that
WTO Members are free to pursue any objective they deem appropriate but at the same
time have to choose the means that will have the least possible negative repercussions on
international trade while pursuing this objective. Necessity does not oblige WTO
Members to target their objectives in the sense that they must always use the first-best
instrument to realize a social preference. It does go some way towards this direction,
however, since the underlying assumption of this principle is that the most onerous of
international trade measures will not be utilized. Hence, necessity emerges as a proxy
relevant in the investigation whether or not protection was indeed intended and/or
provided.
The SPS Agreement goes even further in this respect. Besides what has been described
above, the SPS Agreement obliges WTO Members to base their interventions on
scientific evidence and a process of risk assessment (only if there is no relevant scientific
evidence may governments invoke the so-called “precautionary” principle) and also to
ensure some coherence (Art. 5.5 SPS) in their health- and/or environmental policies.
Science and policy coherence are additional proxies that could help distinguish wheat
from chaff and enable adjudicating bodies to minimize both false positives and negatives.
Of course, the system is still far from providing a “one size fits all” solution in this
context. But some significant steps in the right direction were taken when enacting the
TBT and SPS Agreements.
A brief summary
To sum up our discussion so far, adherence to the WTO contract implies a ban on the use
of QRs, an obligation (the extent of which depends on domestic preferences and
international negotiations) to consolidate customs duties (and thus avoid volatility which
in this respect is synonymous to increased transaction costs), and an obligation to avoid
two forms of subsidization (pecuniary, with respect to export subsidies, and regulatory by
virtue of the national treatment provision). Otherwise, WTO Members are essentially free
to unilaterally pursue trade policies, subject to the constraint that national regulators do
not discriminate between domestic and foreign like goods. This limited reach reflects the
intentions of the founders of the GATT to limit their cooperation to so-called “negative
integration,” as opposed to more far-reaching positive integration (harmonized policies).
Since we still live in a world with import duties, the role of national treatment in its
current, static dimension is to ensure that the value of negotiated concessions are not
undermined through unilateral policies. Eventually however, we will move to a world
without customs duties. This does not necessarily entail that WTO Members will lose the
incentive to pursue beggar-thy-neighbor policies through instruments other than customs
duties. Indeed, this is exactly what internal fiscal and non-fiscal instruments can achieve.
In its dynamic dimension, the national treatment obligation might have to be re-evaluated
and redrafted. In a tariff-free world, internal non-discriminatory instruments might prove
to be a formidable obstacle to trade liberalization. A dynamic approach towards national
treatment might lead to international negotiations that focus on the efficiency of domestic
policies, that is, go beyond the existing parameters of non-discrimination.