Jump to ContentJump to Main Navigation
The Political Economy of the World Trading SystemWTO and Beyond$

Bernard Hoekman and Michel Kostecki

Print publication date: 2001

Print ISBN-13: 9780198294313

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/019829431X.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 29 September 2020

Safeguards and Exceptions

Safeguards and Exceptions

(p.303) 9 Safeguards and Exceptions
The Political Economy of the World Trading System

Bernard M. Hoekman

Michel M. Kostecki (Contributor Webpage)

Oxford University Press

Abstract and Keywords

The various mechanisms allowing for the re‐imposition of trade barriers are discussed in this chapter, which summarizes the rules on––and the economics of––the use of instruments of contingent protection. These have been very important in dealing with domestic political pressures and allowing the pursuit of non‐economic objectives. In practice, they have often been abused, to the detriment of both national and global welfare. The chapter focuses mainly on the safeguards and exceptions embodied in GATT (General Agreement on Tariffs and Trade); those of the GATS (General Agreement on Trade in Services) are very similar or still in development. The different sections are as follows: Renegotiation of concessions; Waivers; Emergency protection and VERs (voluntary export restraints); Antidumping actions; Measures to countervail subsidized imports; Trade restrictions for balance of payments purposes; Infant industry protection; General exceptions; and Conclusion.

Keywords:   anti‐dumping, balance of payments, concessions, contingent protection, GATS, GATT, General Agreement on Tariffs and Trade, General Agreement on Trade in Services, imports, infant industry protection, safeguards, subsidized imports, trade barriers, trade restrictions, VER, voluntary export restraint, waivers

Virtually all international trade agreements or arrangements contain safeguard provisions and exceptions. Broadly defined, the term ‘safeguard protection’ refers to a provision in an agreement permitting governments under specified circumstances to withdraw—or cease to apply—their normal obligations in order to protect (safeguard) certain overriding interests. Safeguard provisions are often critical to the existence and operation of trade‐liberalizing agreements, as they function as both insurance mechanisms and safety valves. They provide governments with the means to renege on specific liberalization commitments—subject to certain conditions—should the need for this arise (safety valve). Without them governments may refrain from signing an agreement that reduces protection substantially (insurance motive). This chapter focuses primarily on the safeguards and exceptions embodied in the GATT. Those of the GATS are either very similar or still in an embryonic stage.

WTO provisions in this area can be separated into two categories. The first are those that can be used in the event of the occurrence of a predefined set of circumstances which legitimize temporary increases in import barriers. The second constitute permanent exceptions to the general obligations. The first category can be further divided into those dealing with so‐called ‘unfair’ trading practices (such as exports benefiting from actionable subsidies) and those that can be applied without having to demonstrate ‘unfairness’. Provisions that allow for the temporary suspension of obligations include:

Antidumping (AD): measures to offset dumping—pricing of exports below what is charged in the home market—that materially injures a domestic industry (Article VI GATT);

Countervailing duties (CVDs): measures to offset the effect of subsidization that materially injures a domestic industry (Article VI GATT);

Balance of payments (BOP): restrictions on imports to safeguard a country's external financial position (Articles XII and XVIII:b GATT; Article XII GATS);

Infant industries: governmental assistance for economic development, allowing import restrictions to protect infant industries (Articles XVIII:a and XVIII:c GATT);

Emergency protection: temporary protection in cases where imports of a product cause or threaten serious injury to domestic producers of directly competitive products (Article XIX GATT);

Special safeguards: provisions embodied in the Agreements on Agriculture and Textiles and Clothing allowing for actions to be taken to restrict trade; and


General waivers: allowing members to ask for permission not to be bound by an obligation (Article IX WTO). In contrast to the other mechanisms, this requires formal approval by the WTO Council.

Provisions allowing for permanent exceptions from the general obligations of the Agreement include:

General exceptions: measures to safeguard public morals, health, laws and natural resources, subject to the requirement that such measures are nondiscriminatory and are not a disguised restriction on trade (Articles XX GATT; XIV GATS);

National security: allowing intervention on national security grounds (Articles XXI GATT; XIV bis GATS; 73 TRIPS); and

Renegotiation or modification of schedules: allowing for the withdrawal of concessions (bound tariff reductions or specific commitments) if compensation is offered to affected members (Articles XXVIII GATT; XXI GATS).

Most of these provisions allow for protection of a specific industry. Only three have an economy‐wide rationale (balance of payments, general exceptions, national security). All the industry‐specific instruments are (imperfect) substitutes for each other, as to a large extent they all address the same issue: protecting domestic firms from foreign competition. In practice the balance of payments provision was also used by developing countries to protect specific industries.

The GATS does not have provisions allowing for contingent or infant industry protection, and an analogue to GATT Article XIX remains to be drafted (see chapter 7). In large part this reflects the difficulty of applying these concepts to trade in services. The GATS does contain provisions allowing for actions to safeguard the balance of payments, for general exceptions and for renegotiation of commitments. These provisions are similar to those of the GATT, except that the language on modification of schedules differs from GATT by calling for mandatory arbitration if no agreement can be reached on compensation.

The goal of the drafters of the GATT was that renegotiation would be the primary mechanism to deal with a need for permanent rebalancing of concessions, and that Article XIX would be used to grant temporary protection to industries finding it too difficult to confront increased import competition. The AD and CVD provisions were included in large part at the behest of the US, Canada and several European nations, which had such statutes on the books, although they were rarely used. During the first 20 years of the GATT renegotiations and Article XIX were the major instruments used (Figure 9.1). Over time, however, industrialized country lobbies increasingly (p.305)

                      Safeguards and Exceptions

Fig. 9.1 Use Of Safeguard Instruments, 1948–93

Source: Finger (1999).

shifted away from Article XIX actions towards VERs to obtain relief from import competition. VERs became a major instrument of protection in the 1970s as they ensured some compensation for affected exporters and were often directed against countries that did not have initial negotiating rights or principal supplier status (see chapter 4). It was estimated that in the early 1980s, VERs covered some 10 percent of world trade, and that the trade‐weighted (p.306) average tariff equivalent of the VERs was on the order of 15 percent (Kostecki, 1987).

As of the mid‐1980s the instrument of choice became AD. Between 1980–6, the EU imposed 213 AD actions, as compared to only 10 Article XIX measures. In the same period, the US imposed five Article XIX measures, as compared to some 195 AD actions (Finger and Olechowski, 1987). The revealed preference for AD and VERs reflected the fact that the conditions that needed to be satisfied to invoke Article XIX protection were relatively stringent. As discussed below, until this was changed in the Uruguay Round, Article XIX actions had to be nondiscriminatory and affected exporters had the right to compensation (or failing adequate compensation, could seek authorization from the GATT Council to retaliate). Governments preferred VERs and AD, as these instruments allowed them to discriminate across exporting countries and did not require compensation.

Developing countries have frequently invoked Article XVIII:b of the GATT to justify temporary protection, often because of a desire to use QRs. If developing countries desired to impose tariffs for BOP reasons, they usually would not have to invoke Article XVIII, because most had either not bound their tariffs or had bound them at high ceiling rates. In such cases countries are free to impose higher tariffs without being confronted with a compensation requirement. Over time the use of Article XVIII:b by developing countries declined, in part due to efforts by the IMF and the World Bank to induce a shift towards more effective and efficient instruments to deal with BOP problems. Table 9.1 provides a brief summary of the frequency with which various instruments have been invoked. Whatever the political rationale for safeguard instruments, their mere existence may reduce competitive pressure on domestic import‐competing firms. They are also all inefficient, in the sense that the costs to consumers are almost invariably larger than the benefits that accrue to the protected industry. In addition, industries can be expected to exploit substitution possibilities across instruments if these exist, making it more difficult for governments to control trade policy.

The various provisions allowing for protection under the GATT can seriously undermine the liberalizing dynamic of the WTO, and limit the usefulness of the WTO to governments that seek protection from protectionist lobbies. Governments (and their advisors) find it very difficult to sell the argument that it makes no economic sense to draft legislation that allows the various WTO provisions to be invoked. Invariably the response will be to point to the US, Canada or the EU—all active users of contingent protection. ‘If they use it, why should we refrain’ is a frequently heard argument. In (p.307)

Table 9.1 Frequency of Use of Safeguard Provisions

Instrument and GATT article

Frequency of use

Periodic—three year—renegotiations at the initiative of the country desiring to raise a bound tariff rate, Articles XXVIII:1 and XXVIII:5.

1955–99: 207 instancesa

‘Special circumstances’ renegotiations (requires GATT authorization), Article XXVIII:4.

1948–2000: 64 instancesb

Waivers under GATT Article XXV.

113 (through March 1994)

40 (through August 2000)

Waivers under Article IX WTO.

Withdrawal of a concession for infant industry purposes, Article XVIII:a.

9 (through July 2000)b

Infant industry protection (requiring a release from bindings), Article XVIII:c.

9 (through July 2000)b

Measures by developing countries for BOP purposes, Article XVIII:b.

Used by 24 countries at least once during 1974–86 (total of 3,434 restrictions)

Emergency protection, Article XIX.

1950–94: 150 actions (3.4 per year) 1995‐May 2000: 49 cases (9 per year)

Special safeguards under the ATC.

1995–97: 31 requests

Countervailing duties, Article VI.

July 1985–92: 187 actions 1995–2000: 110 measures (including undertakings) as of August 2000

Antidumping duties, Article VI.

July 1985–92: 1148 investigations 1995–June 2000: 1105 measures

(a) Renegotiations were minimal during 1995–2000 as tariffs were modified under rectification procedures or in the context of adopting the Harmonized System.

(b) Zero instances under the WTO through 2000.

Source: Finger (1996, 1999), Finger and Schuknecht (2000), WTO secretariat.

1996 developing countries as a group overtook industrialized countries in terms of the number of antidumping investigations initiated. The trends are therefore worrisome. Some progress was made to strengthen the rules in the Uruguay Round. Much remains to be done. Views on the impact of contingent protection depend significantly on whether these measures are seen as ‘facilitating devices’, allowing liberalization to proceed, or as mechanisms (p.308) allowing for backsliding. The debate in this area is analogous to that between optimists and pessimists looking at a glass that is partially filled: an optimist says the glass is half full, the pessimist says it is half empty. Although there is general recognition that contingent protection plays an important political role, the pessimists are concerned that there are excessive opportunities to reimpose protection. Economists also emphasize that some of the instruments that are legal under the WTO make no economic sense (antidumping in particular).

9.1. Renegotiation of Concessions

The GATT allows governments to renegotiate tariff concessions and schedules (Article XXVIII). Renegotiation centers on the compensation that must be offered as a quid pro quo for raising a bound rate. Modification of schedules takes three basic forms: ‘open season’, which may be conducted every three years following a binding; ‘special circumstances renegotiations’, which may take place when approved by GATT contracting parties; and ‘reserved right renegotiations’, which may occur anytime during the three year period following a binding if a notification is made by interested governments to that end (Dam, 1970).

Developing countries may follow a simplified procedure to modify or withdraw concessions. In negotiating the compensation required, account is taken of the interests of the country with which the concession was originally negotiated (which has so‐called ‘initial negotiating rights’—INRs), the interest of the country having a ‘principal supplying interest’, as well as that of countries having a ‘substantial interest’. Principal or substantial supplying interest requires a major or a sizeable share, respectively, in the market concerned, determined on the basis of import statistics for the last three years for which information is available.1

Countries having a substantial interest in the concession concerned (the negotiated tariff binding) have consultation rights only, whereas countries that have INRs or are principal suppliers, have negotiation rights. In disputed cases it is up to the Council to determine whether a given country is a principal supplier or whether it has a substantial interest. No such cases arose under GATT 1947. The main objective of the principal supplier rule is to provide for the participation in the negotiations, in addition to the country with the INRs, of countries with a larger share in the trade affected by the (p.309) concession than the country with INRs might have. This allows a balance to be maintained between the old, previously negotiated situation and new trade patterns that emerge over time. Exceptionally, when the concession to be withdrawn affects trade which constitutes a major part of the total exports of a given country, the country may also enjoy principal supplier status (Article XXVIII:1). The Understanding on the Interpretation of Article XXVIII enhanced the opportunities of affected exporters to participate in tariff renegotiations. The WTO member for which the relative importance of exports of the product on which a tariff is increased is the highest (defined as exports of the product to the market concerned as a proportion of the country's total exports) is considered to have a principal supplying interest if it does not already have so (or an INR) under GATT 1947 procedures. If no agreement is reached on compensation, affected countries may withdraw equivalent concessions

GATS Article XXI is analogous to GATT Article XXVIII, allowing for members to withdraw commitments after a three‐year period has elapsed from the time that the commitment entered into force. The intent to modify must be notified to the GATS Council, and gives rise to compensation discussions. If agreement cannot be reached on compensation, the GATS provides for arbitration (no retaliation is allowed until the arbitration process has been completed). If the recommendations resulting from the arbitration are not implemented, affected members that participated in the arbitration may retaliate without needing authorization by the GATS Council. In this respect the GATS goes beyond GATT, which only provides for countries concerned to refer disagreements regarding compensation to the Council for Trade in Goods, who may in turn ‘submit their views’.

The mechanisms for—and disciplines on—modification of tariff schedules are important. Before the completion of the Uruguay Round, on average renegotiation of concessions occurred every year with respect to some 100 items, as compared to some 80,000 tariff lines bound. During the 1955–95 period over 30 GATT contracting parties utilized the renegotiation option more than 200 times (Table 9.1). In the first five years of the WTO, renegotiations did not occur due to the fact that adjustments occurred in the context of members adopting and implementing the Harmonized System (see chapter 5).

9.2. Waivers

Tariff renegotiations are limited in nature: by definition they only pertain to instances in which a country wants to raise tariffs above previously bound (p.310) levels. Article XXV:5 GATT allows a member to request a waiver from one or more other obligations. The conditions under which waivers are granted are negotiated. Over 100 waivers were granted in the first 45 years of GATT history (Table 9.1), of which 44 were still in effect in 1994. From a systemic perspective, the waiver option allows for members to obtain an exemption from a specific rule in situations where they might otherwise have been forced to withdraw from the Agreement because of political imperatives at home. Waivers can be good or bad from an economic perspective. For example, a number of waivers were granted under GATT 1947 to countries allowing them to impose surcharges on imports for BOP purposes. Although this is an inferior instrument to deal with a BOP problem, at least it is better than the instrument called for by the relevant GATT provision—that is, QRs.

By far the most famous waiver was one requested by the US in 1955. As noted in chapter 5, QRs are allowed under Article XI of the GATT for agricultural commodities as long as concurrent measures are taken to restrict domestic production or to remove a temporary domestic surplus. Although it was the US that drafted this rule when negotiating the GATT, it proved too stringent for Congress. The latter did not wish to be bound by any international agreement and forced the Administration to ask for a waiver of this obligation in 1955. The waiver was necessary as existing US programs supported domestic industries such as sugar and dairy without incorporating any incentives to reduce output. The root of the problem was Section 22 of the Agricultural Adjustment Act, which states that the Secretary of Agriculture must advise the President if he believes any agricultural commodity is being imported so as to interfere with Department of Agriculture price support programs. Depending on the finding of an investigation into the matter, tariffs or QRs may be imposed. Because Section 22 violated GATT rules, US Administrations were reluctant to apply it. However, Congress had no such inhibitions, and amended Section 22 in 1951 to require the President to carry out its provisions regardless of international agreements, that is, the GATT (Evans, 1972: 72).

Under the WTO, disciplines on waivers were tightened. Article IX WTO allows waivers to be requested for any obligation imposed under a Multilateral Trade Agreement. Waivers under the WTO are time bound—in contrast to the GATT 1947—and are reviewed annually to determine if the exceptional circumstances requiring the waiver continue to exist. Any waiver in effect at the entry into force of the WTO was to expire by January 1997, unless extended by the WTO Ministerial Conference. Some 40 waivers were granted in the first five years of the WTO.

(p.311) 9.3. Emergency Protection and Vers

Article XIX is GATT's general safeguard clause. It permits governments to impose emergency measures to protect domestic producers seriously injured by imports. The main rationale for the general safeguard clause is to allow some flexibility with respect to tariff commitments, thereby promoting trade liberalization efforts. Article XIX is a safety valve. Designing a safeguard mechanism in such a way that a balance is achieved between making it difficult to open the safety valve and avoiding an explosion of the boiler is not easy. The drafters of the GATT chose to be rather strict in this regard. Necessary conditions for the invocation of Article XIX under GATT 1947 were: (1) the existence of increased imports; which (2) resulted from unforeseen developments; (3) were the consequence of trade liberalization negotiated in a MTN; and (4) caused or threatened serious injury to domestic producers.

Safeguard measures were to be imposed on a nondiscriminatory basis. The interests of affected exporting countries were protected by a requirement that they be compensated. If no agreement was reached in consultations on compensation, an exporting country could be authorized to retaliate (suspend equivalent concessions or other obligations) against the safeguard‐taking country. The compensation requirement made Article XIX a substitute for Article XXVIII, the main difference being that the latter allows for a permanent change. Although Article XIX actions were to be temporary in principle, no formal time limits were imposed. As a result some actions taken by contracting parties lasted for many years (Sampson, 1987).

GATT contracting parties took only 150 official safeguard actions during the 1948‐94 period. Of these, only twenty involved payment or offer of compensation—mostly in earlier cases—while retaliation occurred in thirteen instances (GATT, 1994b). Article XIX was therefore used relatively infrequently. Reasons for this included the requirement that safeguard actions be nondiscriminatory (affect all exporters), a preference for QRs (much more difficult to implement in a nondiscriminatory manner than a tariff), the need to offer compensation, and the fact that in some jurisdictions (such as the US) granting of emergency protection is subject to the discretion of the President, who is required to take into account the impact of taking action on the economy. The relatively stringent conditions for obtaining Article XIX cover for protection reflects the fact that such protection violates earlier tariff commitments. This is not the case under AD. As dumping is defined to be unfair, actions are legitimized as long as it is shown that dumping occurred and that it materially injured domestic industries.

In addition to AD—discussed below—VERs became a favorite safeguard instrument in the 1970s and 1980s. VERs were used to restrain exports of (p.312) steel and automobiles. While GATT‐illegal (GATT, 1994b, p. 434)—with the exception of the MFA restraints, which had been sanctioned by the GATT—VERs did not give rise to formal dispute settlement cases. The reason for this was that no one had an incentive to bring cases. Third country exporters, including the principal suppliers with which original tariff concessions on the goods involved had been negotiated, did not oppose VERs restricting their (new) competitors, while affected exporters tended to accept VERs because they allowed them to capture part of the rent that was created. Instead of being confronted with an import tariff, the revenue of which is captured by the levying government, a VER involves a voluntary cut back in volume by exporters. This reduction in supply will raise prices—assuming that other exporters do not take up the slack. Exporters therefore may get more per units sold than they would under an equivalent tariff. Essentially they obtain what would be the quota rents if QRs were to be used (see annex 2). There is a very large literature on VERs that will not be discussed here (see section 9.10). The key point to remember about VERs is that they imply some direct compensation of affected exporters and selectively target exporters. Thus, they partially meet GATT 1947 compensation requirements, while allowing for the circumvention of its nondiscrimination requirement.

By the time of the Uruguay Round, the major objective of ‘target’ countries was to constrain the use of AD and VERs and reassert the dominance of Article XIX in instances where the need is to slow the growth of imports: the majority of cases. The problem was how to achieve this goal. Two options were available: either to tighten the disciplines on the use of VERs and AD, or to reduce the disincentives to use Article XIX. Both approaches were pursued. Little progress was achieved on the AD front, but VERs were banned and Article XIX made more attractive to import‐competing industries. Progress on the latter front was facilitated because importing country governments increasingly recognized that VERs were costly—something that economists did not stop from pointing out in study after study (for example, De Melo and Tarr, 1992).

The Uruguay Round Agreement on Safeguards

A major achievement of the Uruguay Round Agreement on Safeguards is a prohibition of VERs and similar measures on the export or the import side (such as export moderation, export‐price or import‐price monitoring systems, export or import surveillance, compulsory import cartels and discretionary export or import licensing schemes). Any such measure in effect as of January 1995 was to be brought into conformity with the new rules or phased out by mid‐1999. The agreement requires that safeguard measures be (p.313) taken only if an investigation demonstrates that imports have increased so much as to have caused or threaten serious injury to an import‐competing domestic industry. Investigations must include reasonable public notice to all interested parties and public hearings or other mechanisms through which traders and other affected parties can present their views whether a safeguard measure would be in the public interest. Investigating authorities must publish a report setting forth their findings and reasoning. Serious injury is defined as a significant overall impairment in the situation of a domestic industry. In determining injury, the domestic industry is defined as those firms whose collective output constitutes a major share of total domestic output of the product concerned. Factors to determine whether increased imports have caused serious injury include the magnitude of the increase in imports, their change in market share, and changes in the level of sales, production, productivity, capacity utilization, profits, and employment of the domestic industry. The Appellate Body in its case law has made clear all mentioned factors must be examined. A causal link must be established between increased imports and serious injury or threat thereof. When factors other than increased imports are causing injury to the domestic industry at the same time, such injury may not be attributed to increased imports.

Protection must be limited to what is necessary to prevent or remedy serious injury. If a QR is used, it may not reduce imports below the average level of the last three representative years, unless clear justification is given that a different level is necessary to prevent or remedy serious injury. While in principle safeguard actions must be nondiscriminatory, quota rights may be allocated on a selective basis if the Committee on Safeguards accepts that imports from certain members have increased disproportionately in comparison to the total increase in imports, and the measures imposed are equitable to all suppliers of the product. Such ‘quota modulation’ may be maintained for four years at the most. QRs may be administered by exporters if this is mutually agreed. Thus, although VERs are prohibited, something analogous may be used if implemented as part of a GATT‐conform procedure. Safeguard actions based on absolute increases in imports that are consistent with the provisions of the Agreement do not require compensation of affected exporting countries for the first three years. In principle, safeguards should be degressive—the level of protection should decline over time—and not last more than four years. All actions are subject to a sunset clause. The maximum total number of years a safeguard may be applied is eight years. If an action is extended beyond four years, a necessary condition is that the industry is adjusting. If individual market shares of developing countries are less than three percent of total imports, and the aggregate (p.314) share of such countries less than nine percent of total imports, they are exempt from safeguard actions.

The Agreement on Safeguards brought existing practices that were GATT‐inconsistent inside the tent, but subjected their use to multilateral surveillance and rules. Thus, VER‐type measures are allowed by the WTO under certain conditions—in contrast to GATT 1947. While this implies a move away from economically preferred policies in an abstract sense, this is an inappropriate benchmark. A better counterfactual for comparison purposes would be continued circumvention of GATT 1947 disciplines, which had become increasingly irrelevant. The ban on VERs is a major achievement. Much will depend, however, on its implementation, which in turn will be a function of the willingness of WTO members to cross‐notify VERs and initiate dispute settlement procedures against countries requesting VERs. Past experience suggests that governments may not have a great incentive to do so. Another test is the extent to which quota modulation is applied.

Table 9.2 Safeguard Measures (1995–9)







OECD countries







Developing countries














Note: OECD member countries include Mexico and Korea). Annual data span November 1 to October 31.

Source: WTO, Safeguards Committee Annual Reports.

The use of safeguards has expanded substantially compared to the pre‐Uruguay Round period (Tables 9.1 and 9.2). Much of the increase reflects greater invocation of the instrument by developing countries. India, Korea and Argentina are among the countries that used safeguard actions most frequently in the post‐1995 period. The US is the primary user among industrialized countries, also being a leader in the use of quota modulation. This is a worrisome development as it suggests safeguards may be used in an antidumping‐like way.

Special safeguard actions have been used under the auspices of the WTO Agreement on Textiles and Clothing (ATC). The ATC contains a special safeguard clause in Article 6, which can be invoked during the implementation period of the ATC (that is, up to 2004) for products being integrated into the WTO. Actions may be taken if imports of a product increase so much as to cause or threaten serious damage to the domestic industry producing like (or directly competitive) products. Damage indicators include standard economic (p.315) variables such as output, productivity, capacity utilization, inventories, market share, exports, wages, employment, domestic prices, profits and investment (Article 6.3 ATC). Transitional safeguard actions can be applied on a discriminatory basis, in contrast to measures taken under the Agreement on Safeguards. They require demonstration of a sharp and substantial increase in imports, actual or imminent, from the targeted countries. Measures are not to exceed three years duration or until the product is integrated into GATT 1994, whichever comes first. Over 30 actions under the ATC were taken between 1995‐7, mostly by the US against developing country textile exporters but also by developing countries against each other (for example, Brazil has been an active user). The measures are reviewed by the Textiles Monitoring Body, and many were rescinded. Two actions led to dispute settlement cases, both involving the US. In both cases, the panels, supported by the Appellate Body, concluded that the US had violated the provisions of the ATC.2 The panels signaled that the transitional safeguards in ATC were to be regarded as exceptional instruments and that members invoking this provision of the ATC had to be in full compliance with the various criteria laid out in the agreement.

9.4. Antidumping Actions

Loosely defined, dumping occurs when a firm sells products on an export market for less than what is charged on its home market for the same product. Dumping is also said to occur if the export price of a product is below the cost of production. WTO rules allow action to be taken against dumped imports if dumping causes or threatens material injury to a domestic import‐competing industry. This is a weaker standard than the serious injury criterion used in the case of safeguards. AD is an option—there is no requirement to have an AD mechanism. WTO disciplines apply only if the option is invoked. Over 3,000 AD investigations have been initiated by WTO members since the late 1970s, of which over 2,000 occurred in the 1985–99 period. Between 1995 and 1999, during the first five years of the WTO, over 1,000 investigations were initiated (Table 9.3). The main users of AD have traditionally been Australia, Canada, the EU, and the US. Since the creation of the WTO, the use of AD by developing countries has increased dramatically. The top users of AD in 1999 included Argentina, Brazil, India and (p.316) South Africa. Mexico, an OECD member, became a major user during 1992–95. China is the most frequent target, followed by the US. If expressed in terms of cases per dollar of exports, AD tends to be much more important against developing economies. As of the late 1990s, AD was disproportionally used by developing economies and against developing economies (Finger and Schuknecht, 2000). AD duties tend to be very high when compared to the average tariff applied on manufactures in OECD nations. Duties of 30 to 50 percent are common.

Table 9.3 Antidumping Actions (1995–9)







OECD countries







Developing countries














Note: OECD includes Mexico and Korea).

Source: WTO, Antidumping Committee, Annual Reports.

At the time the GATT was first negotiated, AD was a non‐issue. It was included because a number of countries had legislation on the books, but it was rarely used. In 1950, there were only 37 cases, of which 21 taken by South Africa (GATT, 1958, p. 14). AD became increasingly controversial during the 1980s as its use expanded. Fifteen AD actions taken by contracting parties were challenged in GATT between 1989–94, of which five led to the creation of panels. In all cases the panel found against the country that had taken the AD action, and in a number of instances recommended that illegal AD duties be removed and reimbursed (Petersmann, 1994). This was very controversial, and led to dispute settlement in this area becoming a major issue in the Uruguay Round. Users sought to circumscribe the power of panels to conclude that specific AD actions violated GATT requirements, or to recommend restitution of duties in cases where AD measures had been imposed in violation of GATT rules.

A Brief Summary of GATT Disciplines

The basic GATT provision dealing with AD is Article VI. Starting in the Kennedy Round, reflecting increasing use of AD, attempts began to be made to further define multilateral disciplines in this area. A code on AD was negotiated in the Kennedy Round. It was opposed by the US Congress, and in practice was only applied insofar as its provisions were consistent with US (p.317) legislation. In the Tokyo Round, the AD code was renegotiated. This code—which only bound signatories—became the basis of the Uruguay Round agreement on AD (formally entitled the Agreement on Implementation of Article VI).

Dumping is defined in GATT as offering a product for sale in export markets at a price below normal value. Normal value is defined as the price charged by a firm in its home market, in the ordinary course of trade. Trade is considered not to be ordinary if over an extended period of time (normally one year) a substantial quantity of goods is sold at less than average total costs (the sum of fixed and variable costs of production plus selling, general and administrative costs). If sales on its domestic market are too small to allow price comparisons, the highest comparable price charged in third markets is used. Alternatively, the exporting firm's estimated costs of production plus a reasonable amount for profits, administrative, selling and any other expenses may be used to determine normal value (the so‐called constructed value). In cases where there is no export price or where it appears to the investigating authorities that the export price is unreliable because of a relationship between the parties to a transaction, the export price may also be constructed. Constructed values should be based on the price at which the imported products are first resold to an independent buyer, or if they are not resold to an independent buyer, ‘on such reasonable basis as the authorities may determine’. The comparison of the export price and the normal value must be made at the same level of trade (normally ex‐factory) and as close as possible to the same time. Allowance is to be made for differences in factors such as the conditions and terms of sale, the quantities involved, physical characteristics, and differences in relevant costs. In an investigation, exporters must be allowed at least 60 days to adjust their export prices to reflect sustained movements in exchange rates during the period of investigation.

Actions against dumping may only be taken if it can be shown that it has caused or threatens material injury of the domestic import‐competing industry. Injury determinations must be based on positive evidence and involve an objective examination of the volume of the dumped imports, their effect on prices in the domestic market, and the impact on domestic producers of like products. A significant increase in dumped imports, either in absolute terms or relative to production or consumption in the importing country, is a necessary condition for finding injury. Significant price under‐cutting of domestic producers, a significant depressing effect on prices, or the level of the dumping margin are other indicators that may be used. The term significant is not defined. While differences in views as to what is significant might be dealt with through the dispute settlement process, this (p.318) possibility is limited as panels are constrained in their ability to overrule substantive decisions taken by domestic investigating authorities (see below).

An illustrative list of injury indicators is given in the agreement. These include actual and potential decline in sales, profits, output, market share, productivity, return on investments, or utilization of capacity; factors affecting domestic prices; the magnitude of the margin of dumping; actual and potential negative effects on cash flow, inventories, employment, wages, growth, ability to raise capital or investments. This list is not exhaustive, and no single or combination of factors is decisive. Dumped imports must be found to cause injury because of dumping. The necessary causality must be established on the basis of all relevant evidence before the authorities. Any other known factors that are injuring the domestic industry must be taken into account, and may not be attributed to the dumped imports. Factors that may be relevant include the volume and prices of imports not sold at dumping prices, contraction in demand or changes in the patterns of consumption, trade restrictive practices of—and competition between—the foreign and domestic producers, developments in technology, and the export performance and productivity of the domestic industry.

This very brief summary of the major elements of the AD agreement illustrates that the wording is technical and complex. Many of its articles are only decipherable for lawyers specialized in this particular area of trade law. The wording of the agreement reflects numerous compromises reached in the Tokyo and Uruguay Rounds. It is a combination of elements of the domestic laws and practices of major WTO members and periodic attempts by target countries to limit the protectionist biases inherent in the use of AD in most jurisdictions. These biases have proven difficult to eradicate. The reason for this is simple: AD is fundamentally flawed from an economic perspective and cannot be fixed by tinkering with the methodological arcana of investigations. For all practical purposes, there is nothing wrong with dumping, as it is a normal business practice. The problem is antidumping. As argued by Finger (1993a), the only way to deal with AD is through efforts by negatively affected parties (consumers, users) to alter domestic implementing legislation to allow their interests to be represented in AD cases.

What's Wrong With Dumping?

Dumping is not prohibited by the WTO. All the WTO does is to establish certain rules that apply to governments that seek to offset dumping. Why dumping occurs is not considered relevant under GATT rules. From a normative, economic, perspective this is important, however. A typology of business motivations for dumping is presented in Table 9.4. (p.319)

Table 9.4 Motivations for Dumping

Type of dumping

Objectives of the exporting firm

Sporadic or random

No deliberate intention to dump

Price discrimination

Maximize profits given differences in demand across markets


Cover at least variable costs and maintain capacity during periods of slack demand


Minimize losses due to excess capacity or to deter entry by competitors

Scale economies

Price below cost initially with expectation of recouping investment outlays (fixed costs) over time as sales expand


To establish a new product as the market leader—revenue, not profit maximization


To attack a dominant supplier in an export market


To establish a monopoly on an export market

Sporadic dumping may occur without any deliberate intention if the exporting firm has to decide on how much to produce before demand conditions or exchange rates are known. Sporadic dumping may also arise from a lack of experience in pricing a new product. The trading environment facing a firm is usually uncertain, so dumping will often be beyond the control of a firm. For example, unexpected changes in exchanges rates may lead a firm to dump even if it had no intention of doing so. However, in most cases dumping reflects a deliberate business strategy followed by exporting companies and constitutes a conscious, premeditated pricing practice aimed at the accomplishment of specific business objectives.

The best known motivation is simple price discrimination between markets. The wording of GATT rules specifically targets this rationale. A firm having some control over prices and operating in two separate markets may find it advantageous to discriminate in its price formation in favor of foreign consumers in order to maximize profits. Price discrimination across markets will occur whenever demand for a product is more elastic in export markets than at home (that is, for any given change in price, foreign consumers change their demand more than domestic customers. Dumping in the sense of spatial price discrimination requires that there are barriers to reimportation of the dumped product into the exporter's home market. Otherwise, price differentials across markets would tend to be eliminated through arbitrage, allowing for transport and transactions costs. As discussed (p.320) further below, this suggests that the problem with dumping—if any—are the trade barriers that prevent such arbitrage.

As noted above, selling below average total cost can also constitute dumping, and it is not strictly necessary that prices charged in the export market be below those charged at home. Issues of trade policy and differing price elasticities then become irrelevant.3 The cost‐dumping case is important in practice, because many exporters produce exclusively or predominately for export, or sell only specific products for export. Often it may be in a firm's interest to sell below variable or even marginal costs (the cost of producing an extra unit of output) for a while. Although by doing so the firm will make a loss (because its fixed costs are not recovered), this may be necessary in the short run to establish or increase market share, or to enable the firm to move down its learning curve, thus increasing expected long‐run profits.

A firm may engage in cyclical dumping to stabilize its production over the business cycle. Dumping arises as the firm reduces prices to cover only average variable costs during periods of slack demand. This can be perfectly rational, insofar as the firm expects better times in the future and perceives the costs of laying off workers and reducing capacity to be higher than continuing production without covering all costs. Indeed, differences in labor markets and employment practices of firms across countries are one factor explaining why firms from some countries are more prone to engage in cyclical dumping than others (Ethier, 1982). If it is very costly for a firm to lay off workers—because of high legislated redundancy payments for example—it will continue to produce more than a firm which confronts a very flexible and less regulated labor market.

Certain forms of dumping have a pronounced strategic dimension. Exporting at prices below production cost may help deter entry by potential competitors into a firm's home market. This can be called defensive dumping (Davies and McGuinness, 1982). A firm may also price exports below total production cost on a longer term basis if such a strategy permits it to realize economies of scale (these exist if unit costs decline as output expands). A firm may need to move down its learning curve as fast as possible. As output increases, production workers tend to become more efficient and unit costs of production fall. Dumping in these cases is part of a strategy to attain an optimal scale of production.

(p.321) A related rationale for dumping arises in cases of new high‐technology products (such as video recorders in the 1970s and early 1980s). Here a firm may attempt to discourage domestic firms from engaging in the development of a competing product by establishing a large market share. Products where there is proprietary technology may foster dumping in the early stages of the product cycle. Profitability in such products often depends on consumers choosing a specific standard. In the video case, there were two main competing standards or formats in the 1970s, VHS and Betamax. Both needed to attract enough adherents to form a large enough customer base to recoup investment in R&D. When firms are trying to establish market share quickly, dumping may be part of an effective competitive strategy. A firm may also choose to export at prices that do not cover even marginal cost when, instead of maximizing profit, it prefers to maximize sales. Such ‘head‐on dumping’ may be used as part of an attack on a price leader in a given export market. Head‐on dumping was practiced extensively by Japanese semi‐conductor and electronics producers in export markets in the 1980s (Kotler et al., 1987).

The foregoing rationales for dumping are driven by market structures, business cycles, or the characteristics of the products that are produced. Of these deliberate strategies, only one is potentially detrimental to the welfare of the country importing the dumped product: predation. Predation was the original rationale for US AD legislation, passed in 1916. The fear was that a foreign firm (or cartel) could deliberately price products low enough to drive existing domestic firms out of business and establish a monopoly. Once established, the monopolist could more than recoup its losses by exploiting its market power. For predation to work, the monopolist (cartel) must not only eliminate domestic competition, it must also be able to prohibit entry by new competitors. For this to be possible it must either have a global monopoly or it must convince the importing government to impose or tolerate entry restrictions. It is not clear why a government would do this. Not surprisingly, in practice, post‐Second World War cases of successful predatory dumping are the exception, not the rule. Research by economists has demonstrated that over 90 percent of all AD investigations would never have been launched if a competition standard—potential threat of injury to competition, as opposed to injury to competitors—had been used as a criterion (Messerlin, 2000). Proponents of AD often have a narrower definition of predation in mind than the economic one described above. The fact that competition from other, outside sources will in most realistic circumstances prevent the formation of a monopoly is considered irrelevant. What matters is the (continued) existence of the domestic industry. But AD will not help in (p.322) achieving this objective. What is needed is adjustment of the industry, something that AD is unlikely to encourage.

Import‐competing firms usually object to underselling, and not to price discrimination or selling below cost. This has been emphasized by de Jong (1968), who noted that popular opinion refers to dumping when foreign producers are able to undersell the domestic supplier in his own market. As described by de Jong, this notion was translated as social dumping in early discussions concerning AD legislation. While this term was not clearly defined, it was apparent that it referred to underselling by foreign firms in the domestic market, made possible by lower labor costs abroad (that is, comparative advantage). In practice, it is underselling that importing‐competing firms consider unfair, reflecting their inability or unwillingness to meet the price set by a foreign competitor.

That predation has very little to do with AD as it is practiced is perhaps best illustrated by the United States, which has two antidumping statutes. One, the Antidumping Act of 1916, maintains a predation standard for antidumping, and is very rarely invoked. The other, the Tariff Act of 1930, as amended, has a price and cost‐discrimination standard, and is the one usually invoked by import‐competing industries. Interestingly, the 1916 Act was the subject of dispute settlement cases in 1998. The cases, brought by the EU and Japan, alleged that the existence of the 1916 Act violated the national treatment rule and WTO disciplines on AD (which stipulate that the remedy to offset dumping is limited to AD duties and undertakings). Japan objected to the provisions of the Act that the importation or sale of imported goods that is found to be unlawful (predatory) may constitute a criminal offence and give rise to claims for damages. The trigger for bringing the case to the WTO was a court action brought under the 1916 Act against affiliates of Japanese companies. These cases are somewhat ironic, in that there is a valid rationale for imposing civil liabilities or criminal penalties in cases where predation is found to exist, as long as the same disciplines apply to domestic firms. A problem in this instance is that US competition authorities may not intervene against pricing practices of firms that are in violation of the 1916 Act, as competition law enforcement has evolved significantly since that statute was passed.

What's Wrong With Antidumping?

AD constitutes straightforward protectionism that is packaged to make it look like something different. By calling dumping unfair, the presumption is that AD is fair and thus a good thing. This is good marketing, but bad economics. From an economic perspective there is nothing wrong with most (p.323) types of dumping. AD is not about fair play. Its goal is to tilt the rules of the game in favor of import‐competing industries (Finger, 1999).

Advocates of AD policies sometimes argue that AD is a justifiable attempt by importing country governments to offset the market access restrictions existing in an exporting firm's home country that underlie the ability of such firms to dump. Such restrictions may consist of import barriers preventing arbitrage, but may also reflect the nonexistence or weak enforcement of competition law by the exporting country. For example, the US has claimed that lax Japanese antitrust enforcement permits Japanese firms to collude, raise prices, and use part of the resulting rents to cross‐subsidize (dump) products sold on foreign markets. AD is clearly an inferior instrument to address foreign market closure because it does not deal with the source of the problem—the government policies that artificially segment markets, or allow this to occur. An AD duty may put pressure on affected firms to lobby their government to eliminate such policies—or to abolish private business practices that restrict entry—but does so in a very indirect manner. Once investigations are initiated, any changes in policies or practices cannot have an impact on the finding. In many cases there will not be significant barriers to entry, so there is not much to be done by exporters to improve access to their home markets. Indeed, under current procedures no account is taken of whether price discrimination or selling below cost is the result of market‐access restrictions. Building this into the analysis would appear to be a first necessary step.

A key problem with AD is the discretion that often is granted to investigating authorities—or, alternatively, the guidelines under which such authorities are forced to operate by law—to follow procedures that can make the instrument blatantly protectionist. In practice methodologies used to determine whether dumping has occurred and calculate the size of the dumping margin are such that high positive margins can be found in many circumstances. An often‐used practice is to calculate dumping margins by using methodologies that raise the normal value and lower the average export price, thereby increasing the dumping margin. Normal values can be biased upward by not including sales in the home market made at prices considered to be below cost, and by excluding sales in the export market that are above the calculated normal value. The latter procedure has been justified on the basis that ‘sales at a high price should not be allowed to conceal dumped sales’ (Hindley, 1994: 97). In cases where the normal value is constructed on the basis of costs, dumping margins can be inflated through the inclusion of high profit and overhead margins in the calculation of the normal value, but not allowing for this in the calculation of the costs of sales for exports.

(p.324) Another problem is that injury criteria may be manipulated by firms. This is potentially of some importance, as the injury test tends to be the main factor constraining the access of import‐competing industries to protection (Finger and Murray, 1990). Indicators of injury include trends in market share, employment, profits, capacity, capacity utilization, import penetration, and price underselling (the exporters charging prices below what is charged by the import‐competing industry). Many of these variables will not be closely linked to trends in imports, but depend on business cycle influences. While all of these indicators may to some extent be correlated with injury, many can be manipulated by firms, thus creating an incentive for ‘indirect rent‐seeking’ by either feigning that criteria have been met, or by deliberately taking actions that will induce injury as defined in the law (Leidy and Hoekman, 1991). This enhances the threat effect of AD, and may foster so‐called cascading of protection.

Under AD, protection follows automatically if the criteria are satisfied. Potential countervailing forces—such as users and consumers—remain outside the administrative process and are effectively neutralized by the law—they do not have legal standing in most jurisdictions. By invoking instruments of contingent protection, an upstream industry that produces an input may significantly injure a downstream industry that uses the input. This increases the probability that downstream firms will seek and gain protection in turn. Indeed, by initiating and winning an AD action, upstream suppliers may be able to manipulate the health of downstream firms to the advantage of both (Hoekman and Leidy, 1992a). As a result, instances of contingent protection may cascade along the chain of production.4

This vertical linkage across instances of contingent protection illustrates one way in which AD procedures may facilitate cartelization along the production stream. AD and similar laws may also facilitate tacit or explicit collusion by enforcing existing cartels and substantially reducing price competition in affected markets (Messerlin, 1989, 1990). A credible threat of invocation of unfair‐trade laws may provide a means for industries to engage in implicit collusion that could not otherwise be maintained (Leidy, 1994a). AD and similar procedures are initiated by private parties (firms or industry associations). These interested parties are active in pursuing the adoption of more restrictive rules, and employing credible threats of invocation of procedures (p.325) to negotiate VERs. The expansion of the scope and use of unfair‐trade laws has been characterized as reflecting the de facto privatization of trade policy (Messerlin, 1990).

Matters are made worse from a competition and transparency perspective by the practice of negotiating price undertakings with exporters that are subject to AD investigations. Once a preliminary dumping or injury margin has been established, exporters will have some sense of the likely magnitude of the final AD duty that could be imposed. This provides them with an incentive to offer to raise their prices (or cut back supply). Offering such undertakings is explicitly allowed for in the EU's AD legislation (EEC Regulation no. 2423/88), and AD investigations frequently result in undertakings being accepted by the European Commission. The US does not accept undertakings, but in practice similar deals are struck, the difference being that US industry is given (informal) assurances by exporters that they will reduce exports or raise prices. Such agreements often explain why US AD petitions are withdrawn at some point after an investigation has been launched. Undertakings are akin to VERs, and are concluded for the same reasons. Exporters prefer them to the alternative of paying AD duties. With an undertaking they can at least capture some of the rents that are created by reducing supply. The downside of undertakings is similar to the downside of VERs—they are less transparent to domestic users and consumers. Indeed, often it will not be generally known that there are undertakings in place. They may also hide a variety of collusive practices. Undertakings clearly are more detrimental to the national welfare of countries using them than duties are, raising the question why they are used. Possible explanations include transaction costs—it is easier for an administration to rely on industry to carry the burden of monitoring whether a deal is implemented than having customs apply discriminatory tariffs—and foreign policy considerations. By transferring rents abroad, the negative impact on trading partners is reduced.

It is often pointed out that one way to attenuate the negative effects of AD on national welfare is to impose filters that require determining what the likely impact on the economy would be of imposing AD duties. So‐called public interest clauses can be introduced that require the authorities to do this. They have not been very effective in practice, in part because the ‘public interest’ is not defined clearly enough, and in part because those that will be affected negatively are often not consulted. Although losers may be effectively excluded from the process of determining whether to intervene, they do have access to the political system, and may have a strong enough incentive to seek to change the law so that the interests of users are considered. Developments in this direction occurred in the EU and the US (p.326) during the 1990s. In 1996, the US Congress held hearings on possible modifications to US AD law, pressured to do so by influential industries that relied heavily on imports of components. In the EU, consumer groups have become more vocal and the EU Commission is becoming somewhat more sensitive to the welfare implications of antidumping. For example, in a 1994 case concerning imports of gum rosin, the Commission concluded that ‘the negative effects of antidumping measures on the users of gum rosin would be overwhelmingly disproportionate to the benefits arising from antidumping measures in favor of the Community industry’.5 This finding was motivated by the fact that the EU industry's production capacity was limited, so that imposition of duties would substantially increase the production costs for industries that use gum rosin as an input. Accordingly, the Commission concluded ‘that protective measures would not be appropriate and that it would not be in the Community interest to continue the proceeding’. The EU has also become more restrained in taking AD actions against imports from Eastern European nations, many of which are in the process of accession to the EU. The environment in major OECD countries thus may be slowly turning away from the enthusiastic use of AD that characterized the late 1980s and early 1990s (Finger, 1999). However, as noted, developing countries are stepping into the breach.

The Uruguay Round Agreement

The WTO Antidumping Agreement made some progress in disciplining the use of AD. A sunset clause was added. AD duties are to be terminated within five years of imposition, unless a review determines that both dumping and injury caused by dumped imports continues to persist or that removal of the measure would likely lead to the recurrence of dumping and injury. De minimis rules were agreed to. Duties may not be imposed if dumping margins are less than two percent, or the level of injury is negligible, or the market share of a firm is less than three percent and cumulatively less than seven percent for all exporters supplying less than three percent. Discretion with respect to methodologies used to determine dumping and injury margins was reduced. In effect, practices such as the biased averaging methodologies described earlier were authorized, but subjected to certain constraints. While these changes made AD somewhat less protectionist, many of the practices that were identified in the 1980s as leading to significant protectionist biases remained untouched.

(p.327) For example, although the agreement calls for an ‘average‐to‐average’ comparison of home and export prices in the determination of dumping, authorities may compare a normal value that has been calculated on a weighted average basis with the prices of individual transactions (that is, not take into consideration export sales above normal value), if they ‘find a pattern of export prices which differ substantially among different purchasers, regions, or time periods and if an explanation is provided why such differences cannot be taken into account appropriately by the use of a weighted average to weighted average or transaction‐to‐transaction comparison’ (Article 2.4.2). The fist condition will usually be met, so much depends on the extent to which explanations are demanded. As no objective criteria have been established—only an explanation is required—this does not appear to be much of a constraint. However, given that before the Uruguay Round, US investigating authorities consistently refused to use average‐to‐average comparisons, the requirement under the WTO to do so—even though subject to loopholes—was an improvement (Palmeter, 1995).

Procedural biases and methodological abuses are very difficult, if not impossible to regulate away given the definition of dumping. For example, the requirement that AD duties be terminated within five years would appear to be a major improvement from an economic welfare point of view. In practice this may not be a binding constraint, however, as it is conditional upon whether a review investigation finds that dumping and injury continues (or threatens) to persist. Another example pertains to the definition of an interested party in AD cases. This provides users and final consumers of the import a voice during the investigations, but restricts them to providing evidence that is relevant to the determination of dumping, or injury to domestic firms that compete with the imported product. The fact that a duty or undertaking may injure their proper business is not a factor that can be brought forward. The agreement also does not require any consideration of the economy‐wide impact of AD duties, the state of competition in the domestic market, let alone an investigation into the market access conditions prevailing in the exporter's home market.

Although the WTO's general dispute settlement mechanism applies, Article 17.6 of the AD Agreement restricts the ability of panels to focus on the substance of a case, as they are required to accept any ‘reasonable interpretation’ of the facts put before an AD domestic authority. In cases where the Agreement can be interpreted in more than one way, a decision by investigating AD authorities must be accepted if it is based upon one of the permissible interpretations. New information that was not available or used by investigating authorities may not be used by a panel to overturn an AD action (Finger, 1994). In many cases panels will be limited to determining whether (p.328) the procedural requirements of the Agreement were violated. As noted by Palmeter (1995), a major goal of US user industries in the Uruguay Round was to limit the ability of GATT panels to overturn domestic AD decisions. The standard of review embodied in the Agreement reflects the power of the industries supporting AD. This lobby was strong enough to make this specific issue a deal breaker for the US. It obtained most of what it sought.

Little was done in the Uruguay Round negotiations to discipline AD because the talks were essentially conducted between the users of antidumping measures on the one hand, and the countries that pursued export‐oriented development strategies on the other. This meant that the negotiations differed substantially from those in the Tokyo and the Kennedy Rounds, where the negotiating process mainly involved user countries. The user‐exporter dichotomy made it difficult to come up with a balanced package deal in the Uruguay Round. Exporting country governments had little to offer in the negotiations. As dumping is a private practice, governments cannot and should not prohibit it. Countries such as Japan, Hong Kong, Korea and Singapore considered that their export trade was detrimentally affected by AD measures, and made many proposals to discipline the use of methodologies that were biased toward finding high dumping and injury margins. In this they had some success, in that practices that were tolerated but not explicitly subject to GATT disciplines became subject to multilateral rules specifying the conditions under which they could be used.

A number of contentious practices could not be resolved in this manner. A good example pertains to so‐called anticircumvention measures, which Japan and Korea sought to subject to multilateral rules. Exporting firms may try to circumvent AD actions by establishing assembly plants either in the importing country (where the final product has become subject to AD duties) or in third countries. Anticircumvention became an important issue in AD enforcement as of the late 1980s. In June 1987, the EU adopted legislation allowing measures to be imposed to prevent circumvention of antidumping measures on finished products. Such measures could be applied to products assembled or produced in the EU, using imported materials or parts. In the year following adoption of this Regulation, the EU initiated investigations on electronic typewriters, electronic scales, excavators, and photocopiers. All of these products were assembled or produced by Japanese‐related companies in the EU.

Japan challenged the anticircumvention measures before the GATT, arguing that the existence of dumping and injury related to imports of components was not investigated. It also held that the provision contained GATT‐inconsistent local content requirements as it stipulated that duties could be imposed if the value of components originating in the country (p.329) subject to the initial AD duty exceeded the value of all other parts by a specified margin. Finally, Japan noted that the duties were imposed only on manufacturers associated with foreign companies that were already subject to AD duties. Domestic producers were not affected even if they used the same imported components (discrimination). Australia, Hong Kong and Singapore made detailed submissions critical of the EU circumvention regulation. The EU argued that the anticircumvention provision was adopted after experience had shown that the initiation of AD action was frequently followed by the establishment of assembly operations in the EU motivated by a desire to circumvent AD duties. The US supported the EU's objective of combating circumvention of AD duties, it having similar concerns.

The GATT panel that considered the case concluded that the anticircumvention duties on the finished products, being levied on products manufactured within the EU, were not customs duties but internal taxes. Because these were levied on a discriminatory basis, they were inconsistent with national treatment. The EU was requested to bring the application of its anti‐circumvention mechanism into conformity with GATT obligations. Although Japan won this battle, as adopted panel reports become part of the GATT case law, it did not win the war. No agreement emerged on anticircumvention in the Uruguay Round, as negotiators could not agree on a specific text. The matter was referred back to the Committee on Antidumping Practices.

Numerous disputes have been brought to the WTO regarding the imposition of AD measures. As the agreement's procedural rules are complex and technical, investigating authorities, especially those in developing countries, may find it difficult to jump through all the hoops that are established by the WTO. One can speculate that to some extent the strategy that was pursued in the Uruguay Round was to make it more costly for industry and government agencies to undertake AD actions. Such costs impact disproportionately on developing countries, both as respondents and as users of AD. Space constraints prohibit a detailed discussion of the various dispute settlement cases in the area of AD. Interested readers are referred to the home page of the WTO and the sources mentioned in section 9.10 below.

To recap, dumping is rarely an anticompetitive practice. Predatory pricing is possible, but will not be profitable as long as governments ensure that markets remain contestable. At the same time, AD creates a large number of distortions. The existence of AD induces rent‐seeking behavior on the part of import‐competing firms, and leads exporting firms to alter production, allocation, and production‐location decisions in ways that can easily reduce welfare at home and abroad. The threat effects of AD are important and insufficiently recognized. AD can imply substantial uncertainty regarding the conditions of market access facing exporters and increase the costs of goods (p.330) for importers. The chilling effect on imports of AD threats can be great. The start of an investigation is a signal to importers to diversify away from targeted suppliers. This signal is strengthened once provisional findings have been issued. These create incentives for the conclusion of agreements with affected exporters. The use of AD to arrive at VER‐type undertakings is particularly troublesome. Threat effects can also arise under Article XIX‐type safeguard protection, but these will generally be less distorting. Safeguards are more transparent, nondiscriminatory, less arbitrary, and less prone to capture. AD mechanisms are an option allowed under the WTO; they are not required. The best option for governments concerned with equity and efficiency is not to pass AD legislation, and to abolish it if it exists. Safeguards are a much better and more honest instrument to address the problem AD is used for—providing import‐competing industries with time to adjust to increased foreign competition (Finger, 1999).

Given that current AD procedures make no attempt to determine whether markets are uncontestable, one way to reduce the protectionist bias that is inherent in the status quo is for governments to put greater effort into determining whether the conditions alleged to give rise to ‘unfair trade’ actually exist. Suggestions that have been made in this regard include making antidumping conditional on a determination that the exporters' home market is not contestable, and shifting away from an ‘injury to competitors’ standard towards an ‘injury to competition standard’ (Hoekman and Mavroidis, 1996; Messerlin, 2000). Greater efforts should also be made to consider the economy‐wide effects of taking action by giving users the legal standing to defend their interests. The basic problem is a political economy one: there are powerful vested interests that are in favor of AD. This suggests that a necessary condition for AD reform is greater mobilization of counter‐vailing forces in the domestic political arena. In addition to users, one group that may see its incentives to push for AD reforms increase is the exporter community. The more that developing countries start to use AD against high‐income country exporters, the less inclined such firms may be to accept the ongoing geographic spread of this instrument.

9.5. Measures to Countervail Subsidized Imports

Countervailing duty (CVD) or antisubsidy procedures, similar to AD, are allowed under Article VI of the GATT. The objective is again to ensure ‘fair competition’. As noted in chapter 5, while the GATT prohibits export subsidies on manufactures, it permits other types of subsidies. However, unless a subsidy is in the ‘green box’ (the set of permitted and nonactionable measures), (p.331) governments that consider that their industries are materially injured by imports benefiting from subsidies may impose countervailing measures. A necessary condition is that an investigation determines that imports have been subsidized and have caused material injury to domestic industry. The procedures to be followed in subsidy and injury investigations are described in detail in the Agreement on Subsidies and Countervailing Measures, which in turn is largely based on the Tokyo Round Subsidies Code. As in the case of AD disciplines, the injury test is the key element underlying the agreement's implicit objective of reconciling legitimate national government subsidy policies with the interests of nations affected by those policies. That is, the focus is on dealing with the externality created by foreign government policies.

The US has been the largest user of CVD mechanisms, initiating over 100 investigations since 1985. A substantial number of these CVD cases were not brought to a conclusion, reflecting the introduction of trade‐restricting bilateral agreements, as in the cases of many steel products. Other major players such as the EU and Japan have made little use of the CVD mechanism. In part this is because other policies for safeguarding producers' interests are available and easier to implement (AD), and in part because many countries fear that initiation of CVD investigations could lead to retaliatory investigations. As the US does not devote many resources to explicit subsidization of manufacturing, its government has never felt constrained in using CVDs. The use of CVDs has been relatively constant over time. It dipped in the mid‐1990s, but picked up noticeably in 1998–9 (Table 9.5).

Table 9.5 Countervailing Duties (1992–9)

Total, 1992–94






Total, 1995–99

OECD countries








Developing countries
















Note: OECD includes Mexico and Korea; annual data spans July 1 to June 30.

Source: WTO, Committee Annual Reports.

One reason for the increase was a more active stance on the part of the EU (which initiated no CVD actions during 1990–95) and New Zealand. The EU, Brazil and South Africa are among the main targets of CVDs. One would have expected a decline in the incidence of CVDs since the Uruguay Round was concluded, given the six‐year peace clause contained in the Agriculture Agreement and the greater clarity that was achieved in the Subsidies Agreement regarding what types of subsidies are not countervailable. The US reports (p.332) that cases initiated during 1988–97 affected only 0.7 percent of total imports, a decline by 50 percent compared to the 1984–94 period, when 1.4 percent of imports were affected (Finger and Schuknecht, 2000). Available data suggest that CVDs are lower than AD duties. In the case of the US, CVDs averaged less than five percent in 80 percent of cases launched between 1995 and 1999.

Political Economy of Countervailing Duties

There are two possible rationales for responding to foreign subsidy policies via import restrictions. The first is to offset the injurious effect of such policies on domestic industries. The second is to induce the foreign government to change its policy. The first rationale has little economic merit, as the imposition of import barriers (CVDs) distorts the decisions of consumers, and will generally reduce the welfare of the country taking action. A subsidy granted to a foreign firm will generally only be one aspect of the industrial policy that is applied by a foreign government. Such governments may also pursue direct and indirect tax policies, engage in investments in infrastructure, and so forth. Given the difficulty of determining the real (general equilibrium) effect of any kind of foreign industrial policy, it will always be very difficult, if not impossible, to determine the appropriate counter action.

The argument that restricting imports of products that have benefited from unfair government assistance can be justified as a means of inhibiting the use of such measures has greater economic merit (Deardorff and Stern, 1987). It may be the case that even though a CVD is welfare reducing in the short run, the threat of CVDs induces foreign governments to refrain from subsidizing. The relevance of this argument depends on whether the cost to the foreign country of a CVD is greater than the benefit it realizes from the subsidy policy. This may not be the case, especially if the policy aims to offset a market failure or is driven by noneconomic considerations. As important, even if the subsidy cannot be justified by market failure or public good considerations, the effectiveness of CVDs depend on the ability of the country imposing them to affect the terms of trade of the country granting the subsidy. Small countries are unlikely to be able to have an impact on the subsidizer's exporters that are large enough to induce them to pressure their governments not to impose the subsidy. A better option is to engage in discussions with the subsidizing country and seek compensation. This is an option that in principle is available under the WTO (see below).

CVDs are superior to AD in that the instrument is better targeted at the source of the perceived externality: foreign government intervention. In contrast to the case of dumping by firms, in the subsidy context it is possible (p.333) to build a case for ‘unfair competition’. However, for most countries countervailing subsidies will rarely make economic sense, unless the subsidy is expected to be temporary. In the case of agricultural subsidization by OECD countries, for example, the policy could be regarded as structural, in that affected producers are well advised to adjust to the situation. Imposition of CVDs is equivalent to throwing good money after bad. After all, the subsidy is equivalent to a transfer from a foreign government to the consumers of the importing country. If a tariff is imposed on these imports, the economy will be worse off than if nothing is done. A CVD may benefit the domestic industry, but is equivalent to a tax on the rest of the economy. Consumers lose the benefit of the foreign subsidy as prices are forced up, and at best the country ends up with an additional deadweight loss. However, this loss is smaller than is the case under a regular tariff, as there is no production distortion (see annex 2).

WTO Disciplines

The GATT history on CVDs revolves around the use made by the US of this instrument, it being the primary user. When the US acceded to GATT, it grandfathered its existing CVD legislation. The negotiation of a Code on Subsidies and Countervailing Measures in the Tokyo Round was driven by a desire by targeted countries to see the US adopt an injury test (which was not required under its law). This attempt was somewhat successful, in that the US signed an agreement that required an injury test. However, in practice, the US made this conditional upon bilateral commitments with respect to subsidy policies. US CVD policies therefore continued to be a source of controversy (Stern and Hoekman, 1987).

In the Uruguay Round the issue of subsidies and CVDs was substantially clarified. The WTO makes a distinction between different types of subsidies, depending on their trade impact and their objective (see chapter 5). Subsidies that have an economy‐wide impact and are not specific (education, general infrastructure, basic R&D) or have a noneconomic rationale (regional disparities, income support) are permitted and not subject to the threat of CVDs (nonactionable). For subsidies that are not in this ‘green box’ and that have adverse effects on trade, WTO members have the choice of initiating CVD investigations or invoking dispute settlement procedures. Both routes may be pursued simultaneously. However, only one remedy may be applied.

As noted in chapter 5, adverse effects include injury to a domestic industry, nullification or impairment of tariff concessions, or serious prejudice or threat thereof to the country's interests. Serious prejudice is deemed to exist if the total ad valorem subsidization of a product exceeds five percent, subsidies (p.334) cover operating losses of an industry or enterprise, or there is forgiveness of government‐held debt. Serious prejudice may arise if the subsidy displaces imports of like products on either the subsidizing or third country markets. If actionable subsidies have an adverse effect on a WTO member, it may ask for consultations with the country maintaining the subsidy program. If consultations fail to settle matter within 60 days, the WTO's dispute settlement provisions may be invoked. If a panel deems adverse effects to exist, the subsidy program must be revoked, or the affected country otherwise compensated. If the panel's recommendations are not implemented within six months, the affected country can retaliate by withdrawing equivalent concessions.

Necessary conditions for imposition of CVDs include demonstration of the existence of a subsidy, a finding that a domestic industry producing similar (like) products is materially injured, and a causal link between the subsidization and injury. Injury requires that the volume of subsidized imports has increased, that this has had an impact on price levels or is reflected in price undercutting of domestic firms and that this in turn has had a detrimental effect on the domestic industry. At least 25 percent of the firms in the domestic industry must support the launching of a CVD investigation. Recall that the export subsidy dispute between Brazil and Canada revealed that in the case of an illegal (prohibited) subsidy these rules do not apply. Canada was authorized to retaliate (countervail) up to the amount of the subsidy (as determined by the arbitrators), independently of the injury suffered by the Canadian industry (see chapter 5).

Detailed requirements and deadlines are established regarding the different phases of investigations, including the collection of evidence, the rights of interested parties, the calculation of the extent to which a subsidy benefits the recipient, the determination of injury, possible remedies, and access to judicial review of the CVD decision. As is true for AD, a sunset provision of five years applies, unless a review determines that the abolition of protection would likely lead to the continuation or recurrence of injury. When confronted with CVD investigations, developing countries benefit from de minimis thresholds. If the subsidy is less than two percent of the per unit value of products exported, developing countries are exempt from countervail (for LDCs the threshold is three percent). An exemption also applies if the import market share of a developing country is below four percent, and the aggregate share of all such countries is below nine percent of total imports.

As noted previously, CVDs are used much less frequently than AD, and average tariffs imposed are relatively low. Disputes have arisen on the use of (p.335) subsidies, some of them very high‐profile cases—see chapter 3—but these have not extended to the use of CVDs.

9.6. Trade Restrictions for Balance of Payments Purposes

The GATT permits the imposition of trade restrictions to safeguard a country's external financial position (Articles XII and XVIII:b). The inclusion of such provisions reflects the fact that when the GATT was created, a system of fixed exchange rates prevailed (the so‐called Bretton Woods system). Fixed exchange rates remove an instrument through which governments can seek to address balance of payments (BOP) disequilibria. If a country with a deficit cannot devalue and if wages are relatively inflexible as well, there is a case for imposing temporary import restrictions. An across‐the‐board tariff in conjunction with a subsidy to exports is under certain conditions exactly equivalent to a nominal devaluation of the currency (Vousden, 1990).

The BOP provisions of the GATT were widely used by both industrialized and developing countries. The former used Article XII as cover, the latter Article XVIII:b. Use of QRs for BOP purposes by developed, mostly European, countries occurred mostly in the 1950s when many currencies were not convertible. In contrast, developing country use of Article XVIII:b was fairly constant until the 1990s (Table 9.1). Most developing countries had no need to use Article XIX safeguard actions to protect their industries if needed, as their tariffs were generally not bound or bound at high levels. They did need GATT cover to use QRs, however, and this was the role of Article XVIII—it permits, indeed encourages, the use of QRs for BOP purposes. This was an idiosyncrasy of the GATT 1947, and contradicted the general preference for price‐based instruments such as tariffs. In general, an import surcharge would be less distortionary than QRs (see annex 2).

The challenge to OECD negotiators in the Uruguay Round was to close the BOP loophole, which in practice was simply an avenue to legally impose QRs, albeit subject to GATT surveillance. Closing the loophole should have been facilitated by the move away from the fixed exchange rate system that had occurred in the 1970s. The move towards flexible (more easily adjustable) exchange rates reduces the rationale for resorting to trade restrictions to safeguard a country's external financial position. Exchange rate adjustment provides an automatic and effective mechanism for adjustment of current account imbalances if complemented by supporting measures (fiscal and monetary discipline). Experience clearly demonstrates that QRs are not the right instrument to deal with BOP problems. The IMF and World Bank routinely obtain agreements with borrowing governments not to introduce (p.336) import restrictions for BOP purposes in their adjustment lending to developing countries (Finch and Michalopoulos, 1988).

Most developing countries responded to this line of argument by emphasizing that their foreign exchange shortages did not stem so much from their own policies as from protectionist policies of their trading partners. Although this was a disingenuous argument at best, given that overvalued exchange rates were usually a major cause of foreign exchange shortages and rationing, it was against this background that the issue of BOP escape clauses was considered in the Tokyo Round. A 1979 Declaration on Trade Measures Taken for Balance‐of‐Payments Purposes reinforced scrutiny over the trade and adjustment policies of industrial countries, but asserted that developing countries should be allowed greater latitude in safeguarding their foreign exchange reserves. During the 1980s, industrialized countries argued increasingly that the clause rendered the participation of developing countries in GATT to a large extent meaningless. This damaged both the trading system (because it undermined adherence to the principles on which the system rests) and developing countries, as the latter had no effective means within the GATT context to counter powerful protectionist interests at home. The measures imposed tended to be permanent, whereas BOP difficulties are mainly of a cyclical nature. The trade restrictions were also often imposed on selected products, rather than being applied across the board as would be necessary for BOP purposes.

WTO Disciplines

In the Uruguay Round, new language on Article XVIII was agreed that reduced the scope to use QRs, and strengthened surveillance of BOP actions. GATT 1947 contracting parties committed themselves to publicly announce time‐schedules for the removal of restrictive import measures taken for BOP purposes. They also agreed to give preference to those measures that have the least disruptive effect on trade. Such measures include import surcharges, import deposit requirements or other equivalent trade measures with an impact on the price of imported goods. The use of new QRs for BOP purposes requires a justification why price‐based measures cannot arrest the deterioration in the external accounts. Only one type of restrictive import measure may be applied on a product. The emphasis on the use of price‐based measures is a significant improvement over the old GATT.

Surcharges or similar measures must be applied on an across‐the‐board basis. However, exemptions may be made for certain essential products, necessary to meet basic consumption needs or which help improve the BOP situation, such as capital goods or inputs needed for production. A WTO (p.337) member applying new restrictions or raising the general level of its existing restrictions must consult with the BOP Committee within four months of the adoption of such measures. Each year a member taking BOP actions must provide the WTO secretariat with a consolidated notification providing information at the tariff line level on the type of measures applied, the criteria used for their administration, product coverage and trade flows affected. Countries applying BOP measures must engage in periodic consultations with the BOP Committee. The report prepared for such meetings must include an overview of the BOP situation and the policy measures that have been taken to restore equilibrium, a description of the restrictions that are applied, progress towards removing the restrictions, and a plan for the elimination and progressive relaxation of remaining barriers. The WTO secretariat also prepares a report, using data obtained from the IMF, regarding the macroeconomic situation in the country concerned. The IMF is represented at all BOP Committee meetings.

There is therefore in principle rather close surveillance of BOP actions. Very much depends, however, on the willingness of the BOP Committee to insist that measures are no longer or not justified. Under the GATT 1947 not much could be expected from this Committee (Eglin, 1987). In practice the main source of discipline came from the international capital markets, from international financial institutions and from bilateral pressure by WTO members to stop invoking Article XVIII as cover for trade restrictions. A total of ten developing countries revoked Article XVIII during the 1980s and early 1990s, largely following the adoption of more appropriate macroeconomic policies and unilateral liberalization efforts. Multilateral surveillance exercised by the BOP Committee played only a minimal role a role in this. The GATT dispute settlement system had more teeth. For example, Korea's use of Article XVIII:b was challenged in the 1980s by beef exporters, who alleged that Korea no longer had a BOP problem and that restrictions on beef could therefore not be justified by this Article, as claimed by Korea. The panel that dealt with this case found in favor of the petitioners and recommended that Korea be required to eliminate its import restrictions on beef.

Disciplines under the WTO are more binding. The difference between GATT and the WTO was illustrated by a case brought against India in 1997 by the United States. The US claimed that QRs maintained by India—a long‐standing user of Article XVIII:b—on importation of a large number of products, covering more than 2,700 agricultural and industrial product tariff lines, were inconsistent with GATT Articles XI:1 and XVIII:11 as well as other provisions of the WTO. The panel that considered the allegations found that the measures violated India's WTO obligations and nullified or impaired benefits accruing to the US under the Agreement on Agriculture. The Appellate (p.338) Body upheld the report, and the panel and Appellate Body reports were adopted in September 1999. Noteworthy is that the Appellate Body rejected India's argument that the panel had no jurisdiction given that the Balance of Payments committee had not pronounced on the matter.

India stated its intention to comply with the recommendations and rulings of the DSB, and drew attention to the panel's suggestion that the reasonable period of time for implementation in this case could exceed 15 months in view of India's status as a developing country. After consultations, the US and India agreed most changes would be made by April 2000, with the remainder to be implemented by April 2001. This case would not have been possible under GATT. It signals the end of decades of invocation by India of Article XVIII as a cover for QRs.

9.7. Infant Industry Protection

Article XVIII:a GATT allows for the removal of tariff concessions if necessary to establish an industry in a developing country. It does not differ much in substance from Article XXVIII (re‐negotiation of tariffs), as compensation negotiations must be initiated. Article XVIII:c permits the use of QRs or other nontariff measures by developing countries for infant industry purposes. This provision requires the approval of WTO members, and compensation may also be requested. At the end of the Tokyo Round, the GATT infant industry exception was widened considerably to allow for measures intended to develop, modify or extend production structures more generally, in accordance with a country's economic development priorities. These exceptions have rarely been invoked as a cover for the use of import quotas, probably due to the fact that the BOP loophole embodied in Article XVIII:b was preferred. In comparison to the latter, surveillance and approval procedures under the infant industry provisions are more strict, and the possibility of retaliation more likely.

In most circumstances the economic rationale for invocation of Article XVIII:c is weak, as a QR in itself will do very little to stimulate the establishment of a competitive industry. Any justification for a government to help in the establishment of an industry must be based on market failure. Even if it is assumed that a government can correctly identify the market failure, a QR will never be an appropriate instrument to offset the source of the distortion. Usually a subsidy of some type will be a less inefficient instrument to promote the establishment of an industry. From an economic viewpoint, the drafters of the GATT were therefore justified in placing relatively stringent conditions on the use of infant industry protection. But, as has been the case (p.339) with other GATT disciplines as well, the result of this was to induce a shift towards invoking substitute GATT‐cover: the BOP route and TRIMs (see chapter 5). Given that a number of countries have used TRIMs for industrial development purposes, as the TRIMs agreement becomes a binding constraint on developing countries, an increase in the invocation of Article XVIII may occur for infant industry purposes. However, enough scope exists to intervene in ways that make more economic sense, through use of subsidies, public investment and the use of safeguard protection if deemed necessary.

9.8. General Exceptions

Both the GATT and the GATS contain provisions entitled ‘General Exceptions’ allowing members to take measures that violate a rule or discipline if necessary to achieve noneconomic objectives (GATT Articles XX; GATS Article XIV). Such objectives include protection of public morals (XX:a) and the health and safety of human, plant or animal life (XX:b). GATT Article XX also allows controls to prevent imports of goods produced with prison labor (XX:e), to conserve natural exhaustible resources (as long as the same measures are applied to domestic production or consumption as well—XX:g), to protect cultural heritage (XX:f), and to control exports of goods in short supply or subject to public intervention (XX:i and j). Both GATT and GATS allow for measures needed to secure compliance with laws or regulations that are not inconsistent with multilateral rules (examples mentioned include prevention of deceptive and fraudulent practices, and protection of privacy of individuals). A necessary condition for the invocation of the exception provisions of the WTO is that measures do not result in ‘arbitrary or unjustifiable’ discrimination between countries, and are not a disguised restriction on international trade. In addition to exceptions that focus on the attainment of noneconomic objectives, all three multilateral trade agreements have a national security exception (GATT Article XXI, GATS Article XIVbis and TRIPs Article 73).

The general exceptions articles are purposely worded in rather broad and vague terms. There are no compensation or approval requirements. There is also no notification requirement—it is up to affected parties to raise a measure they perceive to be discriminatory and detrimental to their interests with the member applying them. If that member defends the measure under the exceptions provisions of the WTO, the only recourse is the dispute settlement mechanism. Many panel cases have involved an investigation whether the provisions of Article XX are applicable in specific instances. In (p.340) many such cases it is the task of the panel to decide if the measure in question is necessary to achieve the government's purported objective. This ‘necessity test’ is the main discipline on invocation of the exception provisions. For example, in a case concerning a policy imposed in Thailand prohibiting imports of cigarettes, the panel found that this violated Article XI (prohibition on QRs). An argument by the Thai government that the import ban was justified under Article XX as it was necessary to control smoking was rejected, because other instruments were available to restrict imports (so the QR was not necessary). Moreover, and more importantly, the ban was inconsistent with achieving the government's stated goal, as domestic production was unconstrained, thus violating national treatment.

Pressure has increased substantially on WTO members to extend and clarify the limits of the exceptions provisions of the WTO. These pressures are coming from two directions. On one side, various interest groups in OECD countries have sought to induce their governments to invoke the provisions as cover for the imposition of trade barriers motivated by environmental or social concerns. On the other side, exporting countries are concerned that the exceptions articles do not become loopholes importing country governments can use to argue that a particular trade restriction is ‘necessary’ to achieve a noneconomic objective. Increasing attention is being given by scholars and the trade policy community to alternative approaches that can be considered to ensure that the exceptions articles are not abused. Clearly a critical issue is the application of the criterion whether a measure that discriminates against imported products is necessary to achieve the public policy objective. It has been argued that a useful complement to the ‘necessity test’ would be to introduce an economic rationality test: is the measure the least cost way of attaining the objective (Mattoo and Subramanian, 1998)? A key question here is who will determine whether measures are necessary, WTO members or the WTO judiciary (panels and the Appellate Body)? These issues are discussed further in chapter 13, as they are among the major challenges confronting the trading system.

National Security

The national security exemptions are particularly ill‐suited for dispute settlement, as in such cases panels would have to judge whether trade restrictions are necessary to protect national security. This can obviously be a very sensitive issue, especially as the language of the national security exceptions are particularly vague. Article XXI GATT allows measures to be imposed whenever a government considers this ‘necessary for the protection of its essential security interests’ both in time of war or ‘other emergency in international (p.341) relations’. Sometimes efforts to invoke a national security rationale are blatantly spurious. When Sweden imposed import quotas on footwear in 1975 it argued that this was motivated by ‘national security’ concerns because it needed to have a domestic industry to guarantee the country would not be short of army boots in time of war. This argument did not go over well with Sweden's trading partners.

Not surprisingly, GATT was rarely used as an instrument to contest economic sanctions imposed for foreign policy reasons. A US embargo against Nicaragua in 1985 was contested by the Nicaraguan government, but the panel ruled that it did not violate the GATT. It also noted that it was precluded from judging the validity of—or the motivation for—the US action (Jackson, 1997). More recently, a dispute between the EU and the US has tested the limits of the national security exception (Box 9.1).

9.9. Conclusion

Political realities, especially in countries in the process of moving from highly distorted trade regimes to a more neutral policy stance, often dictate (p.342) that there be mechanisms allowing for the temporary reimposition of protection in instances where competition from imports proves to be too fierce. Safeguard mechanisms are therefore likely to be a precondition for far‐reaching liberalization to be politically feasible. Governments need loop‐holes that permit ‘backsliding’ for a variety of reasons (Finger, 1996). One that has not been discussed so far is a sympathy motive. Societies tend to have sympathy for groups severely affected by large, exogenous shocks. They support granting assistance to such groups because they too may be affected some day. This insurance motive is complemented by what Corden (1974) has called the conservative social welfare function. Governments tend to oppose large absolute reductions in real incomes of any significant portion of society.

Trade policy is generally an inefficient instrument in that it tends to create more distortions than it solves. Indeed, Deardorff and Stern (1987) have likened trade policy to doing acupuncture with a two‐pronged fork; even if one of the prongs finds the right spot, the other prong can only do harm. This applies to protection in response to market disruption as well. Protection is a very costly form of intervention, both in a static sense (as demonstrated by studies of costs per job saved such as Hufbauer and Elliot, 1994 and Messerlin, 2001), and in a dynamic sense (due to the distortions that reduce economic growth). In practical terms, however, given a sociopolitical need to address market disturbance, temporary contingent protection may be the best response in situations where import penetration has increased substantially. The issue then is to design and implement procedures that are effective, equitable, and minimize distortions. The WTO does little in terms of providing guidance to policymakers wishing to rationalize or create an economically sound system to deal with market disruption caused by imports. Allowing for the possibility of emergency protection sends a signal to firms that the government cannot or will not commit itself to a given level of intervention or support. This can negatively influence the performance of particular firms—who may build this insurance into their management decisions. This can in turn give rise to so‐called time‐inconsistency problems. If a government is pursuing a liberalization program, but firms do not adjust because they expect to be able to obtain protection in the future, it may not be optimal (politically) for the government not to grant such protection (or alternatively, to remove the temporary, emergency protection). The design of the mechanism and the rules and criteria that apply are therefore important. External obligations—such as those applying under the WTO—can help in reducing possible time inconsistency problems, but cannot eliminate them.

(p.343) In practice the WTO does a bad job by allowing for too many escape valves, some of which make no economic sense. This has given rise to a form of Gresham's law, in that bad provisions (such as AD) have driven out good provisions (safeguards). Countries in the process of developing or reforming their trade laws are well advised not to implement all the options allowed under GATT to impose trade barriers, as it will make it much more difficult to control the trade policy formation process. All that is required is a good safeguards mechanism—AD and BOP actions are best avoided. Countries already caught in the contingent protection morass are well advised to ensure the economic costs of AD are identified and considered. Providing groups that will lose from protection the legal mandate to make the case against intervention is a key requirement for ensuring balance.

Finally, it should be kept in mind that provisions such as Articles XIX and XXVIII GATT (or Article XXI GATS) are only relevant if tariffs have been bound (specific commitments have been made). If this is not the case, countries will have the latitude to simply raise tariffs if the political need for this arises. For the GATT rules to fully bite, tariffs must be bound. For GATS to bind, specific commitments must be made. As discussed in chapters 5 and 7, on both the goods and services front much still remains to be done to achieve full binding of tariffs at applied rates and schedule all service sectors.

9.10. Further Reading

Safeguard protection for import‐competing industries has been analyzed extensively in the economic literature. Robert Baldwin, ‘Assessing the Fair Trade and Safeguards Laws in Terms of Modern Trade and Political Economy Analysis’, The World Economy, 15 (1992), 185–202 discusses the (political economy) issues and surveys some of the literature. Gary Sampson, ‘Safeguards’, in J.M. Finger and A. Olechowski (eds.), The Uruguay Round: A Handbook (Washington DC: The World Bank, 1987) reviews the history of Article XIX in the GATT through 1986. Alan Deardorff, ‘Safeguards Policy and the Conservative Social Welfare Function’, in Henryk Kierzkowski (ed.), Protection and Competition in International Trade (Oxford: Basil Blackwell, 1987) discusses why governments need safeguard instruments. Brian Hindley, ‘GATT Safeguards and Voluntary Export Restraints: What are the Interests of the Developing Countries?’ World Bank Economic Review, 1 (1987), 689–705 discusses the incentive effects of VERs. A classic paper on VERs is Richard Harris, ‘Why Voluntary Export Restraints are “Voluntary’ ”, Canadian Journal of Economics 18 (1985), 799–809. Patrick Low, Trading Free: The (p.344) GATT and US Trade Policy (New York: Twentieth Century Fund, 1993) discusses is some detail the political economy of the US shift towards the use of VERs, as well as the evolution of US trade policy thinking and practice.

Richard Eglin, ‘Surveillance of Balance‐of‐Payments Measures in the GATT’, The World Economy, 10 (1987), 1–26 reviews the GATT experience with Article XVIII actions and multilateral surveillance. The contributions by Finger, Hindley and L. Alan Winters in The New World Trading System (Paris: OECD, 1994) are good summaries of what was agreed on contingent protection in the Uruguay Round. J. Michael Finger, ‘Legalized Backsliding: Safeguard Provisions in the GATT’, in Will Martin and Alan Winters (eds.), The Uruguay Round and the Developing Economies (Cambridge: Cambridge University Press, 1996) is a comprehensive discussion of the various loop‐holes in the GATT that allow for backsliding. Michael Finger and Ludger Schuknecht, ‘Market Access Advances and Retreats: The Uruguay Round and Beyond’, in Bernard Hoekman and Will Martin (eds.), Developing Countries and the WTO (Oxford: Basil Blackwell, 2000) summarizes the first five years of WTO members' use of safeguards and exceptions.

There is a huge literature on antidumping, both legal and economic. Much of the recent economic work is collected and summarized in Robert Lawrence (ed.), Brookings Trade Forum (Washington DC: The Brookings Institution, 1998). J. Michael Finger (ed.), Antidumping: How it Works and Who Gets Hurt (Ann Arbor: University of Michigan Press, 1993) is an excellent set of papers identifying why AD makes no economic sense. Patrick Messerlin, ‘Antidumping Regulations or Pro‐cartel Law? The EC Chemical Cases’, The World Economy, 13 (1990), 465–92 illustrates on the basis of a particular case how an industry can capture AD procedures to enhance its market power. P.K.M. Tharakan, ‘The Political Economy of Price‐Undertakings’, European Economic Review 35 (1991), 1341–59 analyses the use of and motivations for undertakings in EU application of AD. Thomas Prusa addresses the same topic from a US perspective in ‘Why Are So Many Antidumping Petitions Withdrawn?’, Journal of International Economics 33 (1992), 1–20. Patrick Messerlin, ‘Antidumping and Safeguards’, in Jeffrey Schott (ed.), The WTO After Seattle (Washington DC: Institute for International Economics, 2000), and Bernard Hoekman and Petros C. Mavroidis, ‘Dumping, Antidumping and Antitrust’, Journal of World Trade 30 (1996), 27–52 discuss possible reforms that could be considered in the WTO regarding AD. A database on US AD activity for all cases from 1980–95 is available at: http://darkwing.uoregon.edu/∼bruceb/adpage.html. This includes Excel spreadsheets, documentation, and links to government agencies involved in AD enforcement, researchers and online working papers. For a comprehensive, insider account of the AD Agreement, see Mark Koulen, ‘The (p.345) New Antidumping Code Through its Negotiating History’, in J. Bourgeois, F. Berrod and E. Fouvier (eds.), The Uruguay Round Results: A European Lawyer's Perspective (Brussels: European Interuniversity Press, 1995).

A good source of policy‐oriented papers on contingent protection and related issues is The Journal of World Trade (Kluwer).


(1) Principal supplying interest is determined with reference to the share in the export market; substantial supplying interest is determined in relation to a country's total volume of exports.

(2) Costa Rica vs. United States: Restrictions on Imports of Cotton and Man‐Made Fibre Underwear, WT/DS24/AB/R (10 February 1997); India vs. United States: Measures Affecting Imports of Shirts and Blouses, WT/DS33/AB/R (25 April 1997).

(3) The price elasticity of demand for a product is defined as the ratio of the percentage change in the quantity demanded to the percentage change in the price. For example, if prices increase by five percent and the corresponding quantity decreases by more than five percent, demand is price elastic.

(4) Research by Feinberg and Kaplan (1992) provides evidence of cascading. In a statistical analysis of all antidumping and countervailing duty cases during the period 1980–86 brought by US producers and users of metals, they found that user industries tended to file for protection after upstream industries, and that the share of all cases accounted for by downstream industries increased significantly over time. Analysis of cases involving producers and users of chemicals led to the same finding.

(5) Official Journal, L 41/54, 12 February 1994.