a Percentage of 1989 imports.

Source: Rectifications and Modifications of Uruguay Round Schedules

Chart T1: Coverage of GATT Tariff Bindings, Pre- and Post-Uruguay Round
(Industrial Goods)

Chart T2: Depth of Tariff Cuts Agreed at the Uruguay Round
(Industrial Goods)

The Uruguay Round re-emphasized GATT�s concern to discipline the use of quantitative restrictions and similar instruments, and made considerable progress toward that end. Rules that allow such restrictions ¹¹ have been tightened, and those outside the new rules must be phased out. Grey-area measures such as voluntary export restraints are now illegal and existing measures must be phased out. In this section we will review notifications to the WTO of legal quantitative restrictions, and of those made illegal by the Uruguay Round Agreements and scheduled for elimination. In later sections we will review restrictions that have been imposed under various of the safeguard provisions in the Uruguay Round Agreements - some of which provisions allow for quantitative restrictions.

Several GATT provisions that allow quantitative restrictions do not include notification requirements. Principal among these are Articles XX, General Exceptions, and XXI, Security Exceptions. Article XX covers, inter alia, restrictions to protect public morals (e.g., prohibition of importation of pornography) to apply technical or sanitary standards, to preserve human, animal and plant life, restrictions against imports produced by prison labor.

Some countries notify such restrictions, others do not. Table QR1 provides indicative information about the mix of legal quantitative restrictions that Members apply, but it does not provide reliable information about which Members apply legal quantitative restrictions.¹² Virtually all countries have controls on some imports, e.g., of arms. The WTO Agreements do not require that Members report these restrictions and Table QR1 indicates that few Members have chosen to report them.

In Table QR1 we see that 49 times out of 57 (86 percent) Article XX or XXI was the justification cited for a quantitative restriction. The most frequently cited paragraph of Article XX is the one allowing restrictions to protect human, animal or plant live or health. Outside of Articles XX and XXI, Article XVIII.B, allowing developing countries to use quantitative restrictions to defend the balance of payments, is the most frequently cited Article.¹³

Though Articles XII and XVIII.B still provide for trade restrictions to defend the balance-of-payments, the Uruguay Round Understanding on Balance-of-Payments Provisions limited the scope for use of such. Even before that agreement, developing economies had been under considerable pressure to remove such restrictions, and use of the provision has declined. Only India, Nigeria, Tunisia, Bangladesh and Pakistan maintain restrictions under XVIII.B, and schedules for phasing out balance-of-payment based non-tariff-measures have already been agreed with Tunisia and Pakistan. India�s restrictions justified under XVIII.B are subject to a dispute settlement proceeding, and India has re-notified some of the measures under other GATT articles Nigeria's BOP-based quantitative restrictions have also been questioned.

The Uruguay Round Safeguards Agreement imposes time limits on all new safeguard measures. In concert it requires the phase-out of all existing safeguard measures now legal under GATT Article XIX. It also requires the elimination of all VERs and other restraints that would not be allowed under the Uruguay Round safeguard rules.¹⁴ Table QR2 lists all such measures notified by WTO Members. Only eight of the measures notified are measures taken under Article XIX that now come under sunset provisions. The brevity of this list, say, in comparison with the number of antidumping measures in place (below) reflects how infrequently Article XIX measures have been used to deal with troublesome imports. The other 18 measures listed are notified as inconsistent with the new Safeguard Agreement and subject to phase-out. Most of the measures (of both types) have been eliminated already, the remaining three are scheduled for elimination at the end of 1999.

The information in Table QR2 demonstrate that VERs have disappeared. To confirm that finding we looked for other information on VERs. A GATT tabulation (GATT 1992) identifying 79 such restraints outside agriculture and textiles/clothing that were in place as of 1992. Korea (46) and Japan (23) were most frequently involved on the exporters' side. On the importers' side, the European Union and the US accounted for 33 and 17 VERs each. Notifications since the Uruguay Round indicate that these arrangements are gone or will be gone on the deadline specified in the agreement. Among exporters, Japan submitted no report of continuing VERs,¹⁵ Korea�s report indicates that all VERs with the US and the European Union have been eliminated. Thailand�s notification indicates the same for the Thai VER on manioc to the European Union. Though the WTO does not require such notification, eighteen Members have notified the WTO that they maintain no quantitative restrictions. They are listed in a footnote to Table QR1.

WTO Trade Policy Reviews further support the conclusion that VERs have disappeared. The most recent TPR for Japan reports that all VERs in which Japan participated have been eliminated except the Japan-EU VER on cars - and that one will be gone by end 1999 (WTO, 1998). The 1996 Trade Policy Review for the US reports that the VERs agreed with other countries had been removed (WTO 1997). The accuracy of that finding is confirmed by the fact that it was not contested in discussion of the Trade Policy Reviews.

Despite this "elimination" of VERs, there are still WTO-legal ways for trade disputes to be resolved by exporters agreeing to restrain shipments. The antidumping and subsidies, countervailing measures agreements allow such "undertakings." Following the filing in 1998 of antidumping and countervailing duty cases by the US steel industry, the US reached agreement to a VER with non-WTO member Russia. As of March 1999 when this study was written, Brazil had proposed to curb exports voluntarily if a US antidumping case is suspended, and much pressure is being put on Japan to reduce exports to the US.

Tables QR3 and QR4 provide a view of recent reductions in the application of NTBs. The tabulations summarized in the tables cover export restraints, non-automatic licensing, variable charges and quantitative restrictions and price-quantity measures, but exclude antidumping and countervailing duties.

Among OECD countries, the principal finding is that the share of tariff lines affected by such measures declined significantly, or remained at a very low level (Australia and New Zealand) in all countries covered. In those two, the index remained constant. The decline in the NTB index for Norway is mainly due to the tariffication of a large number of agricultural tariff lines, the elimination of VERs is a key reason for the decline in NTBs in the US and the EU.

Information on developing economies� use of NTBs extracted from WTO Trade Policy Reviews suggests a significant reduction in use of NTBs.¹⁶ Colombia, Indonesia, Korea, Malaysia, Mexico, Morocco, South Africa and Thailand now take much less recourse to such trade restrictions. Among the developed economies listed in Table QR4 (which is based on TPRs) only one, Brazil, had a higher NTB index value in 1995-98 than 1989-94, and for Brazil the increase was small, less than 2 percentage points

WTO Trade Policy Reviews explicitly acknowledge the implementation of the Uruguay Round Agreements in bringing down NTBs in Mexico and Thailand by about 50 percent. In Colombia and Korea, the elimination of quantitative restrictions under the balance of payments provision has resulted in a major reduction of NTBs. The decline in the use of the balance of payments provision has been an important achievement in the post-UR period, although it was not explicitly part of the Uruguay Round package.¹⁷

Source: Notifications submitted by Member Countries to the Market Access Committee
(*) Only GATT Articles
(**) Four countries in addition to the three listed here maintained at the end of 1998 quantitative restrictions under the balance of payment provision (Article XVIII.B). They are Bangladesh, Nigeria (no conclusion of last consultation), Pakistan and Tunisia.
(***) P=Prohibition, Q=Quota, L=Licensing, O=Other
P includes: Prohibition (P), Prohibition except under defined conditions (CP)
Q includes: Global quota (GQ), Global quota allocated by country (GQC), Bilateral quota (BQ)
L includes: Automatic licensing (AL), Non-automatic licensing (NAL)
O includes: QR made effective through state trading (STR), Mixing regulation (MXR), Minimum Price (MPR), Voluntary Export Restraint (VER)
Note: Countries which notified that they do not maintain quantitative restrictions and year of notifications: Bahrain (1997), Bolivia (1997), Brunei Darussalam (1996), Costa Rica (1998), Dominican Republic (1996), Gambia (1997), Haiti (1998), Honduras (1997), Iceland (1996), Liechtenstein (1997), Mongolia (1998), Paraguay (1998), Singapore (1996), Switzerland (1997), Trinidad/Tobago (1996-1998), Uganda (1996), United Arab Emirates (1997), Uruguay (1996)

Table QR2: Pre-existing measures notified by Members as subject to elimination under rules adopted at the Uruguay Round
(grey area measures and measures legal under Article XIX now subject to sunset provisions)

Source: Notifications submitted by Member Countries to the WTO Committee on Safeguards
^a Article 10, GATT XIX measures; Article 11, VERs and other measures inconsistent with the Safeguards Agr.

Table QR3: Core Non Tariff Measures* in Selected OECD Countries
(Share of tariff lines with at least one NTB, in percentage)

Source: OECD 1997
* Core Non-Tariff Measures include: export restraints, non-automatic licensing, other quantitative restrictions, variable charges and other price control measures. The figures do not cover antidumping, countervailing duties and voluntary export/price restraints.

Source: WTO Trade Policy Reviews
^a Core Non-Tariff Measures include: export restraints, non-automatic licensing, other quantitative restrictions, variable charges, and other price control measures. The figures do not include coverage of antidumping, countervailing duties and voluntary export/price restraints.
^b Relative NTB incidence where NTB incidence is measured by the share of 85 broad data categories with at least one NTB.
Revised April 22, 1999

Since the 1930s, developed economies have used negotiated or "voluntary" export restraints to limit their imports of textiles and clothing.¹⁸ Finally, in the Uruguay Round Agreements, the international community decided to put an end to this practice. The major provisions for eliminating quotas and VERs on textiles and clothing are as follows:

• all textiles and clothing products will be "integrated into GATT 1994" in four stages (1/1/1995, 1/1/1998, 1/1/2002, 1/1/2005); encompassing 16 percent, 17 percent, 18 percent and 49 percent (by 1990 volume) of imports of all specified textiles and clothing products;
• import growth rates for all products not so integrated (i.e., that remain under restraint) will be increased in each of the three intervals between stages by 16 percent, 25 percent and 27 percent, respectively;¹⁹
• each of the four groups into which the spectrum of textile products has been divided (tops and yarns, fabrics, made up textile products, clothing) must be included in each of the integration.

The percentages listed in the first bullet above apply to "the total volume of the Member�s 1990 imports of the products in the Annex."²⁰ The annex runs to more than 30 pages of six-digit HS products and includes all textile and clothing products that were subject to MFA or MFA-type restraints in at least one importing country.²¹ Any one Member is unlikely to have restraints on all of the products on the list, and hence in the early stages will have the opportunity to "integrate" products on which it has no restrictions. As a matter of semantics then, the operative phrase in the agreement, to integrate into GATT 1994 is better described as "certifying that a product is clean of restrictions"²² than as "removing MFA restrictions."

• A transitional safeguard measure may be applied by a Member ²³ to any product in the Annex, except for products already integrated into the GATT, i.e., liberalized as per the first bullet above.

Constraints for using the transitional safeguard mechanism are tighter than those for, say, the safeguards agreement - e.g., a transitional safeguard cannot extend beyond three years. Transitional safeguards may however be applied against specific exporters.

The developed economies� policies toward these imports are among their most restrictive. Hufbauer and Elliott estimate, for example, that almost 9/10 of the cost to the US economy of US import restrictions are accounted for by restrictions on imports of textiles and clothing. At the same time, textiles and clothing account for more than 20 percent of developing economies� industrial exports, hence there is much to gain all around from liberalization.

Implementation has proceeded through the first two stages; thus, each importing country has integrated into GATT 1994 products accounting for at least 33 percent of its imports. There have been, however, loud complaints that minimal liberalization has resulted from this implementation. The most often voiced complaints are that each importing Member has weighted its liberalization toward products:

• that were not under restraint in that country,
• with little value added or on which developed economies do not have comparative advantage, e.g., yarns and fabrics rather than clothing,
• with high tariffs,
• and that they have overused transitional safeguards or have applied antidumping and other WTO-legal restrictions disproportionally against textiles and clothing.

We will review below available evidence on each of these.

The tabulation in Table TC1 - taken from information notified to the Textiles Monitoring Board - does indicate that importers have selected items that were not under restriction. Norway, of the countries in the table, is the exception. Norway has decided to liberalize more rapidly than the agreement requires.²⁴

For the US and Canada (combined) textile imports and clothing imports are about equal, for Western European countries (combined) clothing imports are considerably larger than textile imports.²⁵ Even so, clothing has made up a small part of the products the countries listed in Table TC2 which have been integrated into GATT 1994 - among the best of them, hardly more than 10 percent.

Tariffs on textiles and clothing remain high relative to those on industrial products generally (Table TC3). Table TC3 also reports that the Uruguay Round tariff cuts on textiles and clothing were relatively large. Not only were the cuts (measured by dT/(1+T) deeper on textiles and clothing, they were applied to a larger fraction of imports.

Table TC4 lists all transitional safeguard measures implemented from implementation of the Uruguay Round Agreements through March 19, 1999. Soon after the agreement went into force, the United States applied a number of transitional safeguards, all but one of which had been lifted by March 1999. The US has not imposed a transitional safeguard since March 1996. Brazil�s actions, all imposed in January 1996, are mostly still in place. The only other country to apply a transitional safeguard measure was Colombia, but the two measures applied by Colombia in 1998 have both been rescinded.

Table AD6 reports recent antidumping initiations by sector. From these dates it appears that the EU has initiated a significant number of cases against textile and clothing imports. These cases however were not on products that the EU liberalized in the first two rounds of phasing textile and clothing tariff lines into the GATT 1994.

Table TC1: Numbers of MFA Quota Limits Notified and Eliminated in Stages 1 and 2
(Stage 1 plus stage 2 requires integration of 33%, by import volume.)

Source: WTO Doc. G/L/179, page 29. Norway G/C/M/23, p.23.

Source: WTO Doc. G/L/179, page 29.

Note: ^a Averaged over all textiles and clothing tariff lines, including those with zero cuts.

Source: Reports of the Textiles Monitoring Body of the WTO

The agriculture negotiations focused on three categories of policy; import restrictions, domestic support programs, and export subsidy programs. We will pay attention principally to import restrictions.

On import restrictions, the major objective of the agreement is to establish a "tariffs only" regime - to eliminate all forms of import restriction other than bound tariff rates. To do so, all members were required to "tariffy" their non-tariff import restrictions. This conversion was based on the price gap methodology, with the methodological details being set out in technical guidelines on how to measure the gap between the domestic price (the price inside the protection wall) and the world price.²⁶ The base period for the conversion, members agreed, would be 1986-1988, a period when many agricultural prices had been unusually low. Because agricultural policies try to maintain a relatively stable - and high - domestic price, the price gap calculated from this base period coupled, in some cases, with other adjustments allowed by the technical guidelines, frequently resulted in high tariff rates. Developing countries had the option to submit ceiling rates on previously unbound tariff items, with the additional proviso to remove all agriculture-specific non-tariff measures. Each member�s legal obligation is defined by its schedule of tariff rates annexed to the GATT 1994.²⁷

According to the WTO Secretariat, forty countries participated in the tariffication process which covered (in aggregate) about 22 percent of their tariff lines.²⁸ Finger-Ingco-Reincke calculations over the IDB show that tariffication covered, by value, just over one-third of tariffying countries� agriculture imports.

The modalities document also gave targets for tariff reductions. A developed country member was to reduce its duties, including those resulting from tariffication, across all agricultural tariff lines by a simple average of 36 percent over six years, with a minimum reduction of 15 percent on individual products. A developing country member was to reduce its duties by 24 percent over 10 years, and least developed countries were not required to make reductions.²⁹

As part of the tariffication package, WTO Members agreed to maintain, for tariffied products, "current" import access opportunities at levels corresponding to those existing during the 1986-88 base period at terms not less favorable than in that period. Where such current access had been less than 5 per cent of domestic consumption of the product in question in the base period, the agreement required an (additional) minimum access opportunity on a most-favored nation basis ³⁰ at a low tariff rate. This was to ensure that in 1995, current and minimum access opportunities combined represented at least 3 per cent of base-period consumption and are progressively expanded to reach 5 per cent of that consumption in the year 2000 (developed countries) or 2004 (developing countries), respectively.³¹

Implementing the tariff part of the agreement involves the following steps:

• As explained above, Convert to tariffs-only,
  i.e., determine new tariff rates for all tariff lines with NTBs, eliminate the NTBs.
  (Many of the new rates are specific rates, not ad valorem rates.)
• Bind all tariff lines - those on which NTBs have been converted to tariffs and those, on which there were tariffs only, i.e., no NTBs.
• Reduce the bound rates.

legal commitments are those specified in the schedules of rates attached by Members, possibly higher - even including staged reductions - than the tariff equivalent of initial measures. The Finger-Ingco-Reincke calculations that we report here are based on work by Ingco (1995) that used 1986-1988 as the base period for calculating (a) the ad valorem equivalents of overall protection applied in that period, and (b) the ad valorem equivalents of the tariffied bound rates.

Ingco found that many of the new bound rates were above the rates actually applied in the base period, and that some of the post-Uruguay Round bound rates - the rates that incorporate all of the scheduled reductions - are above the tariff equivalents of all protection applied before the Uruguay Round. While the focus of the Finger-Ingco-Reincke calculations we report here is on isolating the instances in which the Uruguay Round schedules do imply a reduction of protection, we will pay attention to her evidence on this "dirty tariffication."

Suppose for example, that the ad valorem equivalent rates for a particular tariffied line, as of the base period, were as follows:

In this instance, the Finger-Ingco-Reincke calculations would take as the "before-UR" rate the 20% applied rate, not the 30 % bound rate. Only if the final-UR bound rate is below the "before-UR" applied rate does the country�s Uruguay Round commitment imply a tariff reduction. Thus the "after-UR" rate in the Finger-Ingco-Reincke calculations is the lower of the "before-UR" rate or the post-UR bound rate.

The guideline of a 36 percent reduction is met in this example - the reduction of the bound rate from 30% to 18% is more than a 36 percent reduction. The Finger-Ingco-Reincke calculations, however, look at how the Uruguay Round has reduced applied protection, and would include only the reduction from the previously applied rate, i.e., from 20%, to 18%. If the final-UR bound rate were at or above the base period applied rate - in this example, at or above 20%, the Finger-Ingco-Reincke calculations would attribute zero reduction to the Uruguay Round.³²

Judging from the sample of countries in the IDB, Table A1 reports that both developed and developing economies have now bound virtually 100 percent of their agricultural tariff lines, overall an expansion of coverage of about two-thirds for the developing economies, one-fourth for the developed economies. Uruguay Round adjustments imply reductions of tariff rates on about one-fourth of developed economies� imports, on about one-fifth of developing economies� imports. The developing economies were expected to make smaller cuts, but our results show a larger depth of cut for developing economies.³³

The figures for scope and depth of cut by the developing economies in the IDB, are probably not representative of developing economies in general. The IDB covers all of the developed economy members of the WTO, but it covers none of the least developed countries, who were not expected to make reductions of their agricultural tariffs. The developing economies in the IDB tend to be those that have implemented significant trade reforms.

Chart A1 converts the extent and depth of tariff cuts by the developing economies into multiples of performance by the developed economies. We see there the same pattern we found for tariffs overall. Developing economies� tariffs are still considerably above those of developed economies; the extent (import coverage) of developing economies� Uruguay Round tariff reductions was smaller. But the depth of cut when we measure by how it will affect trading partners� market access was considerably more by the developing economies -- at least for those who are in the IDB.

The above figures do not take into account how the minimum access commitments (implemented through tariff quotas) have affected the scope of liberalization. This impact depends, of course, on which of a country�s tariffied products have imports below 5 percent of base period domestic consumption, and by how much.³⁴ At the maximum (i.e., if imports of all tariffied products were below the minimum) minimum access opportunities would affect the 22 percent of tariff lines, or approximately one-third of imports, by value, that were tariffied. Hathaway and Ingco (1996, p. 49) estimate that Japan and Korea�s minimum access commitments on rice will result in nearly a million tons per year of new imports, an expansion of world trade in rice of 7.5 percent over its 1992 level. Otherwise, they conclude that "the minimum access commitments will provide relatively little additional access and even less additional trade," no more that 0.5 percent for wheat and sugar. (pp. 48, 49)

On agriculture products as on manufactures, developing economies committed to ceiling bindings above their applied rates. Averaged over the developing economies in the IDB, the average applied rate in the base period was 18 percent, their bound rates average almost 60 percent.

Countries that converted NTBs to tariffs have in some cases posted rates higher than the base year tariff equivalent of those NTBs. Japan, for example, has announced that beginning in April 1999 its tariff on rice will be $3.05 per kilo. International Trade Reporter (1998) estimates that this rate is equivalent to 1,000 percent, ad valorem. This rate applies, of course, only to imports in excess of Japan�s minimum access commitment. Hathaway and Ingco calculate that Japan�s actual base period protection on rice had a tariff equivalent of about 650 percent.³⁵ Ingco identifies other instances in which a developed economy�s post-UR bound rate is above the tariff equivalent of its base year protection.

The provision on special safeguards applies only to products that were subject to tariffication. They allow additional duties to be applied in case of a precisely defined surge of import quantity, or cases of imports at prices below a precisely defined reference level. A member can apply a special safeguard to a product for the remainder of the relevant year, on a shipment by shipment basis, only if the member noted in its schedule of commitments that it claimed the right to (eventually) do so. Thirty-eight members have reserved that right, on varying numbers of products. For all relevant Members combined, these reservations imply a potential for imposing special safeguards on almost 6,100 tariff items. During the period January 1998 through September 1998, volume-based actions were taken by five Members affecting a total of 128 tariff items and price-based actions were also taken by five Members affecting a total of 72 tariff items.

Notes: ^a Tariffied products: percentage (by value) of imports with final-UR bound rates (rates that include reductions) below the tariff equivalent of base period protection. Untariffied products: percentage of imports with final-UR bound rates below base period applied rates.
Source: Finger-Ingco-Reincke, Tables G2.

Notes: ^a Change measured as dT/(1+T), expressed as a percentage. Tariffied products: change from the tariff equivalent of base period protection to final-UR bound rate. Untariffied products: change from base period applied rate to the lower of base period applied rate and final-UR bound rate.
^b Ad valorem equivalent based on 1986-1988 prices.
Source: Finger-Ingco-Reincke, Tables G2.

The Uruguay Round Antidumping Agreement (formally, Agreement on Implementation of Article VI of the GATT 1994) adds considerable detail, e.g.:

• specificity has been added on the required evidence and methodology for determining dumping and injury,
• rules regarding support from domestic industry of the antidumping claim are now spelled out,
• procedural rules for hearing both sides and the time frame of investigations have been tightened,
• a "de minimis" clause (specifying that the dumping margin has to be at least 2 percent) have been added,
• notification requirements are explicit,
• a sunset clause and requirement for review have been added.

The use of antidumping has spread significantly from the five original industrial country users (Australia, Canada, the EU, New Zealand and the US). In a first wave, a number of advanced developing countries including Mexico, Brazil and Argentina starting using antidumping in the late 1980s. More recently, many smaller and poorer developing countries have sporadically taken recourse to antidumping, so that the total number of countries which have used this means of trade protection now exceeds 30. The use of antidumping per dollar of imports is notably higher among developing economies than among developed economies.

Among the traditional users of antidumping, the number of initiations has declined since the Uruguay Round Agreements have been in effect, i.e., since January 1, 1995. The filing of a number of antidumping cases by the US steel industry at the time of this writing (March 1999) may however reverse the decline in the United States. Among new users, for which data are available -- Argentina, Mexico, Korea and Turkey report an increase in the number of tariff lines the trade covered by antidumping measures. Among the developing economies for which data are available, only Brazil applies antidumping measures to a declining share of tariff lines (Michalopoulos, 1999). Table AD1 along with Charts AD1 and AD2 illustrate the development of antidumping investigations over the 1991-98 period.

Table AD2 contains the countries most frequently affected by antidumping cases between 1992-94 and 1995-97. China has been targeted most frequently during both periods. The United States, one of the most frequent users of antidumping, is also one of the most targeted countries, second after China and targeted more than twice as often as Japan. After China and the US, Korea and Brazil were the main targets during the early 1990s; while Korea, Germany and Taiwan attracted the most investigations in the latter period. Industrial countries mostly target other high income and transition countries, with less than one quarter of the investigations involving developing countries (and about half of those, China). Developing countries by contrast initiate an almost equal share of investigations against either of the three country groups (Miranda, Torres and Ruiz, 1998). As to sectoral distribution, we find that producers of base metals, chemicals, machinery, electrical equipment, plastics and textiles frequently seek antidumping protection.

The share of initiated investigations (Table AD3) that lead to a restrictive outcome (provisional measures and affirmative findings) tend to be similar among all the countries in the table -- except Australia, where the proportion of cases there that lead to restrictive outcomes, particularly restrictive final outcomes, is notably below the figure for any other country.

Another important dimension of such protection is the use of price undertakings (exporters agreeing to maintain minimum export prices) versus tariffs. For lack of time-series data, we did a tabulation for seven of the main antidumping users for 1997. In this year, only Korea and the EU report a significant share of final measures being price undertakings (Table AD4). In Korea this share exceeds 50 percent, and in the EU 20 percent. In the other 5 countries for which data was readily available, 98-100 percent of final measures consisted of duties.

While the trade coverage of antidumping is small, the level of tariffs applied on average is quite high and has in some instances reached several hundred percent. The last column of Table AD4 provides average duties imposed as part of final measures in the same seven users in 1997. Colombia applies the highest average tariff of 60 percent and Korea the lowest of 28 percent. Four of the other countries apply average rates of 30 to 40 percent. This is very high compared to "normal" average tariffs of about 4 percent on industrial products in developed economies, 8 percent to 12 percent in developing economies. The EU's 40 percent for 1997 are also much higher than the 23 percent applied for the 1980-87 period (Messerlin, 1989).

Because of the sunset clause in the Uruguay Round Agreement, reviews of measures already in place make up an increasing share of investigations. In the EU, 79 out of 190 investigations have been reviews in the 1995-1998 period, in the US 116 reviews of measures in place when the Uruguay Round agreements came into effect were conducted in 1998.

A variety of disputes regarding procedures, standards for initiating and conducting investigations, as well as substantive elements necessary for imposing duties are emerging. The recent dispute between Guatemala and Mexico on Portland cement has illustrated that the procedural and technical requirements to conduct an antidumping investigation consistent with WTO rules are quite stringent. Many developing countries are likely to find it difficult to meet these requirements and may find themselves increasingly challenged, especially by the high-income countries. Developing economies have not been reluctant to challenge antidumping actions by the developed economies: Korea�s case against the US on D-Ram and India�s dispute with the EU on unbleached cotton fibers are examples.

The real test of the use or abuse of antidumping, however, may still be lying ahead. In recent years, the world economic climate has been quite favorable, and previous studies have shown that the use of antidumping is strongly correlated with the business cycles (see, e.g., Leidy, 1996). The slowdown in economic growth since the outbreak of the Asian crisis could, with a lag, lead to a rebound in antidumping measures and more retreats in market access commitments. The saber-rattling in the US and the considerable increase in cases by several countries in 1998 are, we hope, a transitory phenomenon.

One could hypothesize that the Asian crisis would change the pattern of user and target countries for antidumping. As to the Asian countries affected, each experienced a marked currency devaluation, which is likely to reduce import competition and hence pressures for protection. Consistent with this, we found that the five Asian countries affected by crisis initiated fewer new cases but became the target of investigations more often in 1998 as compared to 1997. Korea and Malaysia which had used antidumping to a moderate extent in 1996/97 almost stopped initiating new investigations in 1998. Indonesia continued using antidumping to some degree while Thailand and the Philippines had never been very active users.

As to antidumping initiations against the Asian crisis countries, there is no uniform development although some increase has probably taken place. The US and South Africa increased the number of antidumping investigations against South East Asia from 2 to 6 and from 2 to 9 respectively. The European Union, on the other hand, initiated only 4 new cases in 1998 (all of them against Korea) after 10 cases in 1997. Brazil and Mexico neither used antidumping in 1997 nor in 1998 against these countries. (Table AD5)

It could be hypothesized that sectors with disproportionately large liberalization would also report a stronger increase in antidumping cases since the end of the Uruguay Round. However, there is little evidence of this for four of the main user countries. In India use of antidumping has jumped since the end of the Uruguay Round, but it is not obvious that there is a correlation between new antidumping cases and sectoral liberalization. Disproportionate liberalization in wood products (EU, Australia) and in textiles/clothing/footwear (EU) may be responsible for an increase in antidumping claims in these two sectors. (Table AD6)

Notes: ^a/ Based on numbers of antidumping initiations 1995-98 and values of merchandise imports for 1996.
^b/ 1995-97 figure.
Source: WTO Secretariat, Rules Division; Antidumping Measures Database

Source: WTO Secretariat, Rules Division; Antidumping Measures Database

Source: WTO Rules Division; Antidumping Measures Database
^a/ In percent; average all final measures per country in 1997; when a decision ended in a range of duties, the mean value was taken.

Source: WTO Secretariat, Rules Division; Antidumping Measures Database

Like the antidumping rules, the GATT rules on subsidies and countervailing duties were superseded by a comprehensive Uruguay Round Agreement. The Subsidy Agreement distinguishes three types of subsidies. Subsidies contingent upon export performance or on the use of domestic over imported inputs are prohibited. They can lead to both countervailing duties and a dispute settlement case. Subsidies not specific to an enterprise or industry are considered "non-actionable." They can not be subject to countervailing duties or dispute settlement challenges although they can be brought to the Committee on Subsidies and Countervailing Measures if they result in serious adverse effects in another country. All other subsidies are "actionable", i.e., they may be subject to multilateral dispute settlement challenge or to countervailing action if they adversely effect another WTO Member.

Industrial countries had to phase out prohibited subsidies by 1998. Least developed countries have until 2003 to eliminate subsidies contingent on the use of domestic inputs, but are not required to eliminate export subsidies. Other developing countries must eliminate subsidies contingent on the use of domestic inputs by 2000, export subsidies by 2003. Special Provisions for transition economies are more complex, involving a requirement that prohibited subsidies be phased out by 2002, but at the same time providing some lee way to apply programs and measures necessary for transformation from a centrally-planned to a market, free-enterprise economy.

Like in antidumping, the subsidy agreement added considerable specificity on procedures and criteria for a countervailing duty investigation. The agreement also requires that countries notify their specific subsidies (as defined under the Subsidy Agreement) on an annual basis, and countervailing duty activities on a semi-annual basis.

There are two notable developments in the use of countervailing duties:

• Overall use has declined considerably since the end of the Uruguay Round, with much of this decline due to the absence of steel cases,
• The EU has become a user of countervailing measures since 1996 after hardly having used this instrument before.

The 1992 to 1994 period saw 137 cases with the US, Australia, Brazil and Mexico being the principal users (Table CV1). The European Community and its individual members, Brazil and South Africa were the main targets of countervailing duty cases for that period. Since the Uruguay Round agreements have come into force, the number of initiations has declined drastically. The 1995-1997 period witnessed only 33 cases, one quarter of the cases during the preceding 3 years. Apart from the US, the EU and New Zealand have become the main users of this provision, with EU countries and South Africa remaining the main targets (Table CV2). Although 1998 data are incomplete at the time of this writing, there are indications that the number of cases will be near the annual rates of 1992-95; considerably higher than the low rates in 1995-97.

Data on the trade coverage of countervailing cases is very limited. The US reports that cases initiated over the past 10 years (1988-1997) affected only 0.7 percent of total US imports. This is a decline by 50 percent compared to the 1984-94 period, when 1.4 percent of imports were affected (US ITC, 1998). We mentioned above that the EU does not report trade coverage of countervailing cases separately from antidumping. However, given that the first EU countervailing case of the 1990s was brought about in 1996 and only 9 cases were reported for 1996/98, trade coverage should be very small. The 4 cases from 1998, for example, affected trade of US$ 140 million.

Data on the share of decisions leading to duties versus undertakings and on the average tariff rates are also scarce, but available data for two countries suggest that countervailing duties are much lower than antidumping duties. New Zealand reports two decisions leading to duties and price undertakings, each with an average duty of 7.5 percent. The US notified over 80 percent of decisions leading to duties, and average tariffs only average 4.2 percent.

There were a number of conflicts over the application of the Subsidy Agreement and the use of countervailing duties during the 1995-98 period. With respect to multilateral challenges to subsidies applied by Members, the only adopted dispute settlement panel report (by March 1999) in this area on Indonesia's national car program was decided in favor of the plaintiff, the EU. However, a number of other subsidies panels are ongoing as of this writing, mostly dealing with subsidies that the plaintiff interprets to be in the prohibited category. There are also in process a number of cases on the application of countervailing measures.

Notifications regarding subsidies and countervailing duties seem to be provided relatively regularly. However, developing countries report subsidy data much less frequently, and all subsidy notifications are frequently late and incomplete.

There are no firm plans for new negotiations in this area, but some challenges remain. Given the upcoming deadlines to bring Members' subsidy schemes in line with WTO rules, there is considerable concern about developing and transition countries' ability to meet these deadlines.

Source: WTO Secretariat, Rules Division; Countervailing Measures Database

Source: WTO Secretariat, Rules Division; Countervailing Measures Database

The Agreement on Safeguards specifies the rules for the application of Article XIX emergency action measures. Under the agreement, action can be taken against imports only a determination of "serious" injury caused or threatened to be caused by imports.³⁶ Either quantitative restrictions or duty increases beyond bound rates are allowed, but measures are time bound and must be progressively reduced. The need to make compensation and the right of exporters to retaliate can be avoided by keeping the restrictive measure within certain parameters.³⁷ This restriction on compensation-retaliation was intended to make the use of safeguards measures more attractive, and aims at preventing the re-emergence of voluntary export restraints or orderly marketing arrangements. The agreement also requires that existing safeguard measures be notified and phased out according a schedule similar to the sunset provision that applies to new measures.

There has been a minor increase in cases since the Uruguay Round agreement came into effect, but safeguards remain an infrequently used trade remedy -- only 19 safeguard investigations initiated 1995-97 as compared with 33 countervailing duty investigations and over 600 antidumping initiations.³⁸ (Table SG1)

Total imports affected by investigations amounted to about 2 billion US$ between 1995 and 1998, or about one quarter of one percent of annual imports of the countries initiating investigations.³⁹ Two of the eight cases which have been decided affirmatively resulted in a quantitative restriction, the other six in tariffs. Import quotas were allocated in both cases on the basis of previous relative import shares. Four cases were terminated or ended in a negative decision, and four decisions were outstanding at the time of this writing.

Our findings on safeguards confirm earlier findings that market access retreats through existing provisions have so far been limited. At the same time, the new UR Agreements have been successful in curtailing protection outside the multilateral framework. Between 1995 and spring 1999, there have been no complaints that the prohibition of gray area measures under Article 11 of the SGA has been violated. Anticipation that degressive and time-bound safeguard measures would replace antidumping as an instrument for dealing with specific import problems has not been have been realized.