In the current Doha Round of WorldTrade Organization talks, negotiations arenow under way on the arcane, highly tech-nical, and intensely controversial subject ofantidumping rules. On the agenda are pos-sible changes to the WTO AntidumpingAgreement, which sets the standards thatgovern national antidumping laws.

In the United States, current antidump-ing rules enjoy strong political support.Defenders of the status quo allege thatchanges to the Antidumping Agreement will“weaken” the U.S. law and thereby exposeAmerican industries to unfair, “dumped”competition. In response to such concernsand at the United States’ insistence, theDoha declaration that launched the newantidumping talks insists on “preserving thebasic concepts, principles and effectiveness of[the Antidumping Agreement] and [its]instruments and objectives.”

Fears of weakening the U.S. law are mis-placed. The object of WTO negotiations isnot to weaken national antidumping lawsbut to improve them—by curtailing ram-

pant abuses that allow trade-restrictiveantidumping remedies to punish normal,healthy, import competition. Such abusesrun afoul of what supporters of antidump-ing claim is the purpose of the laws: name-ly, to ensure a “level playing field” by target-ing “unfair” trade practices that reflectunderlying market distortions. Accordingly,changes to antidumping rules are needed tobring national laws into conformity withthe basic principles, concepts, and objectivesof the Antidumping Agreement.

This paper sets forth a detailed road mapfor the WTO antidumping negotiations. Westart by fleshing out the basic concepts, prin-ciples, and objectives of the AntidumpingAgreement—as elucidated by the justifica-tions for antidumping laws offered by theU.S. government and prominent supportersof the U.S. law. Working from that founda-tion, we then outline 21 reform proposalsdesigned to improve antidumping laws’ aimand limit the collateral damage inflicted onbusiness practices that have nothing to dowith “unfair trade.”

Reforming the Antidumping AgreementA Road Map for WTO Negotiations

by Brink Lindsey and Dan Ikenson

December 11, 2002 No. 21

Brink Lindsey is director of the Cato Institute’s Center for Trade Policy Studies. Dan Ikenson isa trade policy analyst with the center.

Executive Summary

Introduction

In November 2001, representatives of 142countries convened in Doha, Qatar, and launcheda new round of global trade negotiations.Included on the agenda of this Doha Round ofWorld Trade Organization talks is the arcane,highly technical, and intensely controversial sub-ject of antidumping rules. Specifically, talks willfocus on possible changes to the existing WTOAntidumping Agreement, which governs whatWTO member states can and cannot do to pro-tect domestic industries from “dumped” or“unfairly priced” import competition.1

The WTO antidumping negotiations facestrong political opposition in the UnitedStates. According to that opposition, anychange in the WTO Antidumping Agreementthreatens to “weaken” the U.S. antidumpinglaw and so expose American industries tounfair foreign competition. Such concerns arereflected in the “trade promotion authority”(TPA) legislation passed by Congress inAugust 2002. Language in the bill instructs thepresident, in any trade negotiations, to

preserve the ability of the United Statesto enforce rigorously its trade laws,including the antidumping, countervail-ing duty, and safeguard laws, and avoidagreements which lessen the effective-ness of domestic and international disci-plines on unfair trade, especially dump-ing and subsidies, in order to ensure thatUnited States workers, agricultural pro-ducers, and firms can compete fully onfair terms and enjoy the benefits of reci-procal trade concessions.2

Nearly identical language was incorporatedinto a resolution passed by the House ofRepresentatives on November 7, 2001, on theeve of the Doha ministerial conference. Thatresolution passed by a vote of 410 to 4.3 And 62senators signed a letter to the president in May2001 warning him not to agree to any tradedeals that would weaken the antidumping orother trade remedy laws. “Unfortunately, some

of our trading partners, many of whom main-tain serious unfair trade practices, continue toseek to weaken these laws,” the letter stated.4

The TPA legislation almost included an amend-ment that would have provided for the denial of spe-cial “fast-track” voting procedures (i.e., an up-or-down vote by Congress without amendments andwithin specified time periods) to those parts of anytrade agreement that made changes to antidumpingrules. This so-called Dayton-Craig amendmentpassed in the original Senate TPA bill but was even-tually dropped in conference committee.

Mindful of domestic political pressures,both the Clinton and Bush administrationsstrongly opposed inclusion of antidumping onthe negotiating agenda for the new round. TheClinton administration refused to budge onthe issue, and the resulting impasse betweenthe U.S. government and other WTO mem-bers was one of the major contributors to thefailure to launch a new round at the Seattleministerial conference in 1999.5 The Bushadministration sought to continue its predeces-sor’s position but ultimately bowed to over-whelming international pressure and agreed atDoha to put antidumping rules on the table.

Although the U.S. government made thatimportant concession, it still sought to limit thescope of WTO negotiations. Specifically, atU.S. insistence, the provision in the Doha min-isterial declaration that authorizes antidump-ing talks reads in relevant part:

In light of experience and of the increas-ing application of these instruments byMembers, we agree to negotiationsaimed at clarifying and improving disci-plines under the [AntidumpingAgreement], while preserving the basic con-cepts, principles and effectiveness of [theAgreement] and [its] instruments andobjectives. . . . In the initial phase of thenegotiations, participants will indicate theprovisions, including disciplines on tradedistorting practices, that they seek to clarifyand improve in the subsequent phase.6

The commitment to preserve the “basicconcepts, principles and effectiveness” of the

2

The WTOantidumping negoti-

ations face strongpolitical opposition

in the United States.

agreement and its “instruments and objectives”was inserted after an effort by the UnitedStates to limit the scope of permissible changesto antidumping rules. The inclusion of “disci-plines on trade distorting practices” on thenegotiating agenda may open the door tochanges that expand national governments’authority to apply antidumping remedies.

The resistance to changes in the WTOAntidumping Agreement is based on a funda-mental misunderstanding of what antidump-ing laws actually do in practice. Those laws aredefended as necessary bulwarks against unfairtrade practices. But antidumping laws—in theUnited States and dozens of other countries—have little to do with targeting unfair tradeunder any plausible definition of that term.Stiff antidumping duties are routinely imposedagainst products of foreign firms that areengaged in perfectly normal and unexception-able commercial practices. At the root of theproblem are serious flaws in the current rulesfor conducting antidumping investigations.Because of those flaws, there is at present verylittle connection between the stated objectivesof antidumping policy and the actual effects ofantidumping actions.

Accordingly, the fear that changes in WTOantidumping rules will expose Americanindustries to unfair competition is entirely mis-placed. Significant changes in those rules areneeded, not to “weaken” national laws, but toimprove them by closing the yawning gapbetween what they are supposed to do andwhat they actually do.

The newly launched Doha Round of WTOtalks offers the chance to close that gap.Although the U.S. government attempted tolimit the scope of negotiations with restrictivelanguage in the Doha ministerial declaration, aproper reading of that language makes clear thatfar-reaching changes in antidumping rules arenot precluded. Indeed, that language—with itsemphasis on the basic concepts, principles, andobjectives of the Antidumping Agreement—provides an excellent point of departure for pro-ductive negotiations.

WTO antidumping negotiations remain intheir preliminary stages, so the timing is propi-

tious for creative thinking about how negotia-tions should proceed. In this paper we attemptto outline a road map for WTO antidumpingnegotiations in the Doha Round. We beginwith the language of the ministerial declarationand then proceed to identify the basic con-cepts, principles, and objectives of theAntidumping Agreement—as delineated, notby critics of antidumping practice, but by theU.S. government and policy experts who sup-port the use of antidumping remedies. We thenexamine how antidumping laws actually workin practice and compare that reality with thepurposes those laws are supposed to serve.

That examination defines the basic workprogram of current and future WTO antidump-ing negotiations: to reduce the gap between con-temporary antidumping practice and theagreed-upon concepts, principles, and objectivesof the Antidumping Agreement. We then elab-orate upon that basic mission with detailedanalysis of the specific changes that are neededin the Antidumping Agreement.

Identifying Antidumping’sObjectives

WTO antidumping negotiations shouldbegin at the beginning by attempting to definethe basic concepts, principles, and objectives ofthe Antidumping Agreement. No suchattempt has been made in any previous WTOor General Agreement on Tariffs and Tradenegotiations. That is an oversight whose cor-rection is long overdue.

The oversight may be explained by the factthat antidumping has been around muchlonger than the multilateral trading system.Antidumping laws originated in the early yearsof the 20th century; the U.S. law, for example,dates back to 1921,7 and laws in Canada,Australia, New Zealand, Great Britain, andFrance go back even further. These laws thuspredate even the original 1947 GATT treaty,Article VI of which provides basic authority fornational governments to apply antidumpingremedies. Subsequent negotiations to elaboratemultilateral standards for the use of antidump-

3

The resistance tochanges in theWTOAntidumpingAgreement is basedon a fundamentalmisunderstandingof what antidump-ing laws actually doin practice.

ing remedies—the Kennedy Round talks thatproduced an Antidumping Code in 1967, theTokyo Round talks that revised that code in1979, and the Uruguay Round talks that pro-duced the current 1994 WTO AntidumpingAgreement—simply assumed the backgroundfact of national antidumping laws withoutmaking any effort to establish a consensus onwhy such laws are needed or what purposesthey are meant to serve.

With so much water under the bridge, whyfocus now on what might be consideredabstract or theoretical issues? The most obviousreason lies in the language of the ministerialdeclaration that launched the Doha Round.That declaration authorizes negotiations toamend the existing Antidumping Agreement“while preserving the basic concepts, principlesand effectiveness of [the Agreement] and [its]instruments and objectives.” To ensure thatnegotiations do not exceed the scope of thislimiting language, it is necessary for parties todetermine in the first instance what the basicconcepts, principles, and objectives actually are.

Furthermore, antidumping negotiationspromise to play a critical role in determiningthe overall success of the Doha Round. In priorrounds, antidumping was at best a second-tierissue. The United States and what is now theEuropean Union accounted for the over-whelming majority of antidumping cases, andthey were united in opposing anything butmarginal changes in their laws. Although manyother countries may have had an interest inrestricting antidumping abuses, none with anybargaining power made that interest a top pri-ority. Consequently, the United States and theEuropean Union were able to containantidumping reform initiatives within narrowlimits without any real sacrifice of their ownmajor negotiating objectives.

This time, the situation is different. With theproliferation of antidumping laws in recent years,the threat that antidumping abuses pose to theworld trading system has become an issue ofintense and widespread concern. There was over-whelming support for the inclusion of antidump-ing on the agenda of the Doha Round; indeed, theUnited States was completely isolated in opposi-

tion. As negotiations proceed, many countries canbe expected to push antidumping reform as one oftheir top priorities. Accordingly, the course of theantidumping negotiations is likely to have impor-tant implications for the overall outcome of theround. Even if the U.S. government persists inresisting major antidumping reforms, it has a com-pelling interest in avoiding a rancorous deadlockthat jeopardizes its own negotiating priorities.

Unless some common ground is first estab-lished, negotiators with opposing interests willsimply talk past each other. Antidumpingnegotiations are simultaneously highly techni-cal and intensely controversial: the details arecomprehensible only to experienced specialists,and the general subject matter is one on whichviews are sharply conflicting and strongly held.That combination is a recipe for impasse andacrimony, not productive results.

A preliminary focus on defining the basicconcepts, principles, and objectives of theAntidumping Agreement could help avoid adeadlock that might wreck the whole round. Inmany other contentious sectors—for example,agriculture and services—parties have madeprogress by agreeing on basic principles with acommitment to gradual (if unspecified) imple-mentation of those principles in the future.Such an approach might be the only way forthe U.S. government to reconcile its presentopposition to significant changes in theAntidumping Agreement with its overridinginterest in a successful round. If the UnitedStates were to accept a clear definition of thebasic concepts, principles, and objectives of theAntidumping Agreement, other countriesmight be satisfied with fairly modest changesto it in the current round. The U.S. governmentcould then remove the antidumping issue as anobstacle to its own major negotiating objectivesand ultimately bring home an agreement thatkeeps controversial changes to U.S. law withintolerable limits.

Meanwhile, supporters of antidumpingreform have a strong interest in initial discus-sions on basic concepts, principles, and objec-tives. Such discussions, if conducted properly,could significantly strengthen the reformers’bargaining position. Specifically, they could

4

A preliminary focuson defining thebasic concepts, principles, and

objectives of theAntidumping

Agreement couldhelp avoid a

deadlock thatmight wreck the

whole round.

enable reformers to claim the rhetorical highground of support for “fairness” and a “levelplaying field.”

At present, defenders of antidumping staketheir case on the grounds of “fairness” and a“level playing field.” Any efforts to change cur-rent practice, they claim, are really just plots to“weaken” existing laws and create “loopholes” forunfair traders. Defenders of antidumping thusdefine the debate as a conflict between a levelplaying field on the one hand and unfair traderson the other. Those terms, needless to say,strongly favor maintenance of the status quo.

Defenders of antidumping have been able tocreate and hold this rhetorical advantagebecause they have never been required to define“fairness” and “level playing field” or explain howcurrent antidumping rules advance thoseadmirable-sounding goals. They simply assertthe connection between current antidumpingrules and fairness and rely on the complexities ofthe law’s methodologies to shield their assertionfrom scrutiny. For decades that strategy has beentremendously successful.

Negotiations aimed at fleshing out the basicconcepts, principles, and objectives of theAntidumping Agreement could allow antidump-ing reformers to call their opponents’ bluff. If themuch-invoked level playing field were actuallydefined—if the specific circumstances that sup-posedly give rise to unfair trade were spelled outand the criteria for distinguishing those circum-stances from normal conditions of competitionwere clearly delineated—then antidumpingreformers could argue with considerable forcethat the imposition of antidumping duties in anyother circumstances amounts to simple protec-tionism. Antidumping reformers could turn theiropponents’ traditional rhetorical advantageagainst them and claim with justice that they, notdefenders of the status quo, are the ones trulyconcerned with fairness.

They could, in other words, redefine theterms of the debate. Instead of a choicebetween the level playing field and unfairtraders—with defenders of antidumping onthe side of the angels—the debate would nowoffer a choice between a level playing field andold-fashioned protectionism. Supporters of

antidumping would at last find themselves onthe negotiating defensive.

The effects of changing the terms of debatewould be felt within the WTO negotiationsthemselves, and also in U.S. domestic politics.Within the WTO, antidumping reformerswould be better able to recruit allies and isolatetheir opponents. Their position would be bothmore attractive and easier to understand, andtheir ability to persuade fence sitters to join theircause would be correspondingly enhanced.Antidumping reformers were able to force theUnited States to accept antidumping negotia-tions in Doha only because they succeeded inisolating the United States diplomatically. Ifthey hope to achieve significant reforms in thepresent negotiations, they will once again haveto rally world opinion to their side. This timethe task will be considerably more difficult.Building a consensus on what antidumping issupposed to do (and on the fact that it isn’tdoing its job properly) would aid the neededdiplomatic effort immensely.

Meanwhile, in the United States, a debatethat highlighted the contrast betweenantidumping’s objectives and its current prac-tice could drive a wedge between hard-coresupporters of the status quo and more casualsupporters. The hard core is concerned primar-ily with the interests of import-competingindustries (notably the steel industry) that usethe law regularly. To those supporters, resultsare all that count: anything that makes it easierfor domestic industries to win protectionmakes the law better, and anything that makesprotection harder to achieve is a step backward.That standard, of course, has nothing to dowith any notions of fair trade; it is a protec-tionist standard, pure and simple.

On the other hand, many U.S. supporters ofthe antidumping law are not so blatantly resultsoriented. Rather, they are attracted to the ideaof a level playing field and believe thatantidumping remedies work to secure thatnoble-sounding objective. If those casual sup-porters could be made aware of the disconnectbetween the law’s appealing rhetoric and howthe law really works, they might be moreamenable to changes in antidumping rules—or

5

Negotiations aimedat fleshing out thebasic concepts,principles, andobjectives of theAntidumpingAgreement couldallow antidumpingreformers to calltheir opponents’bluff.

at least less hostile to such changes. WTOnegotiations that focused initially onantidumping’s basic concepts, principles, andobjectives would illuminate that disconnect,thereby helping to reduce casual supporters’attachment to the antidumping status quo.That change in the U.S. political climate couldlead in turn to a more accommodating U.S.position at the negotiating table.

Defining the“Level Playing Field”

What are the basic concepts, principles, andobjectives of the Antidumping Agreement? Theagreement itself is silent on those matters. Itestablishes standards for how antidumping inves-tigations are to be conducted and remediesimposed, but it says nothing about why dumpingis a problem in the first place. The agreement, inother words, defines the “solution,” but not theproblem that it supposedly solves. We thereforemust look outside the agreement for guidance.

We propose to look for that guidance in thestatements of the U.S. government and promi-nent supporters of the U.S. antidumping law.There are many sound reasons why this approachmakes sense. The United States is the world’sleading antidumping user.8 In the internationalarena, the U.S. government has been the leadingdefender of the need for “strong” antidumpingremedies. In the current WTO negotiations, theUnited States is expected to be the most formi-dable opponent of any changes to theAntidumping Agreement that might “weaken”national laws. Accordingly, if it can be shown thatexisting antidumping rules do a bad job ofaddressing the problem of dumping as U.S.antidumping supporters define that problem, thestrongest possible case for changing the rules willhave been made.

The U.S. government recently issued a posi-tion paper on the “Basic Concepts andPrinciples of the Trade Remedy Rules” in theDoha Round antidumping negotiations.9 Thispaper provides an excellent starting point forunderstanding the problem that antidumpinglaws supposedly solve.

The Bush administration’s position paperadopts what has become the standard refrain ofU.S. antidumping supporters: that antidumpingmeasures are needed to offset artificial competitiveadvantages created by market-distorting governmentpolicies. According to the document, “Effectivetrade remedy instruments are important torespond to and discourage trade-distorting govern-ment policies and the market imperfections thatresult.”10 Specifically, the U.S. government argues,government policies can create “artificial” competi-tive advantages that may be distinguished from the“real” competitive advantages that arise in normalmarket competition:

Ideally, companies and nations wouldcompete in the international market-place on the basis of real comparativeadvantages such as natural resourceendowments, labor skills and abun-dance, availability of capital, and tech-nological innovation. Faced with thetrue relative prices of these productionfactors, companies and nations wouldgravitate towards producing andexporting those products in which theyhave a relative cost advantage and buy-ing/importing those products in whichthey do not have this advantage. . . .

However, government attempts tocreate artificial advantages distortmarket signals indicating where themost profitable business opportunitiesare found. Such distortions can lead tochronic oversupply by inefficient pro-ducers on the one hand, and the clo-sure of otherwise efficient and com-petitive facilities on the other. . . . Inshort, market-distorting practicesreduce worldwide economic efficiency,thereby diminishing the gains to allMembers from international special-ization and exchange based on com-parative advantage.11

This formulation of the problem that givesrise to antidumping laws differs somewhat fromthe others discussed below in that it focuses onefficiency rather than fairness. The Bush admin-

6

The AntidumpingAgreement estab-

lishes standards forhow antidumping

investigations are tobe conducted and

remedies imposed,but it says nothing

about why dumpingis a problem in the

first place.

istration has focused on the losses to worldwideeconomic efficiency caused by market-distortingpractices; the usual focus, however, is on theunfairness to national industries that must faceforeign rivals with artificial (i.e., government pol-icy–caused) competitive advantages. There is nonecessary conflict between the differingemphases, however.12 And in either case, the bot-tom line is the same: antidumping measures areneeded to neutralize artificial competitive advan-tages and restore the so-called level playing field.

The Bush administration’s paper goes on toidentify dumping as particular pricing practicesthat reflect the underlying existence of govern-ment policy-caused market distortions.Specifically, the paper states that dumping takesthe form of either “international price discrimina-tion” or “export pricing at levels below the cost ofproduction plus a reasonable amount for selling,general and administrative expenses and profit.”13

That definition tracks the one normally suppliedby U.S. supporters of antidumping, although thesecond half of the Bush administration’s defini-tion is more expansive than the usual formulation.Typically, as discussed below, dumping is definedas either international price discrimination orsales below the cost of production.

The pricing practices that constitute dumping,however that term is precisely defined, are prob-lematic because they supposedly are the conse-quence of market-distorting government policies.The Bush administration’s paper offers little detailhere, but the one example it gives is typical of thoseprovided by U.S. antidumping supporters:

A government’s industrial policiesor key aspects of the economic systemsupported by government inaction canenable injurious dumping to takeplace. . . . For instance, these policiesmay allow producers to earn highprofits in a home “sanctuary market,”which may in turn allow them to sellabroad at an artificially low price. Suchpractices can result in injury in theimporting country since domesticfirms may not be able to match theartificially low prices from producersin the sanctuary market.14

As will be seen in the discussion that fol-lows, the association between dumping and“sanctuary markets” figures prominently in thejustifications for antidumping laws offered bysupporters of the U.S. law.

The Bush administration’s interpretation ofthe basic concepts, principles, and objectives ofthe Antidumping Agreement shows consider-able continuity with the line taken by theClinton administration in another documentthat attempted to justify the use of antidump-ing measures. That earlier document goes intogreater depth than the Bush administration’sposition paper and also accords more closelywith the formulations of antidumping policymade by other prominent supporters of theU.S. law. Accordingly, the Clinton administra-tion’s paper merits detailed analysis.

The document in question is a 1998 sub-mission by the U.S. government to the WTOWorking Group on the Interaction of Tradeand Competition Policy.15 Some members ofthat working group had asserted thatantidumping laws should be judged by thestandards of competition policy—in otherwords, on the basis of whether they promoteconsumer welfare by targeting anti-competi-tive conduct. In the submission in question, theClinton administration argued vociferouslyagainst that approach:

Stated simply, the antidumping rulesand competition laws have differentobjectives and are founded on differ-ent principles, and they seek to reme-dy different problems. If theantidumping rules were eliminated infavor of competition laws or modifiedto be consistent with competition pol-icy principles, the problems which theantidumping rules seek to remedywould go unaddressed.16

In making its case, the submission expoundsat considerable length on antidumping rules’underlying objectives and principles—to definethe problems that antidumping rules seek toaddress. Accordingly, the Clinton administra-tion’s paper offers illuminating insights into

7

The associationbetween dumpingand “sanctuarymarkets” figuresprominently in thejustifications forantidumping lawsoffered by support-ers of the U.S. law.

what the basic concepts, principles, and objec-tives of the Antidumping Agreement might be.

The Clinton administration’s paper definesdumping as “a situation where an exporter sellsits product abroad at lower prices than it doesat home or at prices that are below cost, whichcauses ‘material injury’ to producers of theproduct in the importing country.”17 It thenasserts that the need for antidumping rulesarises from “imperfections in the multilateraltrading system.” Specifically, dumping is gen-erally the result of government policies in thedumping exporter’s home market:

Although some dumping may be due tobusiness advantages and market segmen-tation which have arisen in response tocommercial forces, more typically it is agovernment’s industrial policies or keyaspects of the national economic systemwhich a government has created, promot-ed or tolerated that enables injuriousdumping to take place.18

The 1998 U.S. submission identifies thefollowing industrial policies as possibly givingrise to dumping: high tariffs or nontariff barri-ers that exclude foreign competition, regula-tions that restrict domestic competition, theabsence of adequate competition laws to coun-teract private anti-competitive conduct, pricecontrols that set artificially high prices for theexported product or artificially low prices forinputs for the exported product, and govern-ment subsidies that give foreign producers anartificial cost advantage or that result in excesscapacity. “Although these policies take on manydifferent forms,” the paper states, “they providesimilar artificial advantages to the benefittingproducers.” The paper then elaborates:

Specifically, these policies enable the ben-efitting producers to charge higher thancompetitive prices in their home mar-ket—what can be thought of as a “sanc-tuary market”—and, as a result, to realizeincreased profits. If the government’spolicies have the effect of lowering theproducers’ unit costs, the producers may

benefit even when they maintain currenthome market prices. . . . Absent interven-tion by their own government, competingproducers in export markets are at a dis-advantage and often suffer injury, such aslost market share, because they cannotmatch the low pricing from producers inthe home market.19

The 1998 paper goes on to address how dif-ferences in national economic systems canresult in dumping. For example, in countries inwhich social pressures or policies inhibit layoffsduring downturns, labor costs are more fixedthan variable. In such settings, producers maychoose to sell below full costs instead of layingoff unneeded workers. In other words, they will“export [their] unemployment to the othercountry’s industry.”20 In another scenario, pro-ducers rely more heavily on debt in countrieswith poorly developed equity markets. Theymay find it necessary to sell below cost to ser-vice their debt obligations, whereas producerswith lower debt-equity ratios might cut backproduction during slumps. The paper alsoidentifies other situations—the presence oflarge, conglomerate business groupings withnoncommercial access to financing; cut-throatpricing encouraged by a policy of adopting“market stabilization” cartels on the basis ofprecartel production levels; state planningregimes with quantitative export targets—inwhich differences in national economic struc-tures can lead to dumping.

In sum, the Clinton administration’s papertraces the roots of dumping to anti-marketpolicies and institutions. The resulting sup-pression or distortion of market competitionyields either abnormally high prices in thehome market or abnormally low prices inexport markets—or both. Those artificial pric-ing patterns are the problem supposedlyaddressed by antidumping rules. According tothe 1998 U.S. paper:

[T]he antidumping rules simply seekto remove unfairness and create a“level playing field” for producers andworkers. It therefore may be more

8

According toantidumping sup-

porters, dumping isgenerally the result

of government policies in the

dumping exporter’shome market.

appropriate to view the antidumpingrules as a judgment by the importingcountry that it will not accept low-priced or below-cost imports—even ifits immediate overall economic wel-fare would be enhanced—to theextent that acceptance means forcingits producers and workers to competeagainst, and be injured by, foreign pro-ducers receiving unfair advantagesfrom government policies or actionswhich lead to significant differences ineconomic systems.21

Other prominent supporters of the U.S.antidumping law concur in the overall analysisput forward by the Bush and Clinton adminis-trations. Alan Wolff, counsel to the U.S. steelindustry and a leading lobbyist for theantidumping status quo, also identifies twotypes of dumping, “price-to-price dumping”and “below-cost dumping.” In a 1995 speechbefore the Steel Manufacturers Association, heanswered the question, “What gives rise todumping?” as follows:

Price-to-price dumping. Price-to-price dumping can occur because thedumping industry enjoys some degreeof market power in its domestic mar-ket which enables it to maintain ahigher price in the home market thanin export markets. This may arise outof protection of the home market fromimport competition . . . ; the relativeabsence of internal competitionbecause of the existence of a monopo-listic, oligopolistic or cartelized marketstructure; or some combination ofthese factors. Absent such elements,the domestic price and the world pricewill equalize.

Below-cost dumping. Below-costdumping can occur because the indus-try which is dumping possesses astructural characteristic which enablesit to export its products below the costof production for a sustained periodwithout going out of business. Such

characteristics vary widely, but mayinclude the existence of some form ofgovernment support, the ability tocross-subsidize losses in one productarea with profits earned in other areas,or simply enormous resources whichmake it possible to sell at a loss for along period of time.22

Terence Stewart, another prominent attor-ney who represents domestic industries in U.S.antidumping investigations, takes a similarline. The antidumping law, he argues, is“designed to offset any artificial advantage thatflows from closed foreign markets, cross-subsi-dization by multiproduct producers, govern-ment largesse, or other factors having nothingto do with comparative advantage.”23

Greg Mastel is probably the most promi-nent supporter of the U.S. antidumping law inthe public policy community. A former analystwith the Economic Strategy Institute and theNew America Foundation, he is the author ofAntidumping Laws and the U.S. Economy, astout defense of current rules.24 At present heserves as chief trade counsel for the U.S. SenateCommittee on Finance; as a key staffer on theSenate committee with primary jurisdictionover trade policy, he plays an important role inthe ongoing debate over antidumping reform.

Mastel’s views on the need for antidumpingrules align closely with those expressed in the twoU.S. government papers as well as those of Wolffand Stewart. He identifies government interven-tionism—whether targeted subsidies or the per-vasive controls of nonmarket economies—as amajor cause of dumping. In particular, he paysspecial attention to the interrelation betweendumping and sanctuary markets:

A secure closed home market or sanc-tuary market encourages companies tomake aggressive production and expan-sion decisions because they can be cer-tain of selling a percentage of their pro-duction at home at good prices. . . .From a sanctuary market, it is also pos-sible to dump in the markets of foreigncompetitors to depress the profit mar-

9

Other prominentsupporters of theU.S. antidumpinglaw concur in theoverall analysis putforward by theBush and Clintonadministrations.

gins of those competitors and reducetheir funds available for investment inR&D and marketing.

Companies from countries with openmarkets do not enjoy this luxury. . . . Overtime, this puts companies in open-marketcountries, such as the United States, at aserious disadvantage in competition withcompanies with sanctuary home markets.25

According to Mastel, a closed home marketallows a dumping strategy to work by main-taining a price differential between the homeand export markets:

If a company engages in dumping andits home market is open, the price differ-ential will induce the company’s com-petitors or other resellers to reexportdumped products to the dumper’s homemarket. These reexports would quicklypull the home market price down to thedumped price and erase home marketprofits. Thus, a closed or restricted homemarket is also a virtual precondition to asuccessful dumping strategy.26

Mastel argues that sanctuary markets can be creat-ed either by government-imposed trade barriers orby government acquiescence in the “private-sectorprotectionism” of anti-competitive collusion.

The justification of antidumping rulesadvanced in these analyses differs sharply fromthe popular view that antidumping is a remedyfor “predatory” private anti-competitive con-duct. The Clinton administration’s WTO sub-mission states flatly that “antidumping rules arenot intended as a remedy for the predatorypractices of firms or as a remedy for any otherprivate anti-competitive practices typicallycondemned by competition laws.”27 Masteltakes a similar view. “There are only a handfulof cases in recent history,” he writes, “in whichit reasonably can be argued that such a system-atic predatory strategy was being followed.”28

Instead, the U.S. government, through twoadministrations, and leading supporters of theU.S. antidumping law argue that the law isneeded primarily to offset the effects of distor-

tions caused by the anti-market policies of for-eign governments. Interventionist policies, it isargued, can confer an “artificial” or “unfair”competitive advantage on foreign producers,allowing them to charge lower prices in exportmarkets than at home or to charge below theircost of production. Antidumping remedies off-set that artificial advantage and thereby restorea “level playing field.”

This approach to defining the problem ofdumping offers a justification for antidumpingrules that can be distinguished from simple pro-tectionism. Many forms of government inter-ventionism do indeed distort markets and giveparticular firms an artificial competitive advan-tage. Although it remains highly debatablewhether trade barriers are the proper response tosuch market distortions, at least it can be main-tained plausibly that government policy–causedmarket distortions are a legitimate problem ininternational trade. The case for trade barriersnarrowly targeted at artificially advantaged firmsis clearly distinguishable from the case foracross-the-board protectionism.

Some of the analyses cited above, however,suggest that dumping can also be the result ofpurely private conduct. Thus, Wolff suggests thatprice-discrimination dumping can occur becauseof oligopolistic market power and that below-costdumping can be due to cross-subsidization by amultiproduct firm. Stewart also cites cross-subsi-dization as a possible cause of dumping.

This attempted extension of the definition ofdumping’s causes cannot survive careful scrutiny.Many manufacturing industries are characterizedby oligopolistic competition, and virtually allmanufacturing enterprises are multiproductfirms. Yet within the United States, there is noregulation of price premiums earned by oligopo-listic firms. For instance, it is not considered legal-ly actionable that a company with a strong brandname can use the strength of its brand to com-mand a higher price. Likewise, there is no gener-al regulation of cross-subsidies by multiproductfirms. For example, a firm that makes razors andblades may sell the former at or even below costin order to maximize revenue from the latter. Ifdomestic firms are completely free to engage insuch practices, one cannot argue plausibly that

10

The justification ofantidumping rules

differs sharply fromthe popular view

that antidumping isa remedy for

“predatory” privateanti-competitive

conduct.

foreign producers who do exactly the same thingare engaging in unfair trade.29 Any resort toantidumping remedies in such situations is indis-tinguishable from garden-variety protectionism.

Missing the Target?

Now that we have some sense of the problemantidumping rules are supposed to solve, we canturn to judging how well those rules actuallywork. Are antidumping remedies reliably target-ing artificial competitive advantages that resultfrom market-distorting government policies?Or are they frequently missing their target?

Unfortunately, there is a serious mismatchbetween what antidumping rules actually do andwhat their supporters say they are supposed to do.Antidumping rules are supposed to impose tradebarriers only as a response to market distortions,but as currently written and enforced they do aterrible job of distinguishing between market dis-tortions and normal, healthy competition.Consequently, antidumping remedies frequentlypunish foreign competitors for unexceptionablebusiness practices routinely engaged in by domes-tic industries. In other words, there is a huge gapbetween the basic concepts, principles, and objec-tives of the Antidumping Agreement and currentantidumping practice.

First of all, antidumping rules are supposedto identify instances of price discrimination orbelow-cost sales. In fact, however, the method-ologies used in antidumping investigations caneasily find dumping even when export pricesare above cost and the same as or higher thanprices in the home market. Consequently, themargins of dumping calculated in investiga-tions are often largely or totally artifacts offlawed methodologies rather than evidence ofthe targeted pricing practices.

Second, antidumping rules supposedly targetprice discrimination and below-cost sales on theground that they are the consequences of under-lying market-distorting government policies.But the truth is that companies can charge dif-ferent prices in different markets for perfectlyvalid commercial reasons having nothing to dowith government interventionism. Likewise,

sales below cost often indicate the presence ofvigorous competition rather than any kind ofmarket distortion. Accordingly, if antidumpingrules are to target the effects of market distor-tions without collateral damage to normal com-petition, they must have some mechanism forweeding out the “false positives.” Currentantidumping rules, however, lack any suchmechanism. As a result, antidumping rules fre-quently punish foreign producers for businesspractices having nothing to do with unfair trade.

Finally, antidumping remedies are supposedto be limited to offsetting artificial competitiveadvantages caused by market distortions. Inother words, they are supposed to produce a“level playing field,” not slant the field in favor ofthe domestic industry. To that end, antidumpingduties may be imposed only when dumping isfound to cause or threaten injury to the domes-tic industry. Because of flaws in current rules,however, remedies may be applied even whenalleged dumping from a particular import sourceis not harming or even threatening to harm thedomestic industry. Furthermore, antidumpingremedies are frequently far in excess of whatwould be needed to remedy any harm and thusexceed the levels arguably needed to secure a“level playing field.”

The flaws in current antidumping rules areexamined in great detail in two other CatoInstitute papers. First, we coauthored anotherCato Trade Policy Analysis published earlier thisyear titled “Antidumping 101: The DevilishDetails of ‘Unfair Trade’ Law.”30 In that paperwe provide a detailed analysis of the proceduresand methodologies used in U.S. antidumpinginvestigations, illustrating the many subtle andnot-so-subtle ways in which the law is biasedagainst normal, healthy, foreign competition.Second, in a 1999 Cato Trade Policy Analysistitled “The U.S. Antidumping Law: Rhetoricversus Reality,” Brink Lindsey presents a gener-al overview of how dumping is calculated andshows that current calculation methodologiesbear little relation to finding unfair trade.31

Rather than repeat at length the analyses andfindings of those papers here, we simply incor-porate them by reference. Here we take the con-clusion of those papers—that the U.S.

11

There is a seriousmismatch betweenwhat antidumpingrules actually doand what their sup-porters say they aresupposed to do.

antidumping law and, by extension, antidump-ing laws around the world, are woefully deficientin distinguishing between normal, healthy com-petition and government-caused market distor-tions—and examine its implications for theWTO antidumping negotiations.

The fundamental problem with antidump-ing rules today is their failure to limit the appli-cation of antidumping remedies to instances ofunfair trade under any plausible definition ofthat term. That failure defines the gap betweenthe basic concepts, principles, and objectives ofthe Antidumping Agreement and currentantidumping practice. Closing that gap byaltering the provisions of the AntidumpingAgreement ought to be the goal of WTOnegotiations—in the present Doha Round andin future rounds if need be.

Reforming theAntidumping Agreement

Below we offer a number of specific proposalsfor reforming the Antidumping Agreement—notto “weaken” national antidumping laws, but toimprove them. For each proposal, we identify someelement of current antidumping practice that con-flicts with the basic concepts, principles, and objec-tives of the Antidumping Agreement. We explainwhy the element in question is in need of reformand then discuss how the proposed reform orreforms would help to reduce the gap betweenantidumping theory and antidumping practice.

In explaining why reforms are necessary, wefrequently illustrate our analysis with examples ofmethodological distortions in current U.S.antidumping practice. We do this, not becausethe U.S. law is uniquely abusive, but simplybecause data on U.S. antidumping cases are muchmore accessible to us than equivalent data fromother countries. In particular, for purposes of thisstudy, we were able to gain access to the full evi-dentiary record of 18 recent U.S. dumping deter-minations. Accordingly, we were able to calculateprecisely how dumping margins would be affect-ed if various methodological distortions wereremoved—using the companies’ actual case dataand the actual computer programs of the U.S.

Department of Commerce for calculating dump-ing margins. Those calculations offer telling illus-trations of exactly how methodological distor-tions in current antidumping rules can act to gen-erate dumping margins out of thin air.

But reliance on U.S. examples should not betaken as a suggestion that U.S. antidumping rulesare especially flawed or that the U.S. law ought tobe the prime target of WTO negotiations. Onthe contrary, the methodological flaws thatplague the U.S. antidumping law are the normaround the world, and changes in the WTOAntidumping Agreement are needed to restrainabuses across the board. As we documented in anearlier paper, U.S. exports are now a leading targetof antidumping actions abroad—and a leadingvictim of antidumping abuses.32 Accordingly, itshould be remembered that the U.S. governmenthas an “offensive” as well as a “defensive” posturein antidumping negotiations.33

Require Evidence of Market DistortionsCritics of the antidumping status quo have

a long laundry list of complaints, but surely thefundamental problem with current antidump-ing practice is the failure to require any directevidence of underlying market distortions.

The supposed justification for targetingantidumping remedies at price discriminationand below-cost sales is that those pricing prac-tices reflect the existence of underlying, mar-ket-distorting government policies. And,indeed, unusually high home-market pricescan indicate a closed sanctuary market, andsustained red ink can be a sign of subsidies or“soft budget constraints.” But there is a host ofother possible explanations for internationalprice differences, most of which have nothingto do with unfair trade under any plausible def-inition of that term. Likewise with below-costsales: losses can be found in healthy, competi-tive markets as well as in distorted markets.34

Unfortunately, current antidumping practiceincludes no mechanism for weeding out “false pos-itives”—for distinguishing between those instancesof targeted pricing practices that actually reflectunderlying market distortions and those that have aperfectly innocent explanation. Under present rules,sales at less than normal value are simply assumed

12

The fundamentalproblem with

antidumping rulestoday is their failureto limit the applica-

tion of antidump-ing remedies to

instances of unfairtrade under any

plausible definitionof that term.

to be unfair—an assumption that is often com-pletely unsupportable. Consequently, antidumpingremedies are frequently imposed on exporters forengaging in normal commercial conduct that hasnothing at all to do with unfair trade or an unlevelplaying field.

Antidumping remedies will routinely deviatefrom the basic concepts, principles, and objec-tives of the Antidumping Agreement so long asthis fundamental flaw remains uncorrected.There are many possible approaches to address-ing this problem, but we suggest the followingreforms. First, domestic industries should berequired to present evidence of underlying mar-ket distortions in their antidumping petitions:

Reform Proposal 1: Article 5.2 of theAntidumping Agreement should beamended to require domestic industriesto provide credible evidence of underly-ing market distortions in the antidump-ing petition. If price-discriminationdumping is alleged, the evidence mustindicate the existence of (1) tariffs sig-nificantly higher than those in theexport market under investigation, (2)nontariff barriers significantly higherthan those in the export market underinvestigation, (3) government restric-tions on competition in the home mar-ket, or (4) government acquiescence inprivate anti-competitive conduct. Ifsales-below-cost dumping is alleged,the evidence must relate to the existenceof (1) subsidies that allow persistentlosses to continue or (2) other govern-ment policies that create a “soft budgetconstraint” that allows persistent lossesto continue.

In addition, national antidumping authori-ties should be required to make a finding ofunderlying market distortions before initiatingan investigation as well as in their final deter-mination of sales at less than normal value:

Reform Proposal 2: Article 5.3 of theAntidumping Agreement should berevised to require antidumping

authorities, before they initiate aninvestigation, to find that the domes-tic industry has provided credible evi-dence of underlying market distor-tions. Furthermore, Article 2 of theAntidumping Agreement should beamended to provide that authoritiesmay find dumping only if they deter-mine, on the basis of credible evidenceprovided by the domestic industry,that the price discrimination orbelow-cost sales found during theinvestigation reflect the existence ofunderlying market distortions asalleged in the petition.

It should be noted that the reform proposalabove does not confer any new investigatorypowers on national antidumping authorities. Anyopen-ended mandate to investigate “hidden”trade barriers, anti-competitive conduct, or othermarket distortions could easily lead to abusive“fishing expeditions.” Responding to antidump-ing questionnaires is already far too burdensome;requiring respondents, under the threatened useof “facts available,”35 to satisfy antidumpingauthorities’ potentially limitless curiosity aboutconditions in the home market would make a badsituation immeasurably worse. Accordingly, inthe above proposal, findings by antidumpingauthorities of underlying market distortions are tobe based on evidence provided by the domesticindustry in its petition and subsequent submis-sions—not on any independent fact-finding bythe authorities themselves.

Procedural fairness dictates that respondentsin antidumping investigations should be able torebut the evidence of market distortions provid-ed by the domestic industry. In particular, theyshould be allowed to provide affirmative defens-es to refute any causal connection between anymarket distortions in the home market and thepricing practices under investigation:

Reform Proposal 3: Article 6 of theAntidumping Agreement should berevised to give respondents the rightto present evidence that the pricingpractices under investigation are due

13

Antidumpingremedies are frequently imposedon exporters forengaging in normalcommercial conductthat has nothing atall to do with unfairtrade or an unlevelplaying field.

to factors other than market distor-tions in the home market. In investi-gations of price-discriminationdumping, respondents would have theright to show, for example, that (1)high home-market prices are due tonormal commercial factors (for exam-ple, strong brand-name recognition);(2) notwithstanding the existence ofhigh prices, the respondent does notenjoy unusually high home-marketprofits on sales of the subject mer-chandise (and thus does not enjoy anyartificial competitive advantage); or(3) the respondent’s home market istoo small for high profits in that mar-ket to confer an artificial competitiveadvantage in the export market underinvestigation. In investigations ofsales-below-cost dumping, respon-dents would have the right to show,for example, that below-cost saleswere made (1) to maximize the contri-bution to fixed costs; (2) to maximizeoverall revenue of joint products,products that share overhead, or com-plementary goods; (3) to maximizelong-term revenue by exploitinglearning-curve effects or by buildinglong-term market position; or (4) oth-erwise as part of a conscious strategyto maximize long-term profits.National antidumping authoritiesmust take this evidence into accountwhen determining whether the pric-ing practices under investigationactually reflect (rather than merelycoincide with) the existence of under-lying market distortions.

The reform proposals above, if adopted,would mark a significant departure from tradi-tional antidumping policy. But the fact is thattraditional policy has been sharply at odds withthe basic concepts, principles, and objectivesthat it supposedly serves. If the enormous gapbetween antidumping rhetoric and antidump-ing reality is to be closed, proposals along thelines of those suggested above will be an essen-

tial element of the overall reform program.

Eliminate the Cost TestCurrent antidumping rules fail to achieve

their supposed objectives on two basic levels.First, as discussed above, they make no attemptto connect the targeted pricing practices ofprice discrimination and sales below cost tounderlying market distortions. Second, they doa poor job of identifying actual instances ofprice discrimination and below-cost exportsales. Because of methodological distortions inthe rules that define dumping, findings of salesat less than normal value all too frequently havelittle or nothing to do with the presence ofprice discrimination or below-cost export sales.

The use of the “cost test” is probably thesingle most egregious methodological distor-tion in contemporary antidumping practice.36

Like many other distortions, it skews compar-isons of home-market and export prices andthereby artificially inflates dumping margins.What is especially noteworthy about the costtest, though, is that it operates to inflate dump-ing margins under specific conditions that arethe complete opposite of those that supposed-ly give rise to unfair trade.

The existence and extent of dumping are atpresent determined by a comparison of exportprices to “normal value,” which is typicallybased on prices in the foreign producer’s homemarket. If adjusted export prices are lower thannormal value, dumping is said to exist; the dif-ference between normal value and net exportprices, divided by net export prices, is thedumping margin or dumping rate.

Under the cost test, home-market sales foundto be below the cost of production are excludedfrom the calculation of normal value. In otherwords, all export prices are compared to those ofonly the highest (that is, above-cost) home-marketsales. This asymmetric comparison skews the cal-culation in favor of finding dumping.

In the U.S. antidumping cases we haveexamined, the effect on dumping margins wasdramatic. As Table 137 indicates, each of the 17cases against market-economy countries hadmargins that were inflated by the cost test(investigations of countries judged to be non-

14

The use of the “costtest” is probably the

single most egregiousmethodological dis-

tortion in contempo-rary antidumping

practice.

market economies use a different method forcalculating dumping in which the cost test isirrelevant). In two cases, the calculated dump-ing margin would have been zero had the costtest not been administered. Most of theremaining 15 would have had margins at least50 percent lower than the rate ultimately cal-culated. On average, the 17 cases would havehad margins 59.69 percent lower.38

What possible purpose could be served byexcluding below-cost home-market sales fromnormal value? Remember that the theorybehind price-discrimination dumping is thatthe foreign producer is enjoying an artificialadvantage because of a sanctuary market athome. According to the theory, trade barriersor other restrictions on competition causeprices (and profits) in the home market to beartificially high, thus allowing the foreign pro-ducer to unfairly cross-subsidize cheap exportsales. Consequently, price differences betweenthe export market and the home market aresupposedly probative of unfair trade becausethey might indicate the existence of a closedsanctuary market in the foreign producer’shome market. Whether those price differencesexist, though, cannot be fairly determined if allthe lowest home-market prices are excludedfrom the comparison.

Moreover, the existence of below-cost salesin the home market is actually affirmative evi-dence of the absence of a sanctuary market. Asanctuary market, after all, is supposed to be anisland of artificially high prices and profits. Ifhome-market sales at a loss are found in signif-icant quantities, isn’t that a fairly compellingindication that there is no sanctuary market?But because of the cost test, it is precisely underthose conditions that dumping margins areboosted significantly higher than they other-wise would be.

This absurd methodology clearly flies in theface of the basic concepts, principles, andobjectives of the Antidumping Agreement. Yetit is nonetheless specifically authorized underArticle 2.2.1 of the current agreement. Underthis authority, the cost test has become a cen-tral feature of antidumping investigations. Asurvey by one of this study’s coauthors of U.S.

antidumping cases over a three-year periodfound that only 4 of 37 determinations inwhich home-market sales were available as abasis for normal value employed a pure com-parison of home-market and U.S. prices. In 33of 37 determinations, or 89 percent of the time,the Commerce Department excluded some orall home-market sales through use of the costtest. The average dumping margin when thecost test was used was 16.14 percent; by com-parison, in the four determinations when thecost test wasn’t employed, the average dumpingmargin was only 4.00 percent.39

Because the cost test is explicitly authorizedin the current agreement, and because it is sucha regular feature of contemporary antidumpinginvestigations, resistance to reform in this areawill be fierce. And there is only one reform thatis adequate: outright elimination of the cost test.

The pretext for excluding below-cost salesfrom normal value is that such sales are not “inthe ordinary course of trade.” While it mightmake sense to exclude certain aberrant sales—sales of obsolete inventory or of damagedgoods—there is no serious case that unprofitablesales are outside the ordinary course of trade. Innormal, healthy, competitive markets, there isnothing extraordinary at all about red ink—especially on a product-specific rather than acompany-wide basis. It is absolutely routine forcompanies to fail to cover full costs of produc-tion on particular products at particular times.Selling below full cost is often the rational, prof-it-maximizing strategy. As long as variable costsof production are covered by the selling price,any contribution that price makes to coveringfixed costs is more than would be received if theproduct did not sell at all, which is often thealternative if the price is incrementally higher.

Accordingly, the Antidumping Agreementshould be revised to prohibit use of the cost test:

Reform Proposal 4: Article 2.2.1 ofthe Antidumping Agreement shouldbe rewritten to make clear that exclu-sion of home-market sales from thecalculation of normal value is permit-ted only in the case of specified aber-rational sales. In particular, sales must

15

The existence ofbelow-cost sales inthe home market isactually affirmativeevidence of theabsence of a sanc-tuary market.

16

not be categorized as outside the ordi-nary course of trade simply becausethey are made at less than the full costof production.

Until this reform is made, antidumping practicewill bear little relation to its stated justification ofremedying market distortions.

Revise Criteria for Use of “Constructed Value”;Eliminate Profit Component

Supporters of antidumping generally con-tend that dumping takes two basic forms:price-discrimination dumping and below-costdumping. For the former, the cost test dis-cussed above and many other methodologicaldistortions discussed below ensure that actualprice discrimination is rarely targeted inantidumping investigations, since comparisonsof prices are often badly skewed. Meanwhile,the inclusion of profit in the calculation of“constructed value” means that below-costexport sales are never directly targeted.

The closest that current antidumping rulescome to examining whether export sales arebelow cost is when constructed value is used asthe basis of normal value. Constructed value is theU.S. term for an artificial price that is determinedby calculating the unit cost of production for agiven product and then adding some amount forprofit. Constructed value is used only when (1)virtually all the foreign producer’s sales of the sub-ject merchandise are to the export market underinvestigation (that is, no “viable” home market orthird-country export markets exist); (2) there is aviable comparison market for the merchandise,but no models sufficiently similar to those sold inthe export market are sold there; or (3) all sales ofsimilar comparison-market models have beenexcluded by the cost test.

Accordingly, antidumping rules are egregious-ly misdesigned for the purpose of detectingbelow-cost export sales. First of all, export salesare compared to constructed value only underexceptional circumstances that bear no relation-ship whatsoever to the likelihood that export salesmay be below cost. If the problem to be addressedis below-cost export sales, what does it matterwhether there are viable comparison markets, or

comparable comparison-market products, orabove-cost comparison-market products? Thosecriteria are completely irrelevant and thereforeshould not be preconditions for comparing exportprices to constructed value.

Accordingly, the Antidumping Agreementshould be amended to revise the criteria for useof constructed value:

Reform Proposal 5: Article 2 of theAntidumping Agreement should berevised to provide for two alternativebases of normal value: price-to-pricecomparisons of export and home-market sales (unmodified by any costtest in the home market) and cost ofproduction (known in the UnitedStates as constructed value). Whichbasis is used will depend on the formof dumping alleged by the petitioner.In its petition the domestic industrywill allege either price-discriminationdumping or below-cost dumping,with appropriate corroborating evi-dence. If a price-discrimination case isinitiated, normal value will be basedon home-market prices; if a below-cost case is initiated, normal value willbe based on cost of production.

Furthermore, if the goal is to determinewhether export sales are below cost, thenexport prices should be compared to actual unitcosts of production—not cost plus profit. Anex-factory export price that is lower than thecost of production plus profit indicates onlythat the export price is below a certain level ofprofitability. Yet supporters of antidumpinggenerally define dumping, not as “insufficient-ly profitable” sales, but as below-cost sales. Theremedy should target the problem.

It should be noted that the Bush adminis-tration’s recent position paper on trade remedyrules does define dumping more broadly toinclude “export pricing at levels below the costof production plus a reasonable amount forselling, general and administrative expensesand profit”40—a formulation that tracks thedefinition of constructed value under U.S. law.

17

If the goal is todetermine whetherexport sales arebelow cost, thenexport prices shouldbe compared toactual unit costs ofproduction—notcost plus profit.

In support of this more expansive definition, itcan be argued that a “normal” profit is part of acompany’s cost of capital. In other words, acompany earning a subnormal return is sellingbelow its full economic costs, if above its fullaccounting costs. Nevertheless, the Bushadministration’s definition goes beyond theprevailing characterization of dumping—andis unwarranted in doing so, since any claim thatlow profitability is evidence of market distor-tions is much weaker than is the case withrespect to outright losses. Determining exactlywhat constitutes a normal rate of profit for agiven company in a given industry at a giventime is significantly more difficult than deter-mining whether or not that company is losingmoney. Moreover, low profits are generally sus-tainable over much longer periods than areoutright losses. Persistent failure to earn com-petitive returns can undermine a company’sability to make necessary investments andthereby may lead eventually to outright losses;it may also threaten the employment securityof the company’s management. Unlike sus-tained losses, though, low profitability in and ofitself does not imperil a company’s solvencyand future as a going concern. Accordingly,even chronically low profits are much less sug-gestive of “artificial” market conditions causedby government policies than is genuine red ink.

The present inclusion of profit in constructedvalue serves to inflate dumping margins inappro-priately.41 In 5 of the 18 actual U.S. dumpingdeterminations that we examined, constructedvalue was used for at least some product compar-isons. In 4 of those 5, the profit element influ-enced the outcome. Had profit not been added toconstructed value, the average calculated dump-ing margin would have been 11.02 percent lower(Table 1). In particular, the margin reductionwould have been 22.80 percent in an investiga-tion involving concrete reinforcing bars fromMoldova, 18.25 percent in an investigation of sta-tic random access memories (SRAMs) fromTaiwan, and 13.91 percent in an investigationconcerning dynamic random access memories(DRAMs) from Taiwan.

Those findings agree with earlier findingsmade by one of the coauthors. In the 1998 U.S.

investigation of preserved mushrooms from India,the Commerce Department calculated a dumpingmargin of 7.94 percent for Dieng/Surya Jaya; ifprofit had not been included, the margin wouldhave fallen to 4.88 percent—39 percent lower. Inthe 1997 U.S. investigation of cut-to-length steelplate from China, Liaoning’s dumping margin wasfound to be 17.33 percent; it would have been only5.43 percent—69 percent lower—if profit had notbeen added to constructed value.42

The Antidumping Agreement shouldtherefore be amended as follows:

Reform Proposal 6: Article 2.2 of theAntidumping Agreement should berevised to exclude profit from the calcu-lation of cost of production (known inthe United States as constructed value).

If negotiators are unable to make this reform,an alternative, “second best” proposal is to revise theway profit is calculated. Under current practice,profit is usually calculated on the basis of above-costcomparison-market sales only. Clearly, if an estimat-ed amount for profit must be included in con-structed value, then limiting consideration to onlythe profitable sales distorts the actual profitabilitypicture, artificially inflating profits—and thereforenormal value, and therefore dumping margins.Current practice can result in absurdly highamounts for profit. For example, in the 1997 U.S.investigation of melamine institutional dinnerwarefrom Taiwan, Chen Hao Taiwan was given a prof-it rate of 25.77 percent; in the parallel investigationof dinnerware from Indonesia, PT Multi Raya wasgiven a profit rate of 22.61 percent. The averageprofit rate for the U.S. plastic products industry, bycontrast, was only 5.23 percent.43 The methodolo-gy of using only above-cost sales for calculation ofprofit routinely yields such absurd results.

Accordingly, if constructed-value profit isretained, its calculation should be reformed:

Reform Proposal 7: If profit is notexcluded altogether from the calculationof cost of production, it should be basedon actual representative profit rates forthe subject merchandise. Specifically,profit rates should be based on average

18

The present inclu-sion of profit in

constructed valueserves to inflate

dumping marginsinappropriately.

industry-wide profit rates derived frompublic sources. In any event, profit shouldnever be calculated on the basis of theforeign producer’s (or anyone else’s)above-cost sales only.

While any inclusion of profit in constructedvalue is inconsistent with the basic concepts,principles, and objectives of the AntidumpingAgreement, the reform outlined above would atleast curtail some of the most egregious abusescreated by this methodological distortion.

Eliminate Use of Third-Country Sales inCalculating Normal Value

Article 2.2 of the present AntidumpingAgreement provides for the use of third-countrysales as the basis of normal value under specifiedcircumstances—in particular, when the foreignproducer under investigation does not sell thesubject merchandise in its home market or sellsin insufficient volumes there to “permit a propercomparison.” Although expressly allowed undercurrent antidumping rules, a comparison ofthird-country prices to export prices has norational relation to the basic concepts, principles,and objectives of the Antidumping Agreement.

Dumping, once again, is supposed to con-sist of either international price discriminationthat reveals the existence of a sanctuary marketor below-cost sales that reveal some underly-ing, market-distorting government policies. Acomparison of export and third-country pricesis incapable of identifying either phenomenon.First, and most obviously, a comparison ofprices in the export market under investigationto prices in other export markets says nothingabout whether the investigated sales are belowcost. Second, while differences betweenexport-market prices and third-country pricescan possibly show international price discrimi-nation, they cannot reveal a sanctuary market.Any foreign producer under investigation is an“outsider” as far as all third-country markets areconcerned; it is hindered, not helped, by anygovernment barriers that block access to itsexport sales. If for some reason the company isearning higher prices in that third country, thereason clearly is not that government-imposed

barriers are shielding it from competition. Onthe contrary, it had to overcome any barriersthat were present in that third-country marketto be selling there at all. Meanwhile, pricescharged in a third country indicate nothingabout whether a firm’s home market is closed.

Since a comparison of export-market andthird-country prices cannot possibly identify anyof the practices supposedly targeted byantidumping policy, it follows that third-countryprices are an inappropriate basis for normal value.Accordingly, the Antidumping Agreement needsto be revised along the following lines:

Reform Proposal 8: Article 2.2 of theAntidumping Agreement should beamended to provide that, in theabsence of sufficient home-marketsales, there is no basis for an allegationof price-discrimination dumping.

After this reform, the petitioning domesticindustries would still be able to allege below-cost dumping in those cases where there areinsufficient home-market sales. But it wouldno longer be possible to allege price-discrimi-nation or sanctuary-market dumping when thealleged sanctuary market doesn’t even exist.

Prohibit “Zeroing”The practice of “zeroing” is one of the most

notorious distortions in current antidumpingmethodology.44 It occurs in the final dumpingdetermination, when the foreign producer’sexport prices (whether individual transactions ormodel-specific averages) are compared to nor-mal value (usually average prices of comparablehome-market merchandise). When normalvalue is higher than the export price, the differ-ence is treated as the dumping amount for thatsale. When, however, the export price is higher,the dumping amount is treated as equal to zero.All dumping amounts are then added and divid-ed by the aggregate export sales amount to yieldthe company’s overall dumping margin.

Zeroing thus eliminates “negative dumpingmargins” from the dumping calculation. In sodoing, it can create dumping margins out ofthin air. Consider the results of the 18 U.S.

19

The practice of“zeroing” is one ofthe most notoriousdistortions in cur-rent antidumpingmethodology.

dumping determinations that we were able toexamine. All of the 17 determinations involv-ing market economies had margins inflated byzeroing. In 5 of the cases, the overall dumpingmargin would have been negative. On average,the margin results of the 17 cases would havebeen 86.41 percent lower if zeroing had notbeen employed (see Table 1).

The practice of zeroing has been found toviolate the current Antidumping Agreement.In a case brought by India against theEuropean Union involving bed linen, theWTO Appellate Body ruled in March 2001that the EU’s practice was WTO-inconsis-tent.45 The European Union has since changedits practice as a consequence of the AppellateBody’s ruling, but it still has not abandonedzeroing completely.46

The practice of zeroing continues unabatedin other jurisdictions, most notably the UnitedStates. The U.S. Department of Commercehas thus far refused to alter its practice, dis-missing the EU–Bed Linen case on the groundthat the United States was not a party. Giventhe EU’s continued (if limited) use of zeroing,the United States’ complete intransigence, andthe need generally to provide certainty forworldwide antidumping practice, a revision ofthe Antidumping Agreement to expressly pro-hibit zeroing is called for:

Reform Proposal 9: Article 2 of theAntidumping Agreement should berevised to clarify that the practice ofzeroing is prohibited. Specifically,when calculating dumping margins,negative dumping amounts (i.e.,instances in which export prices arehigher than normal value) should betreated as such and given their fullweight in the calculation of the foreignproducer’s overall dumping margin.

Especially in light of the Appellate Body’s deci-sion in the EU–Bed Linen case, this reform wouldbe in full accord with the basic concepts, principles,and objectives of the Antidumping Agreement.Any measurement of international price differ-ences that recognizes differences in only one direc-

tion and systematically ignores differences in theother direction has no methodological validitywhatsoever. Eliminating zeroing would help toensure that the price differences targeted byantidumping remedies actually exist in reality andare not just artifacts of skewed methodologies.

To close the door on zeroing completely, itwould be helpful to supplement the reformproposal above as follows:

Reform Proposal 10: Article 2.4.2 ofthe Antidumping Agreement shouldbe amended to require that, in bothoriginal investigations and administra-tive reviews, dumping margins must becalculated on the basis of comparingaverage export prices to average nor-mal values or else transaction-specificexport prices to transaction-specificnormal values. Comparisons of indi-vidual export prices to average normalvalues are never allowed.

Under the current Antidumping Agreement,Article 2.4.2 provides that “in the investigationphase” dumping margins shall “normally” beestablished on the basis of comparing averageprices to average prices or transaction-specificprices to transaction-specific prices. It does,however, allow comparisons of individual exportprices to average normal values “if the authori-ties find a pattern of export prices which differsignificantly among different purchasers, regionsor time periods.” The ostensible purpose of thisexception is to address instances of so-called tar-geted dumping, in which unfairly low prices tospecific customers or regions or at specific timesare masked by higher prices otherwise.

The current wording of Article 2.4.2 createspossible openings for the continued use of zero-ing notwithstanding the Appellate Body’s opin-ion in the EU–Bed Linen case. As explainedabove (see endnote 46), the Appellate Bodyconcluded that zeroing is WTO-inconsistentbecause it prevents true average-to-averagecomparisons as called for by Article 2.4.2. Thisreasoning leaves open the possibility that zero-ing may be permissible when dumping is calcu-lated another way. Indeed, since the agreement

20

Eliminating zeroingwould help to

ensure that the price differences

targeted byantidumping reme-

dies actually exist in reality and are not just artifacts

of skewed methodologies.

explicitly allows individual-to-average compar-isons under certain circumstances, and sincethose comparisons would yield exactly the sameresults as average-to-average comparisons unlesszeroing is employed for the former, there is aplausible argument that zeroing is implicitlypermitted under current WTO rules wheneverindividual-to-average comparisons are allowed.

Thus, zeroing may be consistent withArticle 2.4.2 as currently worded in targeteddumping cases. That is the EU’s position atpresent.47 In addition, the United States claimsthat Article 2.4.2’s reference to “the investiga-tion phase” implies that average-to-averagecomparisons are the rule only in investigations,and that individual-to-average comparisons areallowed generally in administrative reviews.Accordingly, the U.S. Department ofCommerce currently uses average-to-averagecomparisons in investigations and individual-to-average comparisons in reviews—withzeroing used in both methodologies.48 Thus,even if the United States eventually loses aWTO challenge along the lines of the EU–BedLinen case, it could continue to argue thatzeroing remains permissible, not only in target-ed dumping situations during original investi-gations, but also in all administrative reviews.49

A clear rule against individual-to-averagecomparisons under any circumstances is neces-sary to eliminate all uncertainty and ensure thatzeroing is completely abolished. As to concernsabout targeted dumping, allowing the imposi-tion of customer-specific antidumping dutyrates would be preferable to the exaggeration ofoverall duty rates that zeroing causes.

Eliminate Asymmetric Treatment of IndirectSelling Expenses

Dumping calculations are not based on acomparison of actual sales prices in the com-parison and export markets. Rather, antidump-ing authorities perform numerous adjustmentsto actual sales prices and then compare theadjusted “net” prices. The adjustments aredesigned to produce “apples-to-apples” com-parisons by taking into account differences intransportation costs, physical characteristics,credit terms, warranty terms, and other selling

expenses. If, however, adjustments are madeasymmetrically—that is, subtractions are madefrom the export price but not from the home-market price—dumping margins can be gener-ated out of thin air.

At present a glaring asymmetry exists in thetreatment of indirect selling expenses in “con-structed export price” situations. Under U.S. law,indirect selling expenses are expenses that do notvary directly with the volume of sales—sales staffsalaries, sales department overhead, and the like.In export price situations—that is, when the for-eign producer sells directly to an unrelated pur-chaser in the export market—no adjustment ismade to export or home-market prices for suchindirect selling expenses. But in “constructedexport price” situations—when the foreign pro-ducer sells to unrelated customers through a relat-ed reseller in the export market—certain indirectexpenses are deducted. Specifically, all indirectselling expenses incurred with specific respect tothe export market are deducted from the exportprice, but the adjustment to home-market pricefor home-market-related indirect selling expens-es is capped at the amount of the export-marketindirect selling expenses. All home-market-relat-ed indirect selling expenses in excess of the cap aresimply disregarded.

There is no possible justification for thisasymmetry. The policy of deducting export-market indirect expenses is apparently based onthe assumption that resales by the reseller in theexport market are on a different level of trade—and therefore that those resale prices includeadditional expenses—than are direct sales by theforeign producer in the home market. Thatassumption, though, is completely arbitrary. Itmay be that the reseller’s customers are largenational distributors, while the foreign producersells directly to small local wholesalers—inwhich case the home-market price actually hasmore of the distribution chain built into it thandoes the export price. Why then should theadjustment for home-market indirect sellingexpenses be limited to the amount of similarexpenses incurred in the export market?

This asymmetry—known as the “CEP off-set cap”—skews dumping calculations in thedirection of higher dumping margins.50 If

21

At present a glaringasymmetry exists inthe treatment ofindirect sellingexpenses in “con-structed exportprice” situations.

export-market indirect expenses are greaterthan equivalent home-market expenses, thenboth are fully taken into account; if, however,the home-market expenses are greater, they arecapped. The result in that case is an artificiallyinflated normal value—and an artificiallyinflated dumping margin.

As shown in Table 1, 10 of the 18 actual U.S.cases we examined would have had lowerdumping margins if the asymmetry of the CEPoffset cap had been eliminated. For example, in2 reviews involving tapered roller bearings fromJapan (one review involved “large” bearings; theother involved “small” bearings), if no indirectselling expenses had been deducted on eitherside, the dumping margin in the large bearingscase would have been 25.55 percent lower, andits rate in the small bearings case could havebeen 15.84 percent lower. For the 10 determina-tions involving CEP transactions, the averageeffect of eliminating the deduction of indirectselling expenses was to reduce the dumpingmargin by 9.06 percent.

It makes no sense to adjust prices automat-ically for indirect selling expenses. There is noreasonable basis for assuming such overheadcosts are built directly into the selling price.Stripping them out of the price, instead of cre-ating fairer price comparisons, produces pricecomparisons even more removed from theactual market reality of real sales prices. Thisartificiality is exacerbated by the asymmetry ofalways deducting all export-market indirectselling expenses while only partially deductingequivalent home-market expenses. The properapproach would be to end the automaticdeduction of any indirect selling expenses fromeither the constructed export price or thehome-market price.

At present the Antidumping Agreement issilent on the specific issue of the CEP offsetcap (although Article 2.4 does require a “faircomparison” between export price and normalvalue). Accordingly, supplemental language isnecessary to make clear that this arbitrary andasymmetric distortion is prohibited:

Reform Proposal 11: Article 2.3 of theAntidumping Agreement (which deals

with constructed export price situa-tions) should prohibit the automaticdeduction of indirect selling expensesfrom either the constructed exportprice or the comparison-market price.In situations where the related reseller’ssales in the export market are deemedto take place on a different level oftrade from sales in the home market,differences in indirect selling expensesmight serve as one possible basis forquantifying the appropriate level-of-trade adjustment.

This reform would eliminate the asymmetry ofthe CEP offset cap by eliminating indirect-selling-expense deductions altogether in nor-mal cases. As a result, price comparisons wouldbe fairer and more realistic—and the chancesof penalizing firms for normal commercialconduct simply because of flawed methodolo-gies would be correspondingly reduced.

Revise the Arm’s-Length TestThe Antidumping Agreement is currently

silent about the treatment of comparison-marketsales to affiliated parties.51 As a result, nationalantidumping authorities have a great deal of lati-tude in addressing this issue—with correspond-ing potential to skew margins significantly.

The practice in the United States has been togive special scrutiny to comparison-market salesmade to customers that are judged to have anaffiliation with the foreign producer.52 Pursuantto what is known as the “arm’s-length test,” theU.S. Department of Commerce includes affili-ated sales in the calculation of normal value onlyif they meet certain standards.

The purpose of the test is to determinewhether sales to affiliated customers in the com-parison market have been made at prices and onterms comparable to those granted to unaffiliatedcustomers. The test involves comparisons of theaverage net selling prices per product for eachaffiliated customer to the average net prices perproduct to all unaffiliated customers. Ratios arecalculated for each unique combination of affili-ated customer and product—provided that theproduct was also sold to at least one unaffiliated

22

It makes no senseto adjust prices

automatically forindirect selling

expenses.

customer—where the affiliated net price is thenumerator and the average unaffiliated net priceis the denominator. From these individual ratiosfor each affiliate, a weighted-average ratio is cal-culated. If that ratio is equal to or greater than99.5 percent, then all sales to the affiliate areaccepted as having been made at arm’s length.Otherwise, all sales to that affiliate are excludedfrom the calculations of average comparison-market prices.53 The apparent theory behind thispractice is that affiliated customers may receivemore favorable sales terms than do unaffiliatedcustomers, and thus the test is designed to findand eliminate sales to affiliated customers thathave paid lower prices on average.

The current U.S. arm’s-length test is seri-ously flawed.54 First of all, it is asymmetrical,since it ignores the equally likely possibilitythat prices to affiliates might be higher thanthose to unaffiliated customers. If the affiliatesare seeking to maximize their combined wel-fare (which is the implicit premise of the arm’s-length test), it might be optimal to show high-er revenues for the seller (to attract investors,improve the stock value, etc.) and higher costsfor the buyer (to avoid tax liabilities, etc.). Yet,only when sales to affiliates are lower than salesto unaffiliated customers are they excluded.The current test’s asymmetry has the effect ofraising average prices in the comparison mar-ket, and hence raising dumping margins.

Not only is the asymmetric application ofthe test inappropriate, but because it ignoresthe reality that prices vary within and betweenindustries and over time, the 99.5 percentthreshold is simply arbitrary. While in someindustries or in some years, there may be verylittle price variation, prices may vary widely inother situations. What this suggests is that thethreshold for deviations from the average priceshould be more liberal when wider price varia-tion is normal. A constant benchmark makesno sense in light of this inherent variability.And why should ratios below that level be con-strued as evidence that prices were manipulat-ed to lower normal value anyway? Could therenot be other reasons? If an arm’s-length test isto be used at all, the relevant threshold shouldreflect the range of prices offered in that par-

ticular market, industry, or time—and it shouldbe applied symmetrically.

As evidenced in Table 1, the arm’s-lengthtest can severely inflate dumping margins.Eliminating the test from the dumping calcu-lation affected the results in 8 of the 13 cases inwhich the test was used (affiliated sales werenot an issue in 5 of the 18 cases), with an over-all average reduction in the dumping margin of6.95 percent. In each of the 8 cases affected, themargin decreased. In an investigation of hot-rolled steel from Japan, the company’s dump-ing margin decreased by 15.97 percent.55 Themargin decreased by 52.60 percent for a com-pany in a review concerning stainless steelsheet and strip from Japan.

Some reform of the U.S. test is necessary evenunder the current Antidumping Agreement. Inthe Japanese Hot-Rolled Steel case, the WTOAppellate Body ruled that the U.S. test’s asym-metry is inconsistent with WTO rules. 56 Inresponse to the WTO’s ruling, the U.S.Department of Commerce is now proposing tochange the test. Under the new approach, all salesunder a 98 percent threshold and above a 102percent threshold would be eliminated.57 Theproposed new “band” approach eliminates thearm’s-length test’s asymmetry, but the fixed 98percent and 102 percent thresholds, like the cur-rent 99.5 percent threshold, are simply arbitrary.

However the United States ultimatelyresponds to the WTO’s ruling, there is a need torevise the Antidumping Agreement and clarifythe standards for dealing with sales to affiliatedparties. Since evidence of price manipulation isnearly impossible to ascertain by comparing aver-age prices, and since the presumption that manip-ulation is evidenced by slightly lower prices toaffiliates is patently unfair, the AntidumpingAgreement should be revised to instruct nationalauthorities to simply eliminate all abnormallyhigh-priced and low-priced sales in the compari-son market to affiliated parties from the calcula-tion of normal value. And rather than allow“abnormally” to be defined subjectively by eachnational authority, valid statistical methodsshould be required. For example, average (mean)prices might be calculated along with the stan-dard deviations, which reflect the variance of each

23

If an arm’s-lengthtest is to be used atall, the relevantthreshold shouldreflect the range ofprices offered inthat particular mar-ket, industry, ortime—and itshould be appliedsymmetrically.

individual price from that average. Those sales atprices exceeding the average plus two (or three)standard deviations and at below the averageminus two (or three) standard deviations shouldbe dropped. Such an approach would reduce theskewing impact of aberrational comparison-mar-ket sales on the overall picture of that market.

Thus, the Antidumping Agreement shouldbe revised along the following lines:

Reform Proposal 12: Article 2 of theAntidumping Agreement should berevised to require that the exclusion ofcomparison-market sales to affiliatedparties be administered in a symmet-rical manner to both higher and lowernon-arm’s-length sales terms.Furthermore, Article 2 should requirea statistically sound methodology fordetermining whether or not salesprices are at arm’s length.

Special Consideration for “Off-Quality” or“Secondary” Merchandise

The Antidumping Agreement is currentlysilent on the treatment of “off-quality” or “sec-ondary” merchandise. This silence needs to beremedied, since export sales of off-quality mer-chandise now tend to generate large dumpingmargins—not because of any unfair trade, butsimply as a result of methodological flaws indumping calculations.

Production processes do not always yieldexactly what was intended. If the output fails tomeet desired specifications or tolerances, it isconsidered off-quality or secondary merchan-dise. Especially in industries where productspecifications reflect inflexible safety standardsor engineering requirements, off-quality outputoften cannot be sold for its intended purpose.Nevertheless, secondary merchandise may havesome value to a customer, as an input for a less-er product or as scrap material, for example. Inthese cases the producer is better off selling thesecondary merchandise rather than simplyscrapping it, but it usually must sell the off-qual-ity products at a substantial discount.

Under current antidumping rules, sales of sec-ondary merchandise in the export market almost

always give rise to high dumping margins. 58

Ideally, such sales would be compared to equiva-lent sales in the home market. Such comparisonsare usually precluded, however, by the operationof the cost test. Although secondary merchandisemust be sold at lower prices than the prime mer-chandise it was intended to be, it costs the sameto produce as prime merchandise. As a result, off-quality merchandise is almost always sold atprices below the cost of production. With allhome-market sales of off-quality goods eliminat-ed by the cost test, export sales of secondary mer-chandise must be compared to much higherpriced home-market sales of prime merchandise.

Antidumping rules do provide for priceadjustments for physical differences in products.When export sales are compared to home-mar-ket sales of nonidentical merchandise, a “differ-ence-in-merchandise,” or DIFMER, adjust-ment is made to compensate for physical differ-ences and thereby, presumably, ensure an“apples-to-apples” comparison. Unfortunately,however, DIFMER adjustments are typicallycalculated as the difference in the variable costsof the models being compared. While there arereal and important physical differences betweenprime and off-quality merchandise, there are nocost differences, and therefore no basis for aDIFMER adjustment. Accordingly, when off-quality export sales are compared to primehome-market sales, typically there is no adjust-ment of the large price difference between them,and the end result is a large dumping margin onthose sales.

Much of the problem with secondary mer-chandise would disappear if the cost test wereeliminated. Export sales of off-quality goodscould be compared to equivalent home-marketsales—provided such sales existed in the homemarket. But if the cost test is retained, some-thing needs to be done about this specificproblem. And even if the cost test is eliminat-ed, the problem still arises whenever there areexport sales of secondary merchandise but nocorresponding home-market sales.

Accordingly, even if the cost test is elimi-nated, but especially if it isn’t, the AntidumpingAgreement needs to be modified along the fol-lowing lines:

24

Under currentantidumping rules,

sales of secondarymerchandise in

the export marketalmost always give

rise to high dumpingmargins.

Reform Proposal 13: Article 2 of theAntidumping Agreement needs toinclude a special provision dealing withsales of secondary merchandise in theexport market. This provision shouldrequire national antidumping authori-ties either to (1) disregard sales of off-quality merchandise in their dumpingcalculations or (2) compare export salesof off-quality merchandise to corre-sponding comparison-market saleswithout regard to whether the compar-ison-market sales are below the fullcost of production. If export sales ofoff-quality merchandise are includedin the dumping calculation, but thereare no corresponding home-marketsales, comparisons of the export salesto comparison-market sales of primemerchandise should receive a specialDIFMER adjustment that reflects theaverage price difference between primeand secondary subject merchandise.

Under current rules, export sales of off-quality merchandise are virtually certain togenerate dumping margins—simply because ofquirks in dumping-calculation methodologies.As a result, companies are being punished fornormal commercial practices that have nothingto do with unfair trade under any plausible def-inition of that term. Fidelity to the basic con-cepts, principles, and objectives of theAntidumping Agreement requires that thisabuse be eliminated.

Tighten Standards on Causation of InjuryIn keeping with traditional practice, the cur-

rent Antidumping Agreement requires morethan simply a finding of dumping beforeantidumping remedies may be imposed. Inaddition, it requires a finding that dumpedimports are causing “material injury” (or threatof material injury) to a domestic industry. Thisinjury requirement is in keeping with the basicconcepts, principles, and objectives of theAntidumping Agreement—namely, that trade-restricting remedies be used to offset artificialcompetitive advantages caused by underlying

market distortions. If imports are not materiallyaffecting the competing domestic industry, thenclearly there is no artificial competitive advan-tage to be offset—and thus no cause forantidumping remedies.

Unfortunately, implementation of the injuryrequirement is seriously flawed under currentantidumping rules. The chief problem is theabsence of clear standards for judging whetherthere is a causal link between dumped importsand injury to a domestic industry. Under U.S.law, the usual approach of the InternationalTrade Commission has been to engage in a so-called bifurcated analysis: first, determiningwhether a domestic industry is injured; and, sec-ond, determining whether subject imports con-stitute “a cause” of that injury. With thisapproach, assuming the domestic industry isdoing poorly and that import volumes are up (orat least substantial, even if flat or falling) andprices are down (or at least consistently lowerthan the domestic industry’s prices), it is alwayspossible for the commission to find that importshave made at least some contribution to thedomestic industry’s condition—and thus tomake an affirmative determination.

This approach has absolutely no analyticalrigor. Any coincidence of significant or risingimports and poor industry performance canserve as the basis for imposing antidumpingremedies. There are no standards for distin-guishing between mere coincidence and actualcausation; accordingly, the commission has vir-tually unchecked discretion to confuse the for-mer with the latter. The analysis of causation isa “black box”—there is no way of predictingwhen the commission will find injury, or even ofbeing sure that commissioners don’t secretlyresort to extrastatutory criteria in making theirdeterminations. That said, U.S. practice is amodel of transparency and high analytical stan-dards compared to what goes on in many otherantidumping jurisdictions.

At a bare minimum, WTO rules should beamended to require, as a necessary but not suf-ficient condition for finding injury, the exis-tence of a clearly established correlationbetween increased imports and decliningdomestic industry performance. Accordingly,

25

The analysis of cau-sation is a “blackbox”—there is noway of predictingwhen the commis-sion will find injury,or even of beingsure that commis-sioners don’t secret-ly resort toextrastatutory crite-ria in making theirdeterminations.

we propose a change in the AntidumpingAgreement along the following lines:

Reform Proposal 14: Article 3.5 of theAntidumping Agreement should berevised to provide that no affirmativeinjury determination shall be made inthe absence of a substantial correlationbetween increased imports during theperiod of investigation and decliningoperating profits for the domesticindustry during the correspondingperiod. The required increase inimports may take the form of either anabsolute increase in import volume or arelative increase (i.e., an increase inmarket share). In codifying thisrequirement, the AntidumpingAgreement should make clear that themere presence of such a correlation,standing alone, does not necessitate anaffirmative determination.

If imports have not increased (whether inabsolute or market-share terms) during the peri-od of investigation, there is no analytically tenablebasis for concluding that any woes suffered by thedomestic industry during that period are due toimports rather than some other factor. Althoughthere are many indicators of industry perfor-mance besides operating profits (including salesvolumes, average prices, investment, and employ-ment), operating profits go to the heart of anindustry’s well-being. Profit levels reflect bothvolumes and prices, and they have a direct impacton investment and employment. A bright-linerequirement of a substantial (i.e., statistically sig-nificant) correlation between increased importsand declining operating profits is therefore emi-nently sensible on the merits and has the addedadvantage of establishing some minimal analyti-cal transparency in the injury process.59

But appropriate standards for guiding injurydeterminations cannot stop here. It is notenough to show that imports could have beenresponsible for the domestic industry’s deterio-rating condition. It is necessary to push beyondmere correlation and require the establishmentof a causal link between imports and injury.

Current WTO rules do make some effortalong these lines. Specifically, under the “non-attribution requirement” of Article 3.5,antidumping authorities are required to “exam-ine any known factors other than dumpedimports which at the same time are injuring thedomestic industry, and the injuries caused bythese other factors must not be attributed to thedumped imports.” In a Japanese challenge to theU.S. antidumping investigation of hot-rolledsteel, the WTO Appellate Body made clearthat, under this provision, antidumping author-ities are not allowed to lump imports with otherfactors and determine that, collectively, all thefactors are causing injury. Rather, authoritiesmust disentangle imports from other factors andjudge their injurious effects separately:

We recognize, therefore, that it maynot be easy, as a practical matter, toseparate and distinguish the injuriouseffects of different causal factors.However, although this process maynot be easy, this is precisely what isenvisaged by the non-attribution lan-guage. If the injurious effects of thedumped imports and the other knownfactors remain lumped together andindistinguishable, there is simply nomeans of knowing whether injuryascribed to dumped imports was, inreality, caused by other factors. Article3.5, therefore, requires investigatingauthorities to undertake the process ofassessing appropriately, and separatingand distinguishing, the injuriouseffects of dumped imports from thoseof other known causal factors.60

The Appellate Body’s interpretation of thenonattribution requirement, though, threatensto lead antidumping investigations intointractable factual and analytical difficulties. Intwo opinions dealing with causation of injuryin the context of the Agreement on Safeguards,the Appellate Body ruled that, once an admin-istering authority isolates the injurious effectsof imports from those of other causal factors,the authority need not find that increased

26

It is necessary topush beyond mere

correlation andrequire the estab-

lishment of a causallink between

imports and injury.

imports alone are causing or threateninginjury. 61 Rather, the authority need find only a“genuine and substantial relationship of causeand effect” between imports and injury thatmay be the result of many causes.62

How this murky formulation should beapplied in practice is unclear. What theAppellate Body’s standard appears to involve issome weighing of different causal factors—that is, assigning relative importance to all thevarious causal factors and then determiningwhether imports alone contribute “enough” tothe combined injurious effect. In many cases,however, such a task would be so analyticallydaunting as to be impracticable.

There is an easier and better way. We suggestthat the approach rejected by the AppellateBody in the context of the SafeguardsAgreement be made an explicit requirement inthe Antidumping Agreement:

Reform Proposal 15: Article 5.3 of theAntidumping Agreement should berevised to require that antidumpingauthorities must find that dumpedimports, considered alone, are causingmaterial injury or threat thereof.

How is this requirement to be administeredwhen there are multiple causal factors involved?The proper approach is something along the linesof the so-called unitary causation analysis, whichwas used in the past by some members of the U.S.International Trade Commission. Specifically,antidumping authorities should use basic tools ofquantitative economic analysis to determinewhether the domestic industry is materially worseoff because of dumped imports—or, in otherwords, whether application of the proposedantidumping measures would make the domesticindustry materially better off.

To make this determination, the administer-ing authority would need to estimate the substi-tutability of subject and nonsubject imports,subject imports and domestic production, anddomestic production and other goods. If non-subject imports substitute easily for subjectimports, then the effect of antidumping reme-dies will be limited, since nonsubject imports

will simply fill the place formerly occupied bysubject imports. If subject imports and domesticproduction are not good substitutes (for exam-ple, if there are quality differences or other formsof product differentiation), then the injuriouseffect of subject imports on the prices of domes-tically produced merchandise will be attenuated.And if other products are good substitutes forthe subject merchandise, the effect of antidump-ing remedies will again be limited, becausedomestic producers will be constrained fromincreasing prices because of competition fromsubstitute goods.

Here then is an analytically sound approachto determining causation of injury. It is straight-forward and administrable: making reasonableestimates of substitutability is far easier thanassessing individually several different causalfactors and assigning to each some level of cul-pability for an industry’s condition. Thisapproach opens up the black box of currentinjury analysis and makes clear exactly whatconditions of competition need to be assessed.

A unitary-style approach to causation is notmechanical: it requires judgment calls and inter-pretation of evidence. Accordingly, it does noteliminate controversy or conflicting interpreta-tions. Nevertheless, it at least specifies the factorsthat bear on causation and makes them transpar-ent. Revising the Antidumping Agreement torequire such an approach would be a dramaticstep toward making the injury requirement oper-ational in an a consistent, administrable, andintellectually credible manner.

Change Standards for “Negligibility”Under the current Antidumping Agreement,

national antidumping authorities are authorizedto “cumulate” imports from multiple countriesfor purposes of making an injury determination.In other words, authorities can group togetherthe imports from some or all countries underinvestigation and determine whether the com-bined effect of those imports is to cause orthreaten injury. Consequently, imports from aparticular country are frequently subject toantidumping duties even though those imports,considered alone, were never found to havecaused any harm.

27

The properapproach is some-thing along thelines of the so-called unitary cau-sation analysis,which was used inthe past by somemembers of theU.S. InternationalTradeCommission.

There are understandable reasons for allow-ing cumulation to some degree. Without it, arti-ficial competitive advantages caused by marketdistortions might go completely unremediedsimply because no one import source, on itsown, is considered injurious. On the other hand,dumping is a company- and country-specificphenomenon: the artificial competitive advan-tages targeted by antidumping policy supposed-ly accrue to particular companies and arise out ofgovernment policies in those companies’ partic-ular home markets. Accordingly, if particularcompanies or even whole countries are suchminor players in an export market that they haveno significant impact on competitive conditionsin that market, then they cannot be said to enjoyany real competitive advantage vis-à-vis thedomestic industry—and therefore cannot beproper targets of antidumping remedies.

Current antidumping rules balance thesecompeting considerations by prohibiting thecumulation of “negligible” imports—importsfrom countries whose combined market sharefalls below a designated threshold. Specifically,under Article 5.8 of the present AntidumpingAgreement, imports from a particular countryare considered negligible if they amount to lessthan 3 percent of total imports of the productunder investigation—unless all the countriesunder investigation that individually fall underthe 3 percent threshold together account formore than 7 percent of total imports.

While the general approach of allowingcumulation except for negligible imports seemsbasically sound, the current threshold fordetermining negligibility is indefensible.Specifically, determining negligibility on thebasis of percentage of total imports makes nomethodological sense. What matters iswhether the arguably negligible imports arecapable of contributing meaningfully to theinjury being suffered by the domestic indus-try—that is, whether they can be said to enjoyany real competitive advantage relative to thedomestic industry. The proper criterion forjudging this question is, not share of totalimports, but share of the overall export market.

Consider the difference between the two cri-teria in the context of a recent U.S. antidumping

investigation. In the investigation of hot-rolledsteel from Argentina, China, India, Indonesia,Kazakhstan, Netherlands, Romania, SouthAfrica, Taiwan, Thailand, and Ukraine, five coun-tries were considered and rejected for negligibili-ty exclusions because they exceeded the collectivethreshold: Argentina, 1.74 percent of imports;Kazakhstan, 2.78 percent; South Africa, 2.26 per-cent; Thailand, 2.40 percent; and Ukraine, 2.65percent. Although each of those countries fellbelow the normal negligibility threshold of 3 per-cent, collectively they accounted for 11.80 percentof imports—and thus their products did notqualify collectively as negligible imports.

In this particular case, however, all importsfrom all sources accounted for only 26.4 percentof the total U.S. merchant market for hot-rolledsteel and only 11.15 percent of total U.S. domes-tic consumption of hot-rolled steel. Accordingly,the import sources in question had the followingmarket shares: Argentina, 0.46 percent of themerchant market (0.19 percent of total domes-tic consumption); Kazakhstan, 0.74 percent(0.31 percent); South Africa, 0.60 percent (0.25percent); Thailand, 0.63 percent (0.27 percent);and Ukraine, 0.70 percent (0.30 percent). Thosecountries thus had a combined market share ofonly 3.13 percent—or 1.32 percent, dependingon how market share is measured for this partic-ular industry. 63 This level of combined importpenetration cannot seriously be considered evi-dence of an unfair competitive advantage—yetunder current antidumping rules such importscan be swept into the maw of a multicountryantidumping duty order.

Furthermore, judging negligibility on thebasis of import share makes the standard a vari-able one—indeed, one that varies in perverseways. The higher the total level of import pene-tration, the greater the market share an importsource can gain and still be considered negligi-ble—despite the fact that overall high importpenetration presumably means that the domes-tic industry is more vulnerable. Meanwhile, thelower the overall import penetration, the small-er the volume of imports that exceeds the negli-gibility cutoff—even though the domesticindustry is presumably less affected by foreigncompetition under such circumstances.

28

While the generalapproach of allow-

ing cumulationexcept for negligibleimports seems basi-cally sound, the cur-

rent threshold fordetermining negligi-bility is indefensible.

Accordingly, Article 5.8 of the AntidumpingAgreement should be revised along the followinglines:

Reform Proposal 16: Article 5.8 of theAntidumping Agreement should berevised to change the threshold for neg-ligibility from 3 percent of total importvolume (and 7 percent collectively) to 2percent of domestic consumption (and 5percent collectively).

This proposal would preserve the negligi-bility rule’s tradeoff between cumulation andexemption of small import sources. However,by redefining the criterion for judging negligi-bility, it would decrease the likelihood thatimports that by reason of small volume cannotmaterially injure a domestic industry getunfairly tangled in protracted antidumpingproceedings. It would also discourage theincreasingly popular but abusive “shotgunapproach” to filing antidumping petitions, inwhich domestic industries pile up allegationsagainst many small exporting countries, manyof which are of no competitive concern what-soever, just for the purpose of pushing past thecollective negligibility threshold of 7 percent.

Raise Initiation StandardsWe have already proposed a major change

in the process of initiating antidumping inves-tigations—namely, that credible evidence ofunderlying market distortions be made arequirement for initiation (see ReformProposals 1 and 2). Even if such fundamentalreform of the definition of dumping is notundertaken, there are less far-reaching but stillvaluable improvements that can be made in theinitiation process. Here again, the guidingprinciple of reform should be to improveantidumping’s aim and limit disruption of nor-mal commercial conduct.

Lax initiation standards can be the cause ofsignificant disruptions. This is so because mereallegations of dumping can wreak havoc withtrade patterns. Under U.S. law, for example,importers are responsible for paying antidumpingduties. Thus, from an importer’s perspective, the

initiation of an antidumping investigation raisesthe prospect of significant extra costs in the formof duty liabilities—a prospect that manyimporters, quite understandably, are anxious toavoid. Consequently, the mere act of launching anantidumping investigation tends to depressimports from investigated countries. And sinceinvestigations last for approximately one year, sig-nificant damage can be done to a foreign produc-er even if it is ultimately cleared of all charges.64

Under current rules, it is too easy to launchantidumping cases. Here in the United States,innocent companies are frequently harassed byill-founded charges and unnecessary investiga-tions. Consider, for example, the fact that about35 percent of U.S. cases result in findings of noinjury or no dumping. In other words, evenassuming that all affirmative dumping andinjury findings are justified, more than one-third of all U.S. investigations result in year-long disruptions in fairly traded imports’ accessto the U.S. market.

Furthermore, it is instructive to comparedumping margins alleged in U.S. antidumpingpetitions with the dumping margins actuallyfound by the U.S. Department of Commerce.Although the Commerce Department’s calcula-tions are rife with methodological distortionsthat inflate dumping margins, they generallyresult in dumping margins substantially lowerthan those alleged in the petition. For example,in original investigations during 2001 in whichadverse facts available were not used, Commercefound an average dumping margin of 38.18 per-cent. By contrast, the average dumping marginalleged in the petitions in those cases was 100.80percent—more than twice as high.65 It is clear,then, that the evidentiary quality of dumpingallegations in U.S. petitions is extremely low—yet it nonetheless passes muster with theCommerce Department.

Hair-trigger initiation of antidumping cases isby no means an exclusively American problem. Forexample, a WTO dispute settlement actionbrought by Mexico against Guatemala addressedthe issue of lax initiation standards. At the center ofthe dispute was Guatemala’s decision to initiate anantidumping investigation of grey portland cementfrom Mexico on the basis of laughably flimsy evi-

29

The mere act oflaunching anantidumping inves-tigation tends todepress importsfrom investigatedcountries.

dence. The only evidence of dumping provided inthe petition was two sets of invoices: two Mexicaninvoices for one bag of cement each and twoinvoices for Guatemalan imports from Mexico ofthousands of bags each. Given the glaring differ-ences in sales volumes and levels of trade of the twosets of invoices, it should have been obvious thatthis documentation provided no evidence ofdumping. Nevertheless, Guatemalan authoritiesinitiated the case. Meanwhile, the two importinvoices also served as the only evidence in supportof the petitioner’s allegation of threatened injury.

In this particular case, the WTO found thatGuatemala’s initiation of an investigationunder these circumstances was improper.66 Wehave no way of knowing, however, how fre-quently antidumping authorities around theworld are initiating investigations in a similar-ly slipshod manner but are never held toaccount before the WTO.

The problem is that the current AntidumpingAgreement provides no standards at all to con-strain the initiation of bogus cases. Article 5.2 ofthe agreement requires antidumping petitions toprovide evidence of dumping, injury, and acausal link between the two; the agreement alsosays that a “simple assertion, unsubstantiated byrelevant evidence, cannot be considered suffi-cient.” Even this nebulous language was enoughto allow the WTO, after the fact, to ruleGuatemala’s egregious actions in the cementcase out of bounds, but it does nothing to restrictabuses before they happen. Clear rules are need-ed to limit authorities’ discretion or create incen-tives for responsible behavior.

Reform Proposals 1 and 2, which wouldrequire credible evidence of underlying marketdistortions before any investigation is initiated,would go a long way toward restricting baselessinvestigations. Those proposals, though, aim togo beyond mere initiation standards torethinking the basic question of how dumpingis defined. Even if such reforms are not adopt-ed in the short term, greater specification of theevidentiary requirements for dumping as tradi-tionally defined would help to limit abuses inthe initiation process. We therefore suggestchanges to the Antidumping Agreement alongthe following lines:

Reform Proposal 17: Articles 5.2 and 5.3of the Antidumping Agreement shouldbe revised to specify concrete evidentiarystandards for initiation. With respect toevidence of dumping, the petitionermust supply documentation on a compa-ny-specific basis of representative pricesof the subject merchandise sold by theforeign producer in the export marketand either (a) representative prices ofcomparable products sold by the foreignproducer in its home market or (b) credi-ble estimates of the foreign producer’scost of production. The petitioner mustsupply such company-specific evidencewith respect to at least four foreign pro-ducers or, alternatively, foreign producersaccounting for a significant portion (forexample, at least 40 percent) of subjectimports. With respect to evidence ofinjury and causation, the petitioner mustsupply documentation of trends in (a)subject import volumes (including mar-ket share); (b) prices in the export market;and (c) the domestic industry’s sales vol-umes (including market share), prof-itability, and employment.

For purposes of the above proposal, “representative”price data means prices of major products that arerepresentative of price levels throughout the periodof investigation. “Credible estimates” of productioncosts should, to the extent possible, be based on theforeign producer’s own data or at least data relatingto the foreign industry under investigation.

The requirements proposed above wouldnot stop the filing of antidumping petitions forharassment purposes. However, they would atleast bring some minimal discipline to the ini-tiation process and thereby afford some addi-tional protection against the harassment ofhealthy import competition. Such an improve-ment in antidumping policy would be in keep-ing with the basic concepts, principles, andobjectives of the Antidumping Agreement.

Mandate “Lesser-Duty Rule”Article 9.1 of the current Antidumping

Agreement states that it is “desirable” that

30

The currentAntidumping

Agreement providesno standards at all toconstrain the initia-tion of bogus cases.

antidumping duties “be less than the [dumping]margin if such lesser duty would be adequate toremove the injury to the domestic industry.” Thisexpress preference for the so-called lesser-dutyrule is in keeping with the basic concepts, princi-ples, and objectives of the AntidumpingAgreement. After all, the avowed purpose ofantidumping remedies is to restore a “level play-ing field”—in other words, to neutralize artificialcompetitive advantages created by market-dis-torting government policies. If a particular dutyrate is deemed sufficient to eliminate injury to thedomestic industry, there is no justification forimposing a higher rate; a higher rate exceeds themandate of creating a level playing field and slantsthe field in favor of the domestic industry.

A number of WTO members—including theEuropean Union—follow the approach recom-mended in Article 9.1 and apply a lesser-duty rulein their antidumping investigations. The basicapproach is to calculate “noninjurious prices”—prices for export sales that would not depress orsuppress the prices charged by the domesticindustry. The difference between the export priceand the noninjurious price is referred to as the“injury margin.” If the injury margin is greaterthan the dumping margin, then the antidumpingduty rate is equal to the dumping margin; if, how-ever, the injury margin is lower than the dumpingmargin, the lesser duty applies and is set at thelevel of the injury margin.

The lesser-duty rule can result in significantreductions in the antidumping duty rates thatwould otherwise apply. Consider the followingexamples of definitive duties imposed by theEuropean Union during 2000. In the investigationof seamless pipes and tubes from Croatia andUkraine, the authorities found final dumping mar-gins of 40.8 percent and 123.7 percent, respective-ly; application of the lesser-duty rule, though,brought the actual duty rates down to 23.0 percentfor Croatia and 38.5 percent for Ukraine—reduc-tions of 44 percent and 69 percent, respectively. Inthe investigation of hot-rolled steel from China,the final dumping margin came to 55.5 percent,but because of the lesser-duty rule the actual dutyrate was only 8.1 percent—an 85 percent reduc-tion. And in the investigation of black colorform-ers (i.e., dyes) from Japan, the dumping margin was

49.8 percent, but the final duty rate was only 18.9percent—or 62 percent lower—because of thelesser-duty rule.67

Because the language in Article 9.1 is notmandatory, WTO members are under no oblig-ation at present to adopt a lesser-duty rule. TheU.S. law, for instance, does not have such a rule.Among jurisdictions that do have some kind oflesser-duty rule, there is no consistency and littletransparency in the manner in which it is applied.As a result, antidumping duties in excess of thosethat can be justified by the basic concepts, princi-ples, and objectives of the AntidumpingAgreement are being imposed routinely.

To cure this glaring defect in current antidump-ing practice, we recommend a change in theAntidumping Agreement along the following lines:

Reform Proposal 18: Article 9.1 of theAntidumping Agreement should berevised to require that antidumpingduties be less than the dumping mar-gin if the lesser duty is sufficient toremove the injury to the domesticindustry. Specifically, antidumpingauthorities should be required to cal-culate noninjurious prices for exportsales, which would be at levels that donot depress or suppress the pricescharged by the domestic industry. Ifthe difference between the noninjuri-ous prices and the export prices(known as the injury margin) is lessthan the dumping margin, theantidumping duty should be set at thelesser rate equal to the injury margin.

Raise de Minimis to 5 Percent in Investigationsand Reviews

The calculation of dumping margins isplagued with methodological difficulties. Mostobviously, there are the distortions that tend toskew the analysis in favor of finding dumping:the failure to require evidence of underlyingmarket distortions, the cost test, the inclusion ofprofit in constructed value, the asymmetrictreatment of related-party sales, the asymmetrictreatment of indirect selling expenses, the use ofzeroing in dumping calculations, and so forth.

31

If a particular dutyrate is deemed suf-ficient to eliminateinjury to thedomestic industry,there is no justifica-tion for imposing ahigher rate.

Even if all those obvious distortions were elim-inated, though, the measurement of dumpingmargins would still be highly inexact. In the typicalinvestigation, antidumping authorities comparehome-market and export prices of physically dif-ferent goods, in different kinds of packaging, soldat different times, in different and fluctuating cur-rencies, to different customers at different levels oftrade, in different quantities, with different freightand other movement costs, different credit terms,and other differences in directly associated sellingexpenses (for example, commissions, warranties,royalties, and advertising). Is it really surprising thatthe prices compared are not identical? Wouldn’t itbe surprising if they were?

Admittedly, antidumping authorities try toadjust for some of those differences, but theadjustments are necessarily crude and impre-cise. For example, when the U.S. Departmentof Commerce compares physically differentmerchandise, it adjusts for differences in mate-rials, direct labor, and variable overhead costs.While this makes a certain amount of sense, ina real-world context it goes without saying thatactual price differences may be more or lessthan the differences in variable manufacturingcosts. As we pointed out in relation to second-quality merchandise, sometimes huge differ-ences in commercial value can exist withoutany measurable differences in manufacturingcosts. Similarly, the Commerce Departmentadjusts for differences in warranty terms on thebasis of differences in repair parts and laborcosts. While this approach is logical enough, itis still extremely unlikely that the actual real-world price differences between products withdifferent warranties are precisely equal to thedifferences in warranty costs.

And in many cases, antidumping authoritiesmake no adjustment at all. Thus, prices of goodssold in the export market may be compared toprices of goods sold many months earlier or laterin the home market, without any adjustment formarket fluctuations. And although unit pricestypically decline with larger order quantities, theU.S. Department of Commerce rarely adjustsfor quantity discounts.

Calculations of unit costs of production aresimilarly rife with more-or-less arbitrary guess-

work. Especially vexing is the allocation of sharedcosts. Consider, for example, coproducts or jointproducts—two or more different goods that areproduced simultaneously in the same manufac-turing process. Examples include different cuts ofmeat from the same animal, different oresextracted in the same mining operation, and dif-ferent chemicals produced by the same reaction.For such products, some allocation of sharedmanufacturing costs is necessary for cost-accounting purposes. But how costs are allocatedmay well determine whether a given coproductshows a profit or a loss. If costs are allocatedequally to high-value and low-value products (forexample, pig’s feet and pork chops), the low-valueproducts will always show a huge loss. On theother hand, if costs are allocated on the basis ofthe relative sales value of the coproducts, allcoproducts will show the same profitability—anequally arbitrary outcome.

The fact is that the accounting treatment isand should be irrelevant to proper businessdecisions. Managers should decide what mix ofcoproducts to target, not on the basis of arbi-trary unit costs, but on the basis of maximizingtotal net revenue. If a low-value product canfetch a price that covers the marginal costs offurther processing after “splitoff” from otherjoint products, the decision to engage in thatfurther processing and sell that low-valueproduct is profit-maximizing regardless of howcosts are allocated among all the joint products.

Joint products are manufactured from thesame raw materials, but there are many otherways for products to share costs. Sharing offactory overhead costs (for example, the costsof electricity, fuel, maintenance, plant andequipment depreciation, engineering support,research and development, selling, and generaland administrative expenses) is the norm inmultiproduct firms. Indeed, economists explainthe very existence of multiproduct firms interms of cost sharing. The ubiquitousness ofsuch cost sharing shows that unit costs mustalways be taken with several grains of salt. Aparticular product that is never profitable whenviewed in isolation may nonetheless contributeto fixed costs that would be incurred anyway onother, profitable products. Paradoxically, then, a

32

Even if obvious dis-tortions were elimi-

nated, the measure-ment of dumping

margins would stillbe highly inexact.

perennially money-losing product can help tomaximize firmwide profits.

In view of all these methodological chal-lenges, a healthy dose of humility is in orderregarding the accuracy of any dumping calcula-tions. Appropriately, such humility informs thecurrent Antidumping Agreement’s concept ofde minimis dumping margins. Dumping mar-gins below a certain threshold are deemed to bede minimis and are treated as equal to zero.Given the tension between the trade-restrictiveeffects of antidumping measures and the overallorientation of the WTO agreement towardmarket opening, a conservative policy of resolv-ing doubts against the imposition of duties isentirely fitting.

The de minimis rule badly needs strength-ening, however. Under Article 5.8 of the cur-rent agreement, the threshold is set at 2 per-cent. And because of awkward draftsmanship,this provision is claimed by the United Statesto apply only to investigations—and not tosubsequent administrative reviews that recalcu-late dumping margins after an antidumpingduty order goes into effect. Under U.S. law, thede minimis threshold in such reviews continuesto be a mere 0.5 percent.

The threshold needs to be raised and appliedequally to original investigations and reviews:

Reform Proposal 19: Article 5.8 of theAntidumping Agreement should berevised to provide that any margin ofdumping of less than 5 percent shouldbe treated as de minimis. The samedefinition of de minimis should applyin both original investigations andadministrative reviews.

In view of the irreducible imprecision of dumpingcalculations, the de minimis threshold should beraised even if all of the reform proposals in thispaper for addressing methodological distortionsare ultimately adopted. If, however, any of theexisting distortions are left intact, the case for rais-ing the de minimis threshold becomes that muchstronger. A relatively high de minimis thresholdwould act as a kind of final check on unremediedmethodological abuses.

Mandate a Public-Interest TestA number of WTO members—including

the EU, Canada, Thailand, and Malaysia—haveincorporated a “public-interest test” into theirantidumping regulations. The basic idea behindsuch public-interest provisions is to make theimposition of antidumping measures permissiverather than mandatory. Specifically, a public-interest provision allows authorities to refuse toimpose duties, even when dumping and injuryhave been found, on the ground that antidump-ing measures in a particular case would be con-trary to the broader public interest.

A public-interest test, if properly devised andimplemented, can help to reconcile a country’santidumping policy with its larger nationalinterests. After all, even staunch defenders ofantidumping remedies must recognize that theresort to such remedies carries costs. Even if adomestic industry is being harmed by allegedlydumped imports, other domestic interests—namely downstream import-using industriesand consumers—are benefited by them. Indeed,the fact that the imports are entering the coun-try in sufficient quantities to injure domesticproducers shows that many domestic interestsprefer those imports to products made at home.Accordingly, antidumping investigations involvemore than a dispute between a domestic indus-try and its foreign rivals; they also involve a con-flict of interest between that domestic industryand other domestic industries.

An antidumping law with no public-interestprovision fails to take account of these conflictinginterests. If the requisite showings of dumpingand injury are made, trade-restrictive remediesfollow automatically—regardless of the conse-quences for the rest of the country. That is hardlya recipe for rational policymaking: if major affect-ed interests are systematically ignored in the deci-sionmaking process, it’s hardly likely that theresulting policy will reflect an optimal accommo-dation of all competing interests.

Furthermore, given the tension between thetrade-restrictive effects of antidumping measuresand the market-opening thrust of the WTOagreements as a whole, due restraint in the appli-cation of antidumping measures is in keepingwith the basic concepts, principles, and objectives

33

A healthy dose ofhumility is in orderregarding the accu-racy of any dump-ing calculations.

of the Antidumping Agreement. In that regard,Article 9.1 of the current agreement states, “It isdesirable that the imposition [of duties] be per-missive in the territory of all Members.”

The present agreement, however, does notrequire any kind of public-interest test, much lessspecify standards for how it should be applied.Consequently, many WTO members—includ-ing the United States—have no public-interestprovision at all. Meanwhile, there is little consis-tency or transparency in the public-interest provi-sions that do exist; by and large they are stan-dardless “black boxes” that occasionally block theimposition of duties for no clearly defined reason.In some countries, such as Canada, the public-interest provision merges with a lesser-duty rule,so that a public-interest determination must bemade before the lesser-duty rule is invoked.

Antidumping policy around the world wouldbe greatly improved by mandating the inclusion ofa public-interest test and then specifying standardsfor how it should be applied. On the latter point,the critical challenge is to find some set of criteriathat give the public-interest test real teeth withoutcausing it to swallow up all of antidumping policy.Thus, if the public interest is defined as “whatever’sgood for domestic import-competing industries,”then a public-interest provision will have no effectat all. On the other hand, if the public interest isdefined as pure economic efficiency, then the testwould work to block the imposition of duties invirtually all cases.

Accordingly, we suggest that the AntidumpingAgreement be amended along the following lines:

Reform Proposal 20: Article 9.1 of theAntidumping Agreement should berevised to require the application of apublic-interest test before antidumpingmeasures are imposed. For purposes ofthis test, antidumping measures wouldbe deemed contrary to the public interestif the harm inflicted by those measures ondownstream import-using interests isdeemed disproportionate to the benefitconferred on the petitioning domesticindustry. “Disproportionate,” for thesepurposes, should be defined explicitly inreference to specified benchmarks.

Note that we suggest the use of some kindof “disproportionate impact” standard for pur-poses of applying the public-interest test. Theexistence of a disproportionate impact could bemeasured in a number of different ways. Forexample, the estimated welfare gain for thepetitioning industry could be compared to theestimated welfare loss for specific downstreamindustries, or for consumers. If the loss is somedesignated multiple of the gain, the impactwould be deemed disproportionate and dutieswould not be imposed. Alternatively, the esti-mated number of jobs saved in the petitioningindustry could be compared to the estimatednumber of jobs lost in downstream import-using industries. If the ratio of downstreamjobs lost to petitioning industry jobs savedcrosses some designated threshold, dutieswould not be imposed on the ground of dis-proportionate impact. Or authorities could cal-culate the deadweight loss to the economy perjob saved in the petitioning industry and com-pare that to average wages in the industry. Ifthe economic cost is some designated multipleof the average wage, disproportionate impactwould be found and no duties would beimposed.

Any of these cost/benefit comparisonscould be made with the use of fairly basic tech-niques of quantitative economic analysis.Relatively easy to administer, such a public-interest test would have real teeth while stillgiving wide scope for the use of antidumpingmeasures. Exactly how sharp the teeth or widethe scope can be settled by choosing a higher orlower threshold for “disproportionate”: thehigher the designated multiple of harms tobenefits is set, the more modest the effect ofthe public-interest test.

Make Termination of Antidumping DutyOrders Automatic

Before the WTO Antidumping Agreement,some jurisdictions—in particular, the UnitedStates—lacked any regularly scheduled “sunset”process for terminating antidumping dutyorders. As a result, the average lifetime of U.S.orders exceeded a decade, and some continuedfor more than 30 years.

34

Antidumping poli-cy around the world

would be greatlyimproved by man-

dating the inclusionof a public-interest

test and then speci-fying standards for

how it should beapplied.

Such a state of affairs was glaringly inconsis-tent with any theory of antidumping policy as aresponse to market distortions. If antidumpingmeasures are to be justified on the ground thatthey offset artificial competitive advantagescaused by market distortions, it follows thatthose measures should be discontinued as soonas the distortions are eliminated or the advan-tages disappear. But if imports can be subject toantidumping remedies year after year despite thefact that they are no longer injuring a domesticindustry, then antidumping has ceased to haveanything to do with a level playing field andcrossed over to simple protectionism.

To address this issue, the current WTOAntidumping Agreement provides for a so-called sunset review process. Specifically, Article11.3 of the agreement mandates the automatictermination of antidumping duty orders afterfive years unless a special review initiated beforeexpiration determines that termination of theorder “would be likely to lead to the continua-tion or recurrence of dumping and injury.”

Unfortunately, Article 11.3 has proved lessthan successful in phasing out old orders. Inthe United States, for example, there were 354sunset reviews initiated between July 1998 andAugust 2002, of which 265 were contested bypetitioners. The outcomes in 2 of those 265contested cases were still pending. TheDepartment of Commerce made affirmativesunset determinations to continue the order inall but 4 of the 263 decided cases, while theITC voted affirmative 72 percent of the time.68

The sunset review process is fundamentallyflawed. At the root of the problem is the fact thatthe review is prospective and counterfactual in itsfocus and thus inherently speculative. It seeks todetermine whether dumping and injury will hap-pen in the future if an order is lifted. It is difficultenough to control antidumping authorities’ abuseof discretion when their investigations are tied toa clear evidentiary record; it is next to impossiblewhen the authorities are allowed to gaze into acrystal ball.

Accordingly, to ensure that antidumping mea-sures do not continue after the artificial competitiveadvantages that are their supposed target have beenneutralized, we suggest the following reform:

Reform Proposal 21: Article 11.3 ofthe Antidumping Agreement shouldbe amended to provide for automatictermination of antidumping dutyorders after five years. Domesticindustries would be able to file newpetitions immediately upon expira-tion, but they would be required toshow evidence of actual injury orthreat of injury by reason of dumpedimports just as in any normal case. Forpetitions filed within one year of theexpiration of a prior order, special pro-cedures would be required to expediterelief for petitioners. Specifically, theadministering authorities would berequired to make a preliminary find-ing as to injury within 45 days of theinitiation of the new investigation. Ifthat preliminary determination isaffirmative, preliminary antidumpingmeasures would go into effect at therates that applied at the expiration ofthe old order.

This proposal strikes a reasonable compromisebetween two competing interests: on the one hand,ensuring that antidumping measures are not main-tained even after the conditions that justified themno longer exist, and, on the other hand, continuingto provide a remedy when those conditions happento persist. Under the suggested reform, automatictermination ensures that all orders will come to anend; at the same time, though, special provisionsfor follow-up investigations ensure reasonable con-tinuity of relief if conditions warrant. As to the pre-liminary injury finding proposed, we envisionsomething along the lines of the U.S. ITC’s pre-liminary injury investigation. An affirmative find-ing in this preliminary phase would trigger theimposition of preliminary measures at the old rateseven before any new finding on dumping margins.Subsequent preliminary and final determinationson dumping in the follow-up investigation wouldreplace the old rates with new ones.

Other than the special provisions for ensur-ing continuity of relief, follow-up investigationswould be just like original investigations in everyrespect: the same evidentiary requirements for

35

It is difficultenough to controlantidumpingauthorities’ abuseof discretion whentheir investigationsare tied to a clearevidentiary record;it is next to impos-sible when theauthorities areallowed to gazeinto a crystal ball.

initiation, the same standards for determiningdumping and injury. The all-too-often bogusguessing games of the sunset process would bereplaced by full-blown dumping and injuryanalysis in accordance with the normal provi-sions—and the basic concepts, principles, andobjectives—of the Antidumping Agreement.

Other Issues

The 21 reform proposals discussed do not targetall of the flaws in current antidumping practice.Indeed, some of the most glaring flaws are notaddressed—at least not directly. In crafting our pro-posals, we chose to focus on problems that are (1) seri-ous and (2) susceptible to reform by changes in theWTO Antidumping Agreement. Unfortunately,some of the worst abuses of antidumping laws cannotbe remedied effectively by adding one particular pro-vision or another to WTO rules.

The use of “facts available” in calculatingdumping margins is one of the most importantissues that we did not address. Normally,antidumping authorities calculate a foreign pro-ducer’s dumping margin on the basis of compa-ny-specific price and cost data submitted duringthe course of the investigation. If, however, theforeign producer declines to participate in theinvestigation, or if the authorities determine thatthe information submitted is either incompleteor inaccurate, the authorities may use facts avail-able to calculate the company’s dumping mar-gin. Those facts available frequently include thealleged dumping margins featured in thedomestic industry’s antidumping petition.

The use of facts available typically results inextremely high dumping margins. In an earlierpaper by one of the coauthors of this study, anexamination of 141 U.S. dumping determina-tions over a three-year period found that theaverage dumping margin calculated on thebasis of facts available was a whopping 95.58percent—compared to 27.22 percent when theforeign producer’s data were used. In that peri-od, the Commerce Department used factsavailable just over 25 percent of the time.69

U.S. exporters are frequent victims of factsavailable determinations. As we discussed in an

earlier study, between 1995 and 2000 five of eightIndian dumping determinations against U.S.products were based on facts available—and theaverage dumping margin in those five cases was83 percent. Three of four South African dumpingdeterminations against U.S. products over thesame period were based on facts available, with anaverage dumping margin of 89 percent.70

Unfortunately, no clear-cut solution toabuses of facts available is apparent. Sinceantidumping authorities do not have subpoenapower, they have to rely on the voluntary coop-eration of investigated companies. If thosecompanies refuse to participate, the authoritiescannot simply give up; otherwise, stonewallingwould be a perfect defense. And authoritiesmust be able to throw out incomplete data;otherwise, respondent firms could submit frag-mentary data that appear to exonerate themand the authorities would again be stymied.

Accordingly, there seems to be no alternativeto allowing authorities the discretion to disre-gard respondents’ price and cost data. Given thatfact, authorities must also have the discretion tochoose the facts available that will substitute forrespondents’ data in the dumping calculations.And where there is discretion, there is the ever-present possibility of abuse of discretion.

We do not see how WTO rules could definewith any clarity either (1) the circumstancesunder which resort to facts available is justifiedor (2) the standards for selecting facts availablefor use in dumping calculations. The proprietyor impropriety of antidumping authorities’ con-duct on either front will inevitably turn on case-specific factual circumstances that cannot bespecified in advance. The best that can be doneis to lay out broad, general standards for whenand how facts available should be used.

The current Antidumping Agreementalready establishes such standards. Article 6.8states that facts available can be used only whena respondent “refuses access to, or otherwisedoes not provide, necessary information within areasonable period or significantly impedes theinvestigation.” Annex II, paragraph 5 of theagreement further states that respondents’ infor-mation should not be disregarded “even thoughthe information provided may not be provided

36

Unfortunately, noclear-cut solution toabuses of facts avail-

able is apparent.

in all respects, . . . provided the interested partyhas acted to the best of his ability.” Also, para-graph 7 of Annex II provides guidance regard-ing sources of facts available to be used byantidumping authorities.

It is possible that the existing language couldbe tightened up and improved. Even if that isdone, however, little will have changed.Authorities will still have broad discretion todisregard respondents’ data, and even wider dis-cretion about what to use in their stead. Ifantidumping authorities are intent on abusingthe law and achieving a protectionist outcome,they will still have wide latitude to do so. Onoccasion, victimized countries might challengethe use of facts available through WTO disputesettlement and win a reversal. But by and large,the facts available loophole cannot be closed.

Another serious problem with currentantidumping practice is the lack of transparencyand basic administrative fairness. With its com-plexity and wide scope for discretion, theantidumping law creates enormous potential forabuse in poorer countries that lack well-estab-lished traditions of transparency and the rule oflaw. Failure to provide respondent companieswith the factual and legal bases for determina-tions, to allow them a fair hearing, to takeaccount of their claims of legal and factual errorsin determinations, and to safeguard the privacyof their confidential business data are all proce-dural irregularities that are in clear violation ofcurrent WTO rules, yet they are alleged to bedistressingly common in many countries.Consequently, substantive flaws in antidumpingrules are all too often compounded by egregiousprocedural unfairness.

Abuses caused by nontransparency or out-right corruption are difficult to remedythrough changes in WTO rules. Since currentrules are being widely ignored, it is doubtfulthat new rules will meet a better fate. Indeed,the very essence of nontransparency and cor-ruption is that government officials don’t fol-low the stated rules.

The only effective way to reduce the abusesof facts available and nontransparency is toreduce the number of unjustified antidumpinginvestigations that are initiated and conducted

in the first place. Because of the glaring flaws inexisting antidumping rules, investigations areroutinely instituted without any evidence ofunfair trade under any plausible definition ofthat term. The harm caused by those unjustifi-able investigations can then be exacerbated byabuses of administrative discretion or outrightmisconduct. If the number of unjustifiableinvestigations can be reduced, the number ofinvestigations plagued by facts available or non-transparency can likewise be expected to fall.

The reform proposals set forth in this paperthus constitute an indirect and partial solution toother problems that resist straight-on efforts atreform. All of our reform proposals take the formof specific rules—as opposed to broad, discre-tionary standards. Such clear-cut rules provideauthorities with little discretion about how toimplement them and thus are hard to circumvent.If the new rules proposed here are adopted andincorporated into the WTO AntidumpingAgreement, dramatic improvements in antidump-ing practice would almost certainly ensue. Withsome coherence achieved between the basic con-cepts, principles, and objectives of theAntidumping Agreement and the specific provi-sions of that agreement, there would be a signifi-cant reduction in unjustifiable antidumping activi-ty—that is, in antidumping investigations andmeasures that have no rational relation to offsettingartificial competitive advantages created by mar-ket-distorting government policies. As a result,there would be fewer opportunities for authoritiesto misuse facts available or run roughshod over therequirements of procedural fairness.

Conclusion

Antidumping reform faces formidable obsta-cles. Use of antidumping laws around the worldis widespread and growing; and wherever theselaws operate, the protectionist status quo enjoysthe support of entrenched bureaucracies andimport-competing corporate interests. In theUnited States in particular, energetic and well-organized protectionist lobbies have mobilizednearly overwhelming political support for theirposition on antidumping issues. As a result, for

37

The only effectiveway to reduce theabuses of factsavailable and non-transparency is toreduce the numberof unjustifiedantidumping inves-tigations that areinitiated and con-ducted in the firstplace.

many years the world’s most powerful countryand leader of the multilateral trading system hasstood as the principal opponent of meaningfulchanges in antidumping rules.

Of all the obstacles hindering antidumpingreform, however, none is greater than ignorance.Failure to understand how antidumping lawsactually operate in practice—and how they fail sospectacularly to do what their supporters say theyare supposed to do—lies at the root of much ofthe resistance to antidumping reform. Many sup-porters of the antidumping status quo honestlybelieve that these laws in their present form arenecessary to combat unfair trading practices andthereby ensure a level playing field. If those sup-porters fully understood the reality of contempo-rary antidumping practice—if they understoodhow frequently trade-restrictive measures areinflicted on normal, healthy competition—theiropposition to needed reforms would likely soften.

Of course, protectionist interests supportthe antidumping status quo so fervently pre-cisely because of its flaws. Their goal is tosquelch foreign competition in whatever waythey can, and the antidumping law in its cur-rent form has proved very handy indeed. Andbecause of ignorance about the law’s complexworkings, protectionist interests are able tocloak their special pleading in the high-mind-ed rhetoric of fairness and concern for a levelplaying field. If they were forced to defend thestatus quo honestly, for the protectionist scamthat it is, they would find it much harder to winadherents to their cause.

Accordingly, supporters of antidumpingreform need to make education and clarifica-tion their top priorities in WTO negotiations.Negotiations that focus exclusively on specificchanges to the Antidumping Agreement aredoomed to achieve disappointing results.Instead, the first order of business ought to beclarifying what exactly are the basic concepts,principles, and objectives of the AntidumpingAgreement. In this paper, we have sought toidentify those basic concepts, principles, andobjectives by relying on the justifications forantidumping measures offered by U.S.antidumping supporters. We believe thatWTO negotiations would ultimately arrive at

more or less the same position—namely, thatthe basic objective of the AntidumpingAgreement is to allow member states to offsetartificial competitive advantages created bymarket-distorting government policies.

A consensus along those lines would be ofenormous value in guiding negotiations aboutspecific provisions of the AntidumpingAgreement. That consensus would provide abenchmark by which to evaluate contemporaryantidumping practice—a benchmark in com-parison to which much of contemporary prac-tice would be found sadly deficient. This criticalevaluation, in turn, would help to define thework program of negotiators—namely, toreduce the yawning gap between antidumping’saccepted goals and its actual practice. We believethat the specific reform proposals outlined inthis study define, at least in broad outline, thework program that needs to be undertaken.

This work program may prove too ambi-tious to be accomplished in a single round. Butat least the work can be started—and thegroundwork for ongoing progress in futurerounds can be laid. In the Uruguay Roundagreements on agriculture and services, forexample, actual reductions of market barrierswere modest, but at least a consensus was builtfor the need to make further progress in thefuture. As a result, in the current Doha Round,there is no dispute about whether market barri-ers in agriculture and services should bereduced; the only question is how much.

By contrast, the Uruguay Round achievedno consensus on the proper objectives ofantidumping policy. Negotiators succeeded inhammering out the Antidumping Agreement,but all it really did was to codify existing U.S.and EU practice with a few technical modifi-cations around the edges. Without any consen-sus on why the agreement exists or what pur-pose it serves, when the time came to launch anew round, supporters of reform had to strug-gle ferociously just to get antidumping on thenegotiating agenda.

Antidumping reform shouldn’t have to startfrom scratch every time. Now is the time to builda durable foundation for an ongoing project ofreform. Now is the time to change the terms of

38

Now is the time tobuild a durable

foundation for anongoing project ofreform. Now is thetime to change the

terms of the debate.

the debate. If supporters of reform play their cardsright, the Doha Round will end the question ofwhether antidumping abuses should be curtailed.How much will be the only topic for negotiation infuture rounds.

Notes1. The agreement is available online at www.wto.org/english/docs_e/legal_e/19-adp.pdf.

2. Trade Act of 2002, Public Law 107-210, sec.2102(b)(14)(A).

3. House Con. Res. 262, 107th Cong., 1st sess.,November 7, 2001. Interestingly, at the same time theresolution also instructed the president to “ensurethat United States exports are not subject to the abu-sive use of trade laws, including antidumping andcountervailing duty laws, by other countries.”

4. “62 Senators Call on President Not to WeakenTrade Remedy Laws,” Inside U.S. Trade, May 11, 2001.

5. See Gary Horlick, “The Speed Bump at Seattle,”Journal of International Economic Law 3 (March 2000):167.

6. World Trade Organization, Ministerial Declaration,November 14, 2001, WT/MIN(01)/DEC/W/1, para.28. Emphasis added.

7. An earlier U.S. antidumping law with criminalsanctions was enacted in 1916 (and is at this writ-ing still on the books despite a WTO ruling thatit is inconsistent with the AntidumpingAgreement), but the current law that imposesantidumping duties traces back to 1921.

8. During 1995–2001, the United States initiated255 antidumping investigations, more than anyother country. The second most frequentantidumping user over the same period was India,with 248 initiations; the European Union fol-lowed closely with 246 initiations. Argentina wasa distant fourth with 170 initiations. For statisticson antidumping initiations worldwide during1995–2001, see www.wto.org/english/tratop_e/adp_e/adp_stattab2_e.htm.

9. “Basic Concepts and Principles of the TradeRemedy Rules,” Communication from theUnited States to the WTO Negotiating Group onRules, TN/RL/W/27, October 22, 2002.

10. Ibid., p. 3.

11. Ibid.

12. On the other hand, a focus on national eco-

nomic efficiency may very well conflict with con-cerns about fairness. Even if worldwide economicefficiency is harmed by trade-distorting practices,it may still be in the national economic interest ofan importing country to allow “unfair” importsto enter freely and thereby reap the gains of theirartificially low prices. In such situations, trade-restrictive antidumping remedies will sacrificenational economic efficiency for the sake of fair-ness to domestic import-competing industries.

13. “Basic Concepts and Principles of the TradeRemedy Rules,” p. 3.

14. Ibid., p. 4.

15. “Observations on the Distinctions betweenCompetition Laws and Antidumping Rules,”Submission of the United States to the WTOWorking Group on the Interaction of Trade andCompetition Policy,” Meeting of July 27–28, 1998.Cited hereafter as U.S. WTO 1998 Submission.

16. U.S. 1998 WTO Submission, p. 1.

17. Ibid., p. 2.

18. Ibid.

19. Ibid., p. 9.

20. Ibid., p. 14.

21. Ibid., p. 4.

22. Alan Wolff, “Role of the Antidumping Laws,”Remarks before the Steel ManufacturersAssociation, May 3, 1995, p. 2. Emphasis in original.

23. Terence P. Stewart, “Administration of theAntidumping Law: A Different Perspective,” inDown in the Dumps: Administration of the Unfair TradeLaws, ed. Richard Boltuck and Robert E. Litan(Washington: Brookings Institution, 1991), p. 288.

24. Greg Mastel, Antidumping Laws and the U.S.Economy (Armonk, N.Y.: M. E. Sharpe, 1998).

25. Ibid., pp. 42–43.

26. Ibid., p. 43.

27. U.S. 1998 WTO Submission, p. 2.

28. Mastel, p. 40.

29. This point is developed at greater length inBrink Lindsey, “The U.S. Antidumping Law:Rhetoric versus Reality,” Cato Institute TradePolicy Analysis no. 7, August 16, 1999, pp. 16–17.

30. Brink Lindsey and Dan Ikenson, “Antidumping101: The Devilish Details of ‘Unfair Trade’ Law,”Cato Institute Trade Policy Analysis no. 20,

39

November 26 2002.

31. Lindsey, “The U.S. Antidumping Law:Rhetoric versus Reality.”

32. See Brink Lindsey and Dan Ikenson, “ComingHome to Roost: Proliferating Antidumping Lawsand the Growing Threat to U.S. Exports,” CatoInstitute Trade Policy Analysis no. 14, July 30, 2001.

33. U.S. Trade Representative Robert Zoellick hasnoted this fact on numerous occasions. See, forexample, “Statement of U.S. Trade RepresentativeRobert B. Zoellick before the Committee on Financeof the U.S. Senate,” February 6, 2002, www.ustr.gov/speech-test/zoellick /zoellick_14.html.

34. For a detailed discussion of alternative causesof price differences and below-cost sales, seeLindsey, “Rhetoric versus Reality,” pp. 11–19.

35. Under certain circumstances, antidumpingauthorities ignore data submitted by companiesunder investigation and instead base their dump-ing findings on “facts available”—which usuallymeans the allegations made by the domesticindustry in its petition. See the section on “OtherIssues” below for a fuller discussion.

36. See Lindsey and Ikenson, “Antidumping 101,”pp. 13–16.

37. Note that this table appeared previously as Table 4in Lindsey and Ikenson, “Antidumping 101.”

38. These findings are consistent with an examina-tion of an earlier U.S. case by one of the coauthors.In that case, the 1998 investigation of static randomaccess memory chips from Taiwan, the CommerceDepartment found a dumping margin of 7.56 per-cent. If no cost test had been used, the margin wouldhave been only 2.74 percent—64 percent lower. SeeLindsey, “Rhetoric versus Reality,” p. 9.

39. These figures are derived from Lindsey,“Rhetoric versus Reality,” Table 2, Appendix.

40. “Basic Concepts and Principles of the TradeRemedy Rules,” p. 3.

41. See Lindsey and Ikenson, “Antidumping 101,” p. 16.

42. Lindsey, “Rhetoric versus Reality,” p. 10.

43. Ibid., Table 4, p. 11.

44. See Lindsey and Ikenson, “Antidumping 101,”pp. 21–22.

45. Appellate Body Report on EuropeanCommunities—Anti-Dumping Duties on Imports ofCotton-Type Bed Linen from India, WT/DS141/AB/R, March 1, 2001.

46. In the EU–Bed Linen case, the Appellate Bodyfound that zeroing is inconsistent with Article 2.4.2of the Antidumping Agreement, which providesthat dumping shall be determined on the basis ofcomparing average normal values to average exportprices except under special circumstances. Zeroing,the Appellate Body concluded, departs inappropri-ately from a true average-to-average comparison andthus runs afoul of WTO rules. Because of theAppellate Body’s ruling, the EU has discontinuedzeroing in cases in which dumping margins arebased on average-to-average comparisons.

However, Article 2.4.2 allows a departure fromaverage-to-average comparisons in cases where“authorities find a pattern of export prices whichdiffer significantly among different purchasers,regions or time periods.” In those so-called targeteddumping situations, authorities may calculatedumping by comparing individual export transac-tions to average normal values. The EU’s presentposition is that, in cases where individual-to-averagecomparisons are used, zeroing is not inconsistentwith the Appellate Body’s ruling in EU–Bed Linen.Accordingly, the EU will continue to perform zero-ing in those kinds of cases. See, e.g., “Proposal for aCouncil Regulation Imposing a Definitive Anti-Dumping Duty and Collecting Definitively theProvisional Duty Imposed on Imports ofRecordable Compact Disks Originating in Taiwan,”Official Journal of the European Communities, 2002/C227E/362, September 24, 2002, para. 32–33.

47. See discussion in note 46.

48. The effect of zeroing is somewhat worse in indi-vidual-to-average comparisons, since when model-specific average prices in the home market and exportmarket are compared to each other, at least negativedumping margins on particular sales of that modelwill offset positive dumping margins on sales of thatmodel. Even here, though, overall negative dumpingmargins for one model cannot offset overall positivemargins of other models because of zeroing.

49. At present, the Appellate Body has not ruled onwhether zeroing is permissible in targeted dump-ing situations, or whether individual-to-averagecomparisons are generally permissible in adminis-trative reviews. It is possible that, if these issues areever litigated, the Appellate Body could decide toclose these loopholes under current WTO rules.Such an outcome, though, cannot be assumed.

50. See Lindsey and Ikenson, “Antidumping 101,”pp. 20–21.

51. By contrast, Article 2.3 of the currentAntidumping Agreement provides that export-mar-ket sales to affiliates may be disregarded, in which casea “constructed export price” based on resales from theaffiliates to unrelated purchasers may be used instead.

40

52. The appropriate standards for determiningwhether companies should be regarded as affiliatedparties raise a host of additional issues not coveredin this paper.

53. Under certain circumstances, after excluding affili-ated sales, the Commerce Department will require theaffiliated party to report its resales to unaffiliated cus-tomers. Those resales will then be included in the cal-culation of normal value. In this paper we do not makeany proposal regarding the practice of includingresales in the calculation of normal value.

54. See Lindsey and Ikenson, “Antidumping 101,”pp. 12–13.

55. This investigation gave rise to the dispute thatresulted in the WTO’s ruling that the U.S. arm’s-length test is WTO-inconsistent. See below for adiscussion of that ruling.

56. Report by the Appellate Body on UnitedStates–Anti-Dumping Measures on Certain Hot-RolledSteel Products from Japan, WT/DS184/AB/R, July24, 2001.

57. See the Commerce Department’s request forpublic comment, 67 Fed. Reg. 53339, August 15,2002.

58. See Lindsey and Ikenson, “Antidumping 101,”p. 20.

59. Note that, under certain circumstances, it maybe necessary to use some kind of time lag whendetermining whether increasing imports anddeclining operating profits are substantially cor-related. Specifically, the effect of increasingimports on a domestic industry’s profitabilitymay not be immediate. In such cases, importsduring one period would have to be compared toprofitability during a somewhat later period.

60. Appellate Body Report on United States—Anti-Dumping Measures on Certain Hot-Rolled Steel Productsfrom Japan, WT/DS184/AB/R, July 24, 2001, para. 228.

61. Although these decisions concern theAgreement on Safeguards rather than theAntidumping Agreement, the conceptual issuesregarding causation of injury are identical in thetwo agreements. Accordingly, Appellate Bodyinterpretations of the Safeguards Agreement onthis point are very likely to be followed when thesame issue arises under the AntidumpingAgreement.

62. Appellate Body Report on United States—Safeguard Measures on Imports of Fresh, Chilled orFrozen Lamb Meat from New Zealand and Australia,WT/DS177/AB/R, WT/DS178/AB/R, May 1,2001, para. 168, quoting Appellate Body Report onUnited States—Definitive Safeguard Measures on Imports

of Wheat Gluten from the European Communities,WT/DS166/AB/R, December 22, 2000, para. 69.

63. These figures were derived from U.S.International Trade Commission, Hot-Rolled SteelProducts from Argentina, China, India, Indonesia,Kazakhstan, Netherlands, Romania, South Africa,Taiwan, Thailand, and Ukraine, Investigation nos. 701-TA-404-408 (preliminary) and 731-TA-898-908 (pre-liminary), USITC Publication no. 3381, January2001; Hot Rolled Steel Products from Argentina and SouthAfrica, Inv. nos. 701-TA-404 (final) and 731-TA-898and 905 (final), USITC Publication no. 3446, August2001. Note that the difference between the mer-chant market and total domestic consumption liesin the fact that approximately 65 percent of hot-rolled steel produced by the U.S. industry is “cap-tively consumed” by U.S. steel producers in the man-ufacture of downstream products.

64. See Thomas J. Prusa, “On the Spread andImpact of Antidumping,” National Bureau ofEconomic Research Working Paper 7404,October 1999, www.nber.org/papers/w7404.

65. The average Commerce dumping margin wascalculated as the simple average of all the compa-ny-specific dumping margins from original inves-tigations during 2001 in which adverse facts avail-able were not used. Since adverse facts availableare generally based on figures in the petition,inclusion of these determinations would have cre-ated circularity problems when comparingCommerce’s findings to petitioners’ allegations.For petition dumping margins, we used compa-ny-specific allegations if available; otherwise, weused either the simple average of the lowest andhighest rates alleged or the countrywide averagerate if one was alleged. We then calculated thesimple average of all the petition rates thusderived from all original investigation determina-tions during 2001 in which Commerce calculateda company-specific rate not based on adverse factsavailable.

66. Panel Report on Guatemala—Definitive Anti-Dumping Measures on Grey Portland Cement fromMexico, WT/DS156/R, October 24, 2000.

67. See Commission of the European Communities,“Nineteenth Annual Report from the Commissionto the European Parliament on the Community’sAnti-Dumping and Anti-Subsidy Activities (2000),”October 12, 2001.

68. These figures were compiled as of August 14, 2002,from data available on the ITC website, www.usitc.gov.

69. These figures were derived from Lindsey,“Rhetoric versus Reality,” Table 2, Appendix.

70. Lindsey and Ikenson, “Coming Home toRoost,” p. 15.

41

42

Trade Briefing Papers from the Cato Institute

“Rethinking the Export-Import Bank” by Aaron Lukas and Ian Vásquez (no. 15, March 12, 2002)

“Steel Trap: How Subsidies and Protectionism Weaken the U.S. Steel Industry” by Dan Ikenson, (no. 14, March 1, 2002)

“America’s Bittersweet Sugar Policy” by Mark A. Groombridge (no. 13, December 4, 2001)

“Missing the Target: The Failure of the Helms-Burton Act” by Mark A. Groombridge (no. 12, June 5, 2001)

“The Case for Open Capital Markets” by Robert Krol (no. 11, March 15, 2001)

“WTO Report Card III: Globalization and Developing Countries” by Aaron Lukas (no. 10, June 20, 2000)

“WTO Report Card II: An Exercise or Surrender of U.S. Sovereignty?” by William H. Lash III and Daniel T. Griswold (no.9, May 4, 2000)

“WTO Report Card: America’s Economic Stake in Open Trade” by Daniel T. Griswold (no. 8, April 3, 2000)

“The H-1B Straitjacket: Why Congress Should Repeal the Cap on Foreign-Born Highly Skilled Workers” by Suzette BrooksMasters and Ted Ruthizer (no. 7, March 3, 2000)

“Trade, Jobs, and Manufacturing: Why (Almost All) U.S. Workers Should Welcome Imports” by Daniel T. Griswold (no. 6,September 30, 1999)

“Trade and the Transformation of China: The Case for Normal Trade Relations” by Daniel T. Griswold, Ned Graham, RobertKapp, and Nicholas Lardy (no. 5, July 19, 1999)

“The Steel ‘Crisis’ and the Costs of Protectionism” by Brink Lindsey, Daniel T. Griswold, and Aaron Lukas (no. 4, April 16,1999)

“State and Local Sanctions Fail Constitutional Test” by David R. Schmahmann and James S. Finch (no. 3, August 6, 1998)

“Free Trade and Human Rights: The Moral Case for Engagement” by Robert A. Sirico (no. 2, July 17, 1998)

“The Blessings of Free Trade” by James K. Glassman (no. 1, May 1, 1998)

From the Cato Institute Briefing Papers Series

“The Myth of Superiority of American Encryption Products” by Henry B. Wolfe (no. 42, November 12, 1998)

“The Fast Track to Freer Trade” by Daniel T. Griswold (no. 34, October 30, 1997)

“Anti-Dumping Laws Trash Supercomputer Competition” by Christopher M. Dumler (no. 32, October 14, 1997)

43

Trade Policy Analysis Papers from the Cato Institute

“Antidumping 101: The Devilish Details of ‘Unfair Trade’ Law” by Brink Lindsey and Dan Ikenson (no. 20, November 26, 2002)

“Willing Workers: Fixing the Problem of Illegal Mexican Migration to the United States” by Daniel T. Griswold (no. 19, October 15,2002)

“The Looming Trade War over Plant Biotechnology” by Ronald Bailey (no. 18, August 1, 2002)

“Safety Valve or Flash Point? The Worsening Conflict between U.S. Trade Laws and WTO Rules” by Lewis E. Leibowitz (no. 17,November 6, 2001)

“Safe Harbor or Stormy Waters? Living with the EU Data Protection Directive” by Aaron Lukas (no. 16, October 30, 2001)

“Trade, Labor, and the Environment: How Blue and Green Sanctions Threaten Higher Standards” by Daniel T. Griswold (no. 15,August 2, 2001)

“Coming Home to Roost: Proliferating Antidumping Laws and the Growing Threat to U.S. Exports” by Brink Lindsey and DanIkenson (no. 14, July 30, 2001)

“Free Trade, Free Markets: Rating the 106th Congress” by Daniel T. Griswold (no. 13, March 26, 2001)

“America’s Record Trade Deficit: A Symbol of Economic Strength” by Daniel T. Griswold (no. 12, February 9, 2001)

“Nailing the Homeowner: The Economic Impact of Trade Protection of the Softwood Lumber Insudstry” by Brink Linsey,Mark A. Groombridge, and Prakash Loungani (no. 11, July 6, 2000)

“China’s Long March to a Market Economy: The Case for Permanent Normal Trade Relations with the People’s Republic ofChina” by Mark A. Groombridge (no. 10, April 24, 2000)

“Tax Bytes: A Primer on the Taxation of Electronic Commerce” by Aaron Lukas (no. 9, December 17, 1999)

“Seattle and Beyond: A WTO Agenda for the New Millennium” by Brink Lindsey, Daniel T. Griswold, Mark A.Groombridge and Aaron Lukas (no. 8, November 4, 1999)

“The U.S. Antidumping Law: Rhetoric versus Reality” by Brink Lindsey (no. 7, August 16, 1999)

“Free Trade, Free Markets: Rating the 105th Congress” by Daniel T. Griswold (no. 6, February 3, 1999)

“Opening U.S. Skies to Global Airline Competition” by Kenneth J. Button (no. 5, November 24, 1998)

“A New Track for U.S. Trade Policy” by Brink Lindsey (no. 4, September 11, 1998)

“Revisiting the ‘Revisionists’: The Rise and Fall of the Japanese Economic Model” by Brink Lindsey and Aaron Lukas (no. 3,July 31, 1998)

“America’s Maligned and Misunderstood Trade Deficit” by Daniel T. Griswold (no. 2, April 20, 1998)

“U.S. Sanctions against Burma: A Failure on All Fronts” by Leon T. Hadar (no. 1, March 26, 1998)

Board of Advisers

James K. GlassmanAmerican EnterpriseInstitute

Douglas A. IrwinDartmouth College

Lawrence KudlowKudlow & Co.

José PiñeraInternational Center forPension Reform

Razeen SallyLondon School ofEconomics

George P. ShultzHoover Institution

Walter B. WristonFormer Chairman andCEO, Citicorp/Citibank

Clayton YeutterFormer U.S. TradeRepresentative

Nothing in Trade Policy Analysis should be construed as necessarily reflecting the views of theCenter for Trade Policy Studies or the Cato Institute or as an attempt to aid or hinder the pas-sage of any bill before Congress. Contact the Cato Institute for reprint permission. Additionalcopies of Trade Policy Analysis studies are $6 each ($3 for five or more). To order, contact theCato Institute, 1000 Massachusetts Avenue, N.W., Washington, D.C. 20001. (202) 842-0200, fax (202) 842-3490, www.cato.org.

The mission of the Cato Institute’s Center for Trade Policy Studies is to increase publicunderstanding of the benefits of free trade and the costs of protectionism. The center

publishes briefing papers, policy analyses, and books and hosts frequent policy forums andconferences on the full range of trade policy issues.

Scholars at the Cato trade policy center recognize that open markets mean wider choicesand lower prices for businesses and consumers, as well as more vigorous competition thatencourages greater productivity and innovation. Those benefits are available to any countrythat adopts free-trade policies; they are not contingent upon “fair trade” or a “level playingfield” in other countries. Moreover, the case for free trade goes beyond economic efficiency.The freedom to trade is a basic human liberty, and its exercise across political borders unitespeople in peaceful cooperation and mutual prosperity.

The center is part of the Cato Institute, an independent policy research organization inWashington, D.C. The Cato Institute pursues a broad-based research program rooted in thetraditional American principles of individual liberty and limited government.

For more information on the Center for Trade Policy Studies,visit www.freetrade.org.

Other Trade Studies from the Cato Institute

“Antidumping 101: The Devilish Details of ‘Unfair Trade’ Law” by Brink Lindsey and DanIkenson, Trade Policy Analysis no. 20 (November 26, 2002)

“Willing Workers: Fixing the Problem of Illegal Mexican Migration to the United States” byDaniel T. Griswold, Trade Policy Analysis no. 19 (October 15, 2002)

“The Looming Trade War over Plant Biotechnology” by Ronald Bailey, Trade Policy Analysisno. 18 (August 1, 2002)

“Rethinking the Export-Import Bank” by Aaron Lukas and Ian Vásquez, Trade BriefingPaper no. 15 (March 12, 2002)

“Steel Trap: How Subsidies and Protectionism Weaken the U.S. Steel Industry” by Dan Ikenson,Trade Briefing Paper no. 14 (March 1, 2002)

“America’s Bittersweet Sugar Policy” by Mark A. Groombridge, Trade Briefing Paper no. 13(December 4, 2001)

CENTER FOR TRADE POLICY STUDIES