The U.S. antidumping law, accordingto its supporters, ensures “fair trade” by off-setting market distortions caused by for-eign governments. Specifically, it allegedlytargets “unfair” pricing practices—pricediscrimination and below-cost sales—thatreflect protectionism, cartelization, subsi-dies, and other structural defects in foreignmarkets.

To evaluate those claims, the author ofthis study reviewed all U.S. Department ofCommerce final determinations throughthe end of 1998 in original antidumpinginvestigations initiated since January 1,1995—a total of 141 company-specificdumping findings in 49 different cases. Inaddition, for particular companies it waspossible to examine highly detailed priceand cost data from the confidential caserecord.

The evidence reviewed in this studyshows that there is a disconnect betweenthe rhetoric of antidumping supportersand the reality of antidumping practice.The law as currently written and enforceddoes not reliably identify either price dis-

crimination or below-cost sales. Of the fivedifferent calculation methodologies usedby the Commerce Department to measuredumping, only one has any relevance todetecting market-distorting price discrim-ination; only 2 of the 107 affirmativedumping findings reviewed in this studyrelied exclusively on this methodology.None of the calculation methodologiesmeasures whether sales are below cost; theone that comes closest merely determineswhether profits are below an often arbi-trary and inflated benchmark.

Furthermore, the law lacks any mecha-nism for determining whether the pricingpractices it condemns as unfair have anyconnection to market-distorting policiesabroad. Although price discrimination andbelow-cost sales can result from govern-ment interventionism, they can also be dueto perfectly normal marketplace behavior.Consequently, the antidumping law fre-quently punishes foreign firms for unex-ceptionable business practices routinelyengaged in by American companies.

The U.S. Antidumping LawRhetoric versus Reality

by Brink Lindsey

August 16, 1999 No. 7

Brink Lindsey is director of the Cato Institute’s Center for Trade Policy Studies.

Executive Summary

2

Introduction

The U.S. antidumping law protectsAmerican industries from supposedly unfairimport competition.1 Specifically, it imposesextra duties on goods from a particular countryor group of countries if two conditions are met:first, the Department of Commerce must findthat the goods are being sold in the UnitedStates at “dumped” prices; second, theInternational Trade Commission must deter-mine that the imports in question are causingor threatening “material injury” to domesticproducers of the “like product.”2

Antidumping advocates hail the law as abulwark against unfair trade practices abroad.They argue that dumping—which they defineas either international price discrimination orexport sales at prices below the cost of produc-tion—results from interventionist governmentpolicies and structural differences betweennational economies. Those market distortionsallegedly give foreign firms an unfair competi-tive advantage in the U.S. market by allowingthem to charge lower prices than would be pos-sible under normal market conditions.Antidumping duties are needed to offset thatunfair advantage and thereby ensure theproverbial level playing field.

The claims of the antidumping law’s sup-porters raise basic questions about the properobjectives of U.S. trade policy. Assuming thatthe antidumping law does indeed target mar-ket-distorting practices, does it really makesense to respond to those practices by protect-ing particular American companies from thecompetitive consequences of those practices?Granted, cheap imports are capable of injuringspecific import-competing firms; those samecheap imports, however, just as clearly benefitthe American companies that buy and usethem, not to mention the millions of con-sumers who buy from those companies. So whyis it appropriate to sacrifice the interests ofsome Americans to the interests of others? Arethe interests of import-competing firms really avalid proxy for the broader national economicinterest?

Those questions go to the heart of the

ongoing debate over free trade versus “fairtrade.” Any complete assessment of antidump-ing policy must ultimately grapple with them.Before reaching those fundamental issues,though, it is necessary to examine whether theantidumping law does in fact uphold someplausible notion of fair trade. This is the narrowand specific focus of this paper: does theantidumping law really target market distor-tions caused by foreign governments? In otherwords, does the antidumping law really dowhat its supporters claim it does?

An examination of those questions reveals adisconnect between the rhetoric of antidump-ing supporters and the reality of antidumpingpractice. The antidumping law as currentlywritten and enforced does not reliably identifyeither price discrimination or below-cost sales.Furthermore, the law lacks any mechanism fordetermining whether the pricing practices itcondemns as unfair have any connection tomarket-distorting policies abroad. Althoughprice discrimination and below-cost sales canresult from government interventionism, theycan also be due to perfectly normal marketplacebehavior. Consequently, the antidumping lawall too frequently punishes normal marketplacebehavior that has nothing to do with “unfairtrade” under any plausible definition of thatterm.

Targeting ArtificialAdvantages

Advocates of antidumping claim thatdumping is an unfair trade practice that takestwo different forms: price discrimination andbelow-cost sales. Both types of dumpingallegedly reflect underlying market distortionscaused by foreign government policies. Thosedistortions confer an artificial advantage onforeign producers when they are selling in theUnited States—they can sell at lower pricesthan would otherwise be possible.

Thus, price discrimination (i.e., selling atlower prices in the United States than at home)supposedly signals the existence of a protected“sanctuary” home market. According to Greg

Antidumpingadvocates hail the

law as a bulwarkagainst unfair

trade practicesabroad.

3

Mastel, formerly a trade policy analyst with theEconomic Strategy Institute and a firm sup-porter of the antidumping law:

If a company engages in dumping inforeign markets and its home market isopen, the price differential will inducethe company’s competitors or otherresellers to reexport dumped products tothe dumper’s home market. These reex-ports would quickly pull the home mar-ket price down to the dumped price anderase home market profits. Thus, aclosed or restricted home market is alsoa virtual precondition to a successfuldumping strategy.3

This situation gives the foreign producer anarguably unfair competitive advantage overU.S. rivals. “A closed home market allows com-panies to charge high prices at home becausethey face no foreign competition,” Mastelexplains. “Foreign companies can then use theprofits from these domestic sales to cross-sub-sidize export sales at dumped prices.”4

As to sales below cost, the contention is thatthe foreign producer could not sustain its loss-es in the absence of market-distorting govern-ment policies back home. Here again, a domes-tic sanctuary market could be the culprit:supranormal profits at home could allow acompany to take losses abroad. Alternatively,government subsidies could prop up a compa-ny in spite of its losses. The subsidies mighttake the form of explicit grants or soft loans, orthey might be considerably more subtle. Under“crony capitalism,” for example, a politicizedbanking system can allow a well-connected butmoney-losing company to receive financingwithout regard to commercial considerations.

Another possibility is that loss-makingexport sales reflect basic structural flaws in aforeign country’s economic policies. For exam-ple, the absence of functional bankruptcy lawscould allow money-losing companies to con-tinue in existence simply because their creditorshave no better remedy than to keep them afloatand hope for a turnaround. In another possiblescenario, hyperinflation or other severe mone-

tary disorder may reduce companies to barteroperations in which concepts of profit and lossno longer obtain.

Note that dumping as described above isnot anticompetitive in the sense that econo-mists use the term. Although politicians andprotectionist business leaders may rail against“predatory dumping,” the more sophisticatedsupporters of antidumping shy away from suchrhetoric. They recognize that true predatorypricing—aggressive underselling of rivals in thehope of driving them out of business and even-tually establishing a monopoly—is rarelyattempted and even more rarely succeeds.“There are only a handful of cases in recent his-tory,” Mastel concedes, “in which it reasonablycan be argued that such a systematic predatorystrategy was being followed.”5 Furthermore, itis clear that antidumping policies do not followcompetition policy standards for dealing withpredation. “The antidumping rules are notintended as a remedy for predatory pricingpractices of firms,” states a U.S. submission tothe World Trade Organization that staunchlydefends the U.S. law, “or as a remedy for anyother private anticompetitive practices typical-ly condemned by competition laws.”6

The primary justification for the antidump-ing law is really more political than economic.The guiding precept is legitimacy rather thanefficiency. Specifically, the argument is thatinternational competition should be subject tocertain agreed-upon “rules of the game”according to which some sources of competi-tive advantage—trade barriers, subsidies, andother market-distorting government policies—are condemned as unfair. In this conception,the legitimacy of international trade flows—and ultimately, political support for maintain-ing those flows—is contingent upon denyingcompetitors the benefits of any unfair advan-tage and thereby ensuring the much-invokedlevel playing field.

The U.S. WTO submission is very explicitin that regard:

The focus of the antidumping rules. . . is not consumer welfare or allocativeefficiency. Rather, consistent with other

The primary justi-fication for theantidumping law isreally more politicalthan economic.The guiding pre-cept is legitimacyrather thanefficiency.

4

WTO agreements, the AntidumpingAgreement implicitly recognizes thatthere is an accepted norm for the behav-ior of governments in the broad multi-lateral trade context, i.e., a governmentshould not pursue industrial policieswhich distort market structures orprocesses and thereby provide artificialadvantages to domestic producers to thedetriment of producers in other coun-tries. The Antidumping Agreement alsorecognizes that there should be a remedyfor certain harms caused when differenteconomic systems interact.7

It is beyond the scope of this paper toexplore whether such rhetoric makes sense—whether the distinction between “natural” and“artificial” competitive advantages is intellectu-ally coherent, and whether erecting trade barri-ers against imports that enjoy those advantagescharacterized as artificial constitutes soundtrade policy or indeed promotes fairness in anymeaningful sense of that term.8 The aim here isnarrower: it is simply to examine whether thereality of antidumping practice matches itsrhetoric. Are antidumping duties, for better orworse, really offsetting the effects of market-distorting government policies?

This question needs to be answered in twostages. First, it is necessary to determine theeffectiveness of current antidumping method-ologies at targeting the supposedly unfair pric-ing practices of price discrimination and sellingbelow cost. Second, to the extent that theantidumping law does indeed find its targets, itmust be ascertained whether those pricingpractices are reliable indicators of underlyingmarket distortions.

How Dumping Is CalculatedThe first step in this inquiry is to examine

how dumping is actually calculated under U.S.law. In general, the Commerce Departmentcompares the prices of imported merchandisesold in the United States to some measure of“normal value.” There are, however, a numberof different ways to perform such compar-isons—and in particular, a number of different

benchmarks for determining normal value.In the most familiar method, Commerce

compares “net” U.S. prices to “net” home-mar-ket prices. To arrive at net values, Commercesubtracts freight charges, brokerage and han-dling fees, commissions, and various other sell-ing expenses; the idea here is to compare priceson an “ex factory” basis.

The antidumping statute indicates thatcomparing U.S. and home-market prices is thepreferred method of calculating dumping mar-gins.9 If specified conditions exist, though, theCommerce Department will employ alterna-tive methodologies. Thus, if the foreign pro-ducer does not sell the subject merchandise inthe domestic market, or its total domestic salesare less than 5 percent of its U.S. sales, thehome market is considered not viable.10 In thatcase the Commerce Department will selectanother export market to serve as the compar-ison market; U.S. prices are then compared toprices in some third-country market.11 If thereare no viable third-country markets,Commerce will compare U.S. prices to “con-structed value”—which is equal to the compa-ny’s total cost of production plus some amountfor profit.12

The Commerce Department can deviatefrom normal price-to-price comparisons evenwhen there is a viable domestic or third-coun-try market. Within the broad category of mer-chandise under investigation, there may bemany different specific product types or mod-els. For each model sold in the United States,Commerce tries to identify sales of identical orsimilar products in the comparison market; if itcannot find any such sales, the U.S. sales of thatmodel will be compared to constructed value.13

More important, Commerce examinescomparison-market prices to determinewhether they are below the full cost of produc-tion. If more than 20 percent of comparison-market prices of a particular model are belowcost, Commerce will exclude all the below-costsales of that model from its calculations on theground that they are “outside the ordinarycourse of trade.” In that case, U.S. prices arecompared to above-cost comparison-marketprices only; if there are no above-cost sales of

Are antidump-ing duties, for bet-ter or worse, really

offsetting theeffects of market-

distorting govern-ment policies?

5

identical or similar merchandise, U.S. prices arecompared to constructed value.14

The Commerce Department employsanother methodology altogether for importsfrom “nonmarket economies” (NMEs), that is,China and members of the former Soviet bloc.15

In NME cases, Commerce rejects home-mar-ket prices as unreliable, since they are not theproduct of genuine market transactions.Constructed value is also rejected on the groundthat the company’s costs are likewise not marketbased. Instead, Commerce obtains the compa-ny’s “factors of production”—the physical quan-tities of all the inputs used in producing themerchandise—and values those inputs on thebasis of prices in a “surrogate country.”Surrogate countries are market economiesjudged to be at a level of economic developmentsimilar to that of the NME country in question.Commerce then compares U.S. prices to a cost-based normal value derived from company-spe-cific factors of production and surrogate-coun-try prices of those factors (including surrogate-country averages for selling, general, andadministrative expenses and profit).16

Finally, the Commerce Department some-times calculates dumping on the basis of “factsavailable” rather than actual company data.17

Determinations are based on facts availablewhen a foreign producer fails to provide all the

price and cost information requested by theCommerce Department, or when the informa-tion provided is judged to be inaccurate orincomplete (an ever-present possibility giventhe byzantine complexity of documentationthat foreign companies are required to pro-vide). In those situations, the facts availableused by the Commerce Department are gener-ally derived from the allegations contained inthe domestic industry’s antidumping petition.18

Missing the TargetWhat do the various calculation method-

ologies have to do with finding either pricediscrimination or sales below cost? As it turnsout, not very much. As to price discrimination,only one methodology even attempts to mea-sure relevant international price differences;and none of the methodologies seeks to deter-mine whether sales below cost are occurring(Table 1).

Of all the different ways that theCommerce Department measures dumping,only the straightforward comparison of home-market and U.S. prices is capable of identifyingprice discrimination that reflects a protectedsanctuary market. On the other hand, theapparent price discrimination may be nothingmore than an artifact of imperfect price com-parisons.

Table 1Antidumping’s Poor Aim

Calculation Relevance to Relevance toMethodology Price-Discrimination Dumping Below-Cost Dumping

U.S. prices to overinclusive nonehome-market prices

U.S prices to none nonethird-country prices

Constructed none overinclusivevalue

NME surrogate-country- none overinclusivebased normal value

“Facts available” none none

What do the vari-ous calculationmethodologieshave to do withfinding either pricediscrimination orsales below cost?As it turns out, notvery much.

6

In the typical antidumping investigation,the Commerce Department compares home-market and U.S. prices of physically differentgoods, in different kinds of packaging, sold atdifferent times, in different and fluctuatingcurrencies, to different customers at differentlevels of trade, in different quantities, with dif-ferent freight and other movement costs, dif-ferent credit terms, and other differences indirectly associated selling expenses (e.g., com-missions, warranties, royalties, and advertising).Is it any wonder that the prices aren’t identical?

Admittedly, the Commerce Department’sdumping calculation methodologies try toadjust for some of the differences, but theadjustments are necessarily imprecise. Forexample, when the Commerce Departmentcompares physically different merchandise, itadjusts for differences in materials, direct labor,and variable overhead costs.19 While this makesa certain amount of sense, in a real-world com-mercial context it goes without saying thatactual price differences may be more or lessthan the differences in variable manufacturingcosts. And in many cases, the CommerceDepartment makes no adjustment. Thus,prices of goods sold in the United States maybe compared to prices of goods sold manymonths earlier or later in the home marketwithout any adjustment for market fluctuationsover the intervening time. And although unitprices typically decline with larger order quan-tities, the Commerce Department rarelyadjusts for quantity discounts.

Critics of antidumping have focused con-siderable attention on asymmetries in theCommerce Department’s methodologies thatproduce a bias in favor of finding price differ-ences.20 Without a doubt, such asymmetriesexist.21 But the more fundamental and toooften neglected problem is that the practice ofcomparing each and every U.S. sale to somesale in the home market will produce spuriousprice differences that are purely the product of“apples-and-oranges” comparisons.

Whatever the problems associated withcomparing home-market and U.S. prices, atleast such comparisons bear directly on thequestion of international price discrimination

and possible sanctuary markets. By contrast,the other methodologies have nothing to dowith finding relevant international price differ-ences.

Thus, a comparison of U.S. and third-coun-try prices can possibly show international pricediscrimination, but it cannot reveal a sanctuarymarket. Any foreign producer under investiga-tion is an “outsider” as far as all third-countrymarkets are concerned; it is hindered, nothelped, by any government barriers that blockaccess to its export sales. If for some reason thecompany is earning higher prices in that thirdcountry, the reason clearly is not that govern-ment-imposed barriers are shielding it fromcompetition. On the contrary, it had to over-come any barriers that were present in thatthird-country market to be selling there at all.Meanwhile, prices charged in a third countryindicate nothing about whether a firm’s homemarket is closed.

Comparison of U.S. prices to a cost-basednormal value—whether it is derived from thecompany’s own costs (in constructed-valuecases) or from surrogate-country prices (inNME cases)—cannot show price discrimina-tion, for the simple reason that price data arenot used for one side of the comparison.Furthermore, a finding of dumping using con-structed value offers no evidence of the exis-tence of a sanctuary home market. All such afinding can show is that U.S. sales are beingmade below some baseline level of profitability;it cannot show that home-market sales are aboveany similar baseline, since home-market salesare excluded from the dumping calculation.

Indeed, when constructed value is usedbecause there are no above-cost sales of identi-cal or similar merchandise in the home market,the available evidence weighs against the exis-tence of a sanctuary market. A sanctuary mar-ket is one in which a foreign company is mak-ing supranormal profits due to governmentintervention; here, though, the company isapparently losing money at home. The sup-posed source of unfair advantage—namely, theopportunity to cross-subsidize low-priceexport sales—is missing.

The situation is similar when U.S. sales are

When con-structed value is

used because thereare no above-cost

sales, the availableevidence weighsagainst the exis-

tence of a sanctuarymarket.

7

compared to above-cost home-market salesonly. A dumping finding based on such com-parisons tells us nothing about the existence ofinternational price discrimination, since thecomparisons are skewed: low-price sales havebeen excluded from the home-market side, butnot the U.S. side. And here again, as in con-structed-value cases, the evidence affirmativelyrebuts claims of a sanctuary market. Below-cost sales are excluded only when they consti-tute at least 20 percent of home-market sales;such widespread losses are inconsistent withthe supposedly supranormal profits of a sanctu-ary market.

Finally, a dumping finding based on factsavailable provides no evidence of either pricediscrimination or a sanctuary market. The factsavailable are generally taken from the domesticindustry’s antidumping petition, hardly asource of objective analysis. Indeed, it isexpressly recognized that determinations onthe basis of facts available are punitive; it is thethreat of such determinations that is used tocompel foreign producers’ cooperation with theCommerce Department’s often onerous infor-mation requests.22 In any event, the dumpingallegations in antidumping petitions are oftenbased on estimates of constructed value, andthus are incapable of substantiating the exis-tence of price discrimination or a sanctuarymarket.

If the antidumping law takes poor aim atprice discrimination, it fires completely blindlywhen it comes to sales below cost. Not one ofthe methodologies employed by theCommerce Department measures whetherimported merchandise is sold at a loss.Commerce does determine whether home-market or third-country sales are below cost indeciding whether to exclude them as “outsidethe ordinary course of trade.” That inquiry,though, is irrelevant to the issue of whetherU.S. sales are below cost.

The closest the Commerce Departmentcomes to determining whether U.S. sales aremade at a loss is in constructed-value andNME cases. In those cases, Commerce doescalculate the production costs of the merchan-dise sold in the United States,23 but then it adds

an amount for profit before the resulting nor-mal value is compared to U.S. sales prices.Thus, the criterion for deciding whetherimports are unfairly traded under this method-ology is, not the existence of losses, but insuffi-cient profitability. Sales at a loss are considereddumped, but so are profitable sales if the profitrate is too low.24

That overinclusiveness is exacerbated by thespecific way in which dumping margins arecalculated in cost-based cases. The CommerceDepartment compares average U.S. prices ofspecific models to a single product-wide orindustry-wide profitability rate. Sales below theprofitability benchmark are considereddumped; sales above the benchmark aredeemed to have dumping margins of zero.Consequently, even if U.S. sales average a “nor-mal” profit, dumping will be found simplybecause profit rates vary by model.

Finally, there is an additional layer ofmethodological distortion in NME cases. Inthose cases, the cost data used are not those ofthe firm under investigation; instead, surrogatevalues from another country are applied to thatfirm’s factors of production. This methodologyis fraught with potential for gross inaccuracy.25

The extent to which the end result bears anyrelation to market-based costs is open to seriousquestion.

Examining the Case Record

To evaluate the problems with currentantidumping practice in fuller detail, the authorof this study examined all CommerceDepartment final determinations throughDecember 31, 1998, in original antidumpinginvestigations initiated since January 1, 1995—the effective date of the Uruguay RoundAgreements Act (see Appendix). This sampleis large enough to allow generalizations aboutpatterns of antidumping practice and has thefurther virtue of including only determinationsunder the law as it currently exists.26 It includes141 company-specific dumping determina-tions in 49 different antidumping investiga-tions.27 Commerce made affirmative dumping

If the antidumpinglaw takes poor aimat price discrimina-tion, it fires com-pletely blindlywhen it comes tosales below cost.

8

findings for 107 of the 141 companies investi-gated and in 48 of the 49 investigations. Theaverage dumping margin in the sample, includ-ing all the zero and de minimis dumping find-ings,28 is 44.68 percent.

The most striking fact that emerges from areview of this case record is how fewantidumping determinations have anything todo with targeting—or even attempting to tar-get—price discrimination associated with pos-sible sanctuary markets. Price discriminationbulks very large in antidumping rhetoric29 butcommands much less attention in actualantidumping practice.

Of the 141 total determinations, 36 arebased on facts available rather than actual com-pany data.30 Another 47 of the determinations

are from the 14 NME investigations includedin the sample. In 16 of the determinations,constructed value was used either because therewas no viable comparison market or becausethere were no identical or similar products soldin the comparison market. For 37 determina-tions, at least 20 percent of the sales of some orall comparison products were below cost, so theCommerce Department compared U.S. pricesto some combination of comparison-marketprices, above-cost comparison-market pricesonly, and constructed value. And one determi-nation is based purely on a comparison of U.S.and third-country prices.

That leaves only 4 determinations in whichthe Commerce Department calculated dump-ing strictly on the basis of comparisons of U.S.

Table 2Summary of Antidumping Investigations, 1995-98

Calculation Determinations Avg. Dumping MarginsMethodology (affirmative only) (affirmative only)

U.S. prices to 4 4.00%home-market prices (2) (7.36%)

U.S. prices to 1 0%third-country prices (0) (0%)

U.S. prices to mixtureof home-market prices, 31 14.59%above-cost home-market prices, (25) (17.95%)and constructed value

U.S. prices to mixtureof third-country prices, 2 7.94%above-cost third-country prices, (2) (7.94%)and constructed value

Constructed value 20 25.07%(14) (35.70%)

Nonmarket economy 47 40.03%(28) (67.05%)

“Facts available” 36 95.58%(36) (95.58%)

Total 141 44.68%(107) (58.79%)

Price discrimina-tion bulks very

large in antidump-ing rhetoric but

commands muchless attention in

actual antidumpingpractice.

9

and home-market prices. Furthermore, in 2 ofthe 4 determinations in question, theCommerce Department concluded that therewas zero or de minimis dumping. Thus, in only2 of the 107 total affirmative determinations(both of which were made in the same investi-gation) did the Commerce Department finddumping by relying exclusively on the only cur-rently used calculation methodology that bearsany possible connection to the existence of mar-ket-distorted price discrimination (Table 2).

Another 31 determinations, encompassing17 different investigations, relied partially oncomparisons of U.S. and home-market prices.31

In all of those determinations, however,Commerce skewed at least some of the com-parisons by using only above-cost home-mar-ket sales, or by substituting constructed valuefor actual price data. In those mixed cases,Commerce found dumping in 25 of the deter-minations. For those determinations, however,it is impossible to tell from the public recordhow much of each dumping margin is attribut-able to normal comparisons of U.S. and home-market prices, how much to comparisons ofU.S. prices and above-cost home-market pricesonly, and how much to comparisons of U.S.prices and constructed value. In other words,there is insufficient publicly available informa-tion to distinguish between the “signal” ofinternational price differences and the “noise”of dumping margins generated by methodolo-gies that do not detect price differences.

There are good grounds for assuming thatthe “noise” is considerable. Mixing methodolo-gies tends to increase dumping margins abovewhat would be found if only normal price-to-price comparisons were made. Comparing U.S.sales to only above-cost home-market salesalways exaggerates dumping margins, since allthe lowest-price home-market sales are exclud-ed from the comparison. And resort to con-structed value often exaggerates dumping mar-gins because of the artificially high profit ratesthat are frequently used.

To illustrate the kinds of distortions thatcan be created by mixing methodologies, theauthor of this study gained access to the fullconfidential record of one of the mixed deter-

minations in the sample. The investigation inquestion was of static random access memory(SRAM) semiconductors from Taiwan, andthe specific company examined was IntegratedSilicon Solution, Inc (ISSI).32 The actual com-pany data submitted in the investigation andthe dumping margin calculation programemployed by the Commerce Department inthe final determination were used to recalcu-late ISSI’s dumping margin; the computerprogram was altered so that only normal price-to-price comparisons were made.33 As a result,the company’s dumping margin fell by almosttwo-thirds, from 7.56 percent to 2.74 percent(Table 3).

In sum, a review of the actual case recordconfirms that the antidumping law as current-ly written and implemented is miserably inef-fective at identifying price discriminationcaused by sanctuary markets. In only 27 of the107 affirmative determinations, or 25.2 per-cent of that total, did Commerce make at leastsome use of the only methodology relevant todetecting price discrimination, and all but 2 ofthose determinations were distorted by resortto other methodologies. Meanwhile, in theother 80 affirmative determinations, or 74.8percent of the total, there is absolutely nothingin the Commerce Department’s findings thatin any way points to the existence of price dis-crimination.

What about the antidumping law’s trackrecord with respect to the other form of dump-ing—below-cost sales caused by market distor-tions? In as many as 100 of the 141 determina-tions in the sample, Commerce relied fully orpartially on cost-based analysis. Nearly half ofthe determinations—67 of 141—dependexclusively on comparisons of U.S. prices tosome cost-based benchmark of normal value.In 20 of those cases, Commerce used the for-eign producer’s own cost information to calcu-late constructed value;34 the remaining 47 wereNME cases in which Commerce calculatedcosts using surrogate-country values. In anadditional 33 determinations, Commercemade at least some use of constructed value inits calculations, although perhaps not in everydetermination.35

In only 27 of the107 affirmativedeterminationsdid Commercemake at least someuse of the onlymethodologyrelevant todetecting pricediscrimination.

10

The most obvious problem with all of thecost-based determinations is that they do notattempt to measure whether U.S. sales arebelow cost. As discussed above, they measureinstead whether U.S. sales are below somemeasure of cost plus profit. Because of theinclusion of profit, sales can be considereddumped even when they are above cost, andthe dumping margins of below-cost sales areexaggerated.

For specific examples of how this method-ological distortion affects dumping margins,access was gained to the confidential caserecords of two cost-based determinations: PTDieng Djaya/PT Surya Jaya Abadi Perkasa(Dieng/Surya Jaya), a respondent in the inves-tigation of preserved mushrooms fromIndonesia,36 and China Metallurgical Import &Export Liaoning Company (Liaoning), arespondent in the investigation of cut-to-length steel plate from China. For both deter-minations, the dumping margin was recalculat-ed by setting profit equal to zero.37 Dieng/Surya Jaya’s dumping margin fell from 7.94percent to 4.88 percent, and Liaoning’s rateplunged from 17.33 percent to 5.43 percent(Table 3).

Even if subnormal profitability, rather thansales below cost, is taken to be the appropriate

threshold indicator of “unfair” trade, currentantidumping practice still exaggerates dump-ing margins. The Commerce Department’scalculation methodologies are biased in favor offinding U.S. sales to be insufficiently profitable.

Most obviously, the profit rates used byCommerce in constructed-value and NMEcases are frequently much higher than any con-ceivable industry norm. Table 4 gives a fewexamples taken from case records. It comparesthe profit rates actually used by Commerce(but expressed as a percentage of sales)38 to theaverage profit rates of the equivalent U.S.industries during the year the respective inves-tigations were initiated.39 In these cases theprofit rates used in the Commerce Depart-ment’s antidumping investigations were gross-ly excessive. Inflated profit rates translatedirectly into inflated dumping margins.

Even when Commerce uses more reason-able profit figures, its practice of comparingmodel-specific prices to product- or industry-average profit rates is skewed in favor of higherdumping margins. Consider a hypotheticalantidumping investigation of widgets, in whichCommerce determines the “normal” profit rateto be 5 percent. The foreign producer in thecase had equal sales of three different models ofwidget—Models A, B, and C. It averaged a 1

Table 3How Dumping Margins Are Inflated

Methodological Commerce’s CorrectedCompany Investigation Distortion Result (%) Result (%)

ISSI SRAMs from Mixing cost-based and 7.56 2.74Taiwan price-to-price methodologies

Dieng/ Preserved mushrooms Inclusion of profit in 7.94 4.88Surya Jaya from Indonesia below-cost investigation

Comparison of model-specific 7.94 0profits to product-wideprofit benchmark

Failure to examine whether 7.94 0.04sales are above variable costs

Liaoning Cut-to-length steel plate Inclusion of profit in 17.33 5.43from China below-cost investigation

Failure to examine whether 17.33 0sales are above variable costs

Sales can be con-sidered dumped

even when they areabove cost, and the

dumping marginsof below-cost sales

are exaggerated.

11

percent profit on U.S. sales of Model A, a 4 per-cent profit for Model B, and a 10 percent prof-it for Model C. Its average profit margin wasthus 5 percent, or equal to the Commercebenchmark. Nevertheless, Commerce deter-mines dumping model by model and treats“negative” dumping margins (i.e., instances inwhich the U.S. price is higher than normalvalue) as equal to zero. Accordingly, it concludesthat sales of Models A and B are dumped.

The case of Dieng/Surya Jaya, theIndonesian producer of mushrooms discussedabove, provides an example of the effect of thisdistortion in actual practice. For purposes ofthis study, the company’s dumping margin wasrecalculated by subtracting “negative” dumpingmargins from the positive margins.40 Therevised dumping calculation makes a properapples-to-apples comparison of product-wideprofitability to a product-wide profit bench-mark, as opposed to the normal method ofcomparing model-specific profitability to aproduct-wide benchmark. In the revised calcu-lation, Dieng/Surya Jaya’s dumping margincompletely disappears: it drops from 7.94 per-cent to zero (Table 3).

Market Distortions Assumed,Not Proven

The evidence reviewed thus far shows that

the antidumping law is highly prone to findingdumping even when there is no price discrim-ination or selling below cost. But there isanother, deeper problem with the law. Namely,it simply assumes that those pricing practices,when found, indicate the existence of govern-ment-caused market distortions. As shownbelow, this assumption is entirely unwarranted.

It is true that international price differencescan reveal a sanctuary home market. Likewise,sales below cost, under certain circumstances,can signal the presence of government-causedmarket distortions. But just because they candoes not mean that they usually do. There aremany other possible explanations—explana-tions that rest entirely on normal business prac-tices and have nothing to do with any “unfair”competitive advantage. By ignoring alternativecauses of the pricing behavior it targets, theantidumping law routinely punishes foreignfirms for normal commercial conduct.

Price Differences and Sanctuary MarketsAs to the connection between affirmative

dumping findings and the existence of sanctu-ary markets, consider Table 5. It identifies, foreach of the 18 investigations in the sample inwhich Commerce relied at least partially onprice-to-price comparisons, the primary U.S.Harmonized Tariff System 10-digit numberunder investigation.41 It then compares the tar-iff rates for that product in the United States

Table 4Comparison of Profit Rates

Company/ Commerce U.S IndustryInvestigation Rate (%) Rate (%)

Chen Hao Taiwan/ 25.77 5.23Dinnerware from Taiwan

Brake drums and rotors 12.50 5.93from China

Cut-to-length steel plate 10.14 3.43from China

PT Multi Raya/ 22.61 5.23Dinnerware from Indonesia

Collated roofing nails 20.50 7.20from China

By ignoring alter-native causes of thepricing behavior ittargets, theantidumping lawroutinely punishesforeign firms fornormal commercialconduct.

12

and the corresponding product in the relevanthome market at the time of the investigation.42

The upshot of this comparison is that inonly 3 of the 18 investigations was the home-market tariff rate greater than 10 percentagepoints higher than the U.S. rate. In only 2 addi-tional cases was the home-market rate morethan 5 percentage points higher than the U.S.rate. In short, as least as far as the most obviousform of protectionism is concerned, there is noevidence that the home market is significantlymore protected than the U.S. market in the vastmajority of relevant cases. Furthermore, there isno correlation between the degree of relativeprotection in the home market and the range ofdumping margins found.43

Especially interesting is the case of open-end spun rayon singles yarn from Austria. Thiswas the only investigation in the entire sampleof 49 in which Commerce made affirmativedumping determinations strictly on the basis ofcomparing U.S. and home-market prices. Andyet in this case, the U.S. tariff rate at the timeof the investigation was actually higher than theAustrian rate.

It is possible, of course, that some of theseforeign product markets may be shielded fromforeign competition by nontariff barriers. Ifsuch barriers were significant, however, onewould expect that they would merit inclusionin the U.S. Trade Representative’s annual com-pendium of foreign trade barriers, the NationalTrade Estimates report. A review of the NTEreports for 1995–98 found allegations thatmight pertain to 2 of the 18 relevantantidumping investigations.44 With respect tothe other 16 cases, though, the NTE reports donot even allege (much less prove) the existenceof protectionist policies that would create sanc-tuary markets.

Even if a foreign producer does enjoy sig-nificantly more protection in its home marketthan U.S. companies do here at home, the casefor an “artificial” and “unfair” competitiveadvantage still has not been established.Although the foreign producer may be able tocharge higher prices at home, it may also beburdened by higher costs; accordingly, its prof-itability may not be superior to that of its U.S.competitors. And even if a company is earning

Table 5

Comparison of Tariff Rates

U.S. Home DumpingCase Name HTS No. Rate (%) Rate (%) Margins (%)

Polyvinyl alcohol from Taiwan 3905.20.00.00 3.2 5.0 19.21Certain pasta from Italy 1902.19.20.00 0.0 11.3 + 31 ecu/100 kg 0.00 - 19.09Framing stock from United Kingdom 3924.90.20.00 3.4 6.5 0.00 - 20.01Dinnerware products from Indonesia 3924.10.20.00 3.4 30 8.95Dinnerware products from Taiwan 3924.10.20.00 3.4 5.0 0.00 - 3.25Reinforcing bars from Turkey 7214.20.00.00 3.9 15.0 9.84 - 18.68Rayon singles yarn from Austria 5510.11.00.00 10.6 7.5 2.36 - 12.36Steel plate from South Africa 7208.52.00.00 4.8 5.0 26.01 - 50.87Steel wire rod from Canada 7213.91.30.00 1.3 0.6 0.91 - 11.94SRAM semiconductors from Korea 8542.13.80.49 0.0 8.0 1.00 - 5.08SRAM semiconductors from Taiwan 8542.13.80.49 0.0 1.0 7.56 - 93.71Steel wire rod from Trinidad and Tobago 7213.91.30.00 1.3 10.0 11.85Stainless steel wire rod from Italy 7221.00.00.15 3.3 4.2 1.27 - 12.73Stainless steel wire rod from Japan 7221.00.00.15 3.3 3.2 21.18 - 34.21Stainless steel wire rod from Korea 7221.00.00.15 3.3 7.0 5.19Stainless steel wire rod from Spain 7221.00.00.15 3.3 4.2 4.73Stainless steel wire rod from Sweden 7221.00.00.15 3.3 4.2 5.71Stainless steel wire rod from Taiwan 7221.00.00.15 3.3 7.5 0.02 - 8.29

Notes: HTS = U.S. Harmonized Tariff System; SRAM = static random access memory

There is no cor-relation between

the degree of rela-tive protection inthe home marketand the range of

dumping marginsfound.

13

supranormal profits, it does not gain any sig-nificant advantage if its domestic market ismuch smaller than its U.S. market. A highprofit rate earned on relatively few sales willnot provide a sufficient “war chest” to offersignficant opportunities for subsidizing itsU.S. sales.

For a concrete illustration of these issues,consider again the case of ISSI, one of therespondents in the investigation of SRAMsfrom Taiwan. As already discussed, most of itsdumping margin was due to deviations from apure price-to-price comparison. But does itsremaining dumping margin of 2.74 percentprovide any evidence of government-causedmarket distortions? As shown in Table 5, theTaiwan SRAM market was not overtly pro-tected: the tariff rate at the time of the investi-gation was only 1 percent. Assuming for thesake of argument that other “hidden” barriersdid in fact shield the Taiwan market, theantidumping investigation nonetheless reveal-ed that ISSI was not enjoying supranormalprofits in Taiwan. The profit rate on ISSI’sabove-cost-only sales in Taiwan was only 7.61percent of sales; by comparison, the averageprofit rate for the U.S. electrical and electronicsproducts industry in 1997 was 10.85 percent.45

Meanwhile, even if ISSI had been earninginflated profits in Taiwan, the fact is that ISSI’sTaiwan sales during the period of investigationwere only about 40 percent of its U.S. sales invalue terms.46 Consequently, its Taiwan marketwas not sufficiently large to serve as a base forsubsidizing export sales.

The lack of connection between affirmativedumping determinations and evidence of sanc-tuary markets is not surprising. As discussedabove, the methodological flaws in pure price-to-price comparisons, compounded by thepractice of using both price-to-price and cost-based comparisons in a single case, can result infindings of dumping even when there is no realpattern of international price differences.

Furthermore, even when antidumpinginvestigations do stumble onto cases of actualprice discrimination, they are incapable of dis-tinguishing between those that reflect the exis-tence of a sanctuary market and those that are

attributable to normal commercial factors.There are in fact many unexceptionable busi-ness reasons for charging more in one marketthan in another, and the persistence of thoseprice differences over time by no means provesthat the high-price market is closed.

International price differences can arisewhen a firm’s status differs between nationalmarkets. A consumer goods firm may enjoybrand recognition in its home market thatallows it to command a premium price, whileabroad its brand name may be less valuable.Similarly, a producer goods firm may have builta reputation at home as a reliable supplier ofhigh-quality products, while remaining a rela-tive unknown in foreign markets. Or it mayhave carefully cultivated long-term businessrelationships with its domestic customers,while serving export markets on more of aspot-market basis. In all of those situations, thefirm is exposed to greater pricing pressureabroad than at home and consequently will beforced to accept a lower price on its exportsales.

Business strategists recognize that, whetherin domestic or international markets, estab-lished “incumbents” enjoy a built-in competi-tive advantage over new market entrants. AsMichael Porter, a leading expert on businessstrategy, puts it:

Product differentiation means thatestablished firms have brand identifica-tion, and customer loyalties, which stemfrom past advertising, customer service,product differences, or simply being firstinto the industry. Differentiation createsa barrier to entry by forcing entrants tospend heavily to overcome existing cus-tomer loyalties.47

New entrants can offset the incumbent’sadvantage and wrest away market share byintroducing a superior new product, or byadvertising frequently or especially effectively,or by offering a lower price. In the internationalsetting, if a new entrant in an export marketenjoys an incumbent position at home, it maywell find that the most effective way for it to

There are in factmany unexception-able business rea-sons for chargingmore in one mar-ket than inanother.

14

gain ground abroad is by pricing more aggres-sively than it does in the domestic market.

Price differences can also result when themarket structures or conditions in which a firmmust operate vary from country to country. Forexample, market concentration may be higherin the firm’s home market than abroad, andpricing pressures may consequently be lesssevere. Or numerical market concentrationmay have nothing to do with it; the vagaries ofbusiness culture and market history may com-bine to render a firm’s home market less proneto aggressive price cutting than a particularexport market. The distinction here is notbetween competitive behavior in one marketand anticompetitive behavior in another;rather, it is a matter of competitive rivalry ofgreater or less intensity.

Variations in competitive intensity among afirm’s customers are also capable of producinginternational price differences. Consider theexample of a foreign manufacturer that sells tosmall, traditional, family-owned distributors inthe home market and highly sophisticated,nationwide retail chains in the United States.The manufacturer’s bargaining position will bemuch weaker when facing a Wal-Mart or aHome Depot than when dealing with a mom-and-pop wholesaler back home; as a result, theprices it charges in the United States are likelyto be lower than those in the domestic market.48

All of the sources of international price dif-ferences discussed above boil down to differ-ences in market power. When a company hasgreater market power in one country thananother—whether due to brand recognition,reputation, the business decisions of its rivals,or the bargaining positions of its customers—itwill be able to command a higher price. Theresulting price differences across national mar-kets reflect purely commercial factors and havenothing to do with government intervention orsanctuary markets.

Not only differences in market power, butdifferences in marketing strategy as well, cancreate price gaps between countries. InCountry A a manufacturer may choose to mar-ket its products (say, cosmetics) as premiumgoods: its strategy is to sell limited volumes at

high prices through a carefully selected upscaledistribution network. Meanwhile, in CountryB the same manufacturer may opt to sell thevery same products as mass-market items: thistime, the strategy is to sell high volumes at lowprices through mass-merchandise outlets.Price points in Countries A and B will be verydifferent, but again sanctuary markets will notbe to blame.

Antidumping supporters argue that thesekinds of commercially caused price differencesare unsustainable: without government-imposed market barriers, they say, all suchprice differences will simply be arbitragedaway. Savvy entrepreneurs in the low-pricemarkets will buy up goods and sell them in thehigh-price markets; increased demand in theformer and increased supply in the latter willcause prices to converge somewhere in themiddle.

Such a scenario makes sense in theory, butin practice things don’t work quite so smooth-ly. It is true that price differences will createincentives for arbitrage, but taking advantageof arbitrage opportunities entails costs. Mostobviously, there are the costs of shipping goodsto the high-price market. In addition, there areall kinds of hidden transaction costs: the costof identifying the price differences in the firstplace, the cost of obtaining supplies in the low-price market, and the cost of finding willingbuyers in the high-price market. Those costsmay not be significant for fungible commodi-ties with well-established spot markets andpublic prices, but for other commodities theyare capable of overwhelming the price gaps inquestion. When transportation costs are sig-nificant, when prices are negotiated and treat-ed as trade secrets, or when distribution isdominated by established relationships andlong-term contracts, price differences acrossnational markets can easily persist in theabsence of government-imposed trade barri-ers. The normal marketplace frictions of rela-tively illiquid product markets can suffice toprevent the forces that push toward priceequalization from reaching their logical, text-book conclusion.

For empirical evidence in support of this

Price differencesacross national

markets can easilypersist in the

absence of govern-ment-imposedtrade barriers.

15

proposition, consider the billions of dollars of“gray-market” imports that flow into theUnited States every year.49 Gray-market goods,also known as parallel imports, are copyrighted,trademarked, or patented products that enterthe country without the intellectual propertyowner’s permission. Often those goods are“reverse imports”—products originally export-ed to other markets and then imported backthrough unauthorized channels. Why do thegoods boomerang? They come back becauseprices are higher in the United States. As alegal commentator explains:

Parallel importations occur becauseof price differentials in the global mar-ketplace. A publisher of computer soft-ware may, for example, have only a smallmarket share in Mexico. As a businessstrategy, that publisher may legitimatelydecide to introduce a new product intothe Mexican market at a substantial dis-count compared to the sales price for thesame product in the United States. If thediscount is large enough, U.S. parties areable to purchase the software in Mexicoand import it into the United States forresale at a discount over the same prod-uct in authorized channels.

In other cases, a manufacturer maylimit its retail distribution to upscalemarkets. This strategy is common in thecosmetics trade, in which some productsare sold only through salons or selectedstores. Discount retailers who would liketo sell the same product often find it onsale abroad at deeply discounted prices.50

The existence of gray-market importsrefutes the assumption that international pricedifferences require government intervention inthe home market. Reverse imports show that,for some products, prices are higher in the rel-atively open U.S. market than elsewhere, andthus that price differentials can arise withoutgovernment-imposed barriers to competition.Furthermore, the fact that gray-market importsof certain products persist year after year provesthat price gaps can continue even in the face of

arbitrage activity. In other words, arbitrage can-not always be counted on to achieve full priceconvergence.

Sales below Cost and Market DistortionsJust as price discrimination can reflect the

existence of market distortions, so can below-cost pricing be associated with “abnormal”market behavior. First, sales below marginalcost generally do not make commercial sense;while sales above marginal cost (but below fullunit cost) at least make some contribution torecovering sunk costs, sales below marginal costonly compound total losses and therefore arealmost always to be avoided. Likewise, firmscannot normally sell below full unit costs for aprotracted period of time. Over the long term,chronic loss-making firms cannot attract thecapital needed to stay in business. In these sce-narios, firms exhibiting a pattern of makinglosses—whether of the acute, marginal costvariety or the chronic, below-unit-cost vari-ety—may be benefiting from some form ofgovernment intervention that allows them toignore normal market signals.

The usual reason for sales at a loss is noth-ing other than a normal, healthy, competitivemarketplace. Here in this country, for example,of the 4.47 million U.S. corporations that filedtax returns in 1995, only 2.46 million—or 55percent—reported any net income.51 Even themightiest corporations are not immune fromred ink. General Motors lost money three yearsin a row in 1990-92, with accumulated pre-taxlosses during that period of $11.4 billion. IBMposted two straight years of negative earningsin 1992 and 1993, racking up a staggering$17.8 billion of pre-tax losses—14 cents in thered for every dollar of sales.52

Sales at a loss can indicate all kinds of nor-mal market phenomena. Companies that aregoing out of business generally leave a trail ofred ink on the way out. Other times, losses areonly temporary, as companies make mistakes orbusiness conditions deteriorate; companies canget back in the black by correcting errors andriding out the storm. During down periods, itmay make good business sense to go on pro-ducing at a loss instead of cutting back produc-

The usual reasonfor sales at a loss isnothing other thana normal, healthy,competitivemarketplace.

16

tion. For example, there may be long-termstrategic benefits that accompany a certainmarket position (and market share); staying themarket leader through a temporary downturnmay maximize long-term profitability.

Also, if a downturn is seen as too temporaryto justify permanent capacity cutbacks, it maypay to continue producing instead of allowingcapacity to go idle. Here the distinctionbetween marginal and sunk costs, or their real-world equivalents of variable and fixed costs, iscrucial. If a company can continue to produceand sell goods above variable costs, it can atleast make some contribution to fixed costs—costs that would be incurred even if thosegoods had not been produced. Under theseconditions—which are typical for industriesthat face cyclical peaks and troughs ofdemand—continuing to produce and sell min-imizes total losses during the downturn.

For young companies, losses are not justcommon; they are the norm. Investment mustcome first, followed (eventually, if all goes well)by returns on that investment. In these circum-stances, even fast-growing companies can gen-erate significant red ink. Consider the case ofAmazon.com, the online retailer. In less thanfour years it has grown into the nation’s third-largest bookseller, yet it has never made a prof-it. In 1998, while sales more than quadrupledfrom $147.8 million to $610 million, Amazon’snet loss was a colossal $124.5 million.53 Thoselosses are part of Amazon’s business strategy: togrow as fast as possible and establish the mar-ket leadership that will bring long-term prof-itability.54 The strategy may succeed or fail, butthat is purely a commercial matter; governmentinterventionism is irrelevant.

Even for established companies, losses arecommon on new products. By virtue of thewell-known phenomenon of the “learningcurve,” production costs tend to decline in linewith cumulative production volume. Knowingthis, businesses often price new goods belowfull current cost in order to increase sales vol-umes and accelerate passage down the learningcurve. Such a strategy is intended to maximizeprofitability over the full life cycle of the prod-uct. This practice of “forward pricing” is partic-

ularly well known in high-tech products likesemiconductors, but learning curves have beenfound in a wide variety of industries.55

Eventually, of course, companies must turna profit on their overall operations if they are tostay in business. Likewise, specific productsmust generally earn a profit sooner or later orelse be dropped from a company’s business line.There are, however, important exceptions. Onsome products, companies can lose moneyindefinitely; indeed, under certain conditionsthey may be well advised to do so.

For example, a multiproduct firm mayintentionally charge a money-losing price forone good to encourage higher sales of anothergood. Such a “cross-subsidization” strategy, ifsuccessful, can actually maximize overall firmprofits. Michael Porter explains:

When a firm offers products thateither are complementary in the strictsense of being used together or are pur-chased at the same time, pricing canpotentially exploit the relatedness amongthem. The idea is to deliberately sell oneproduct (which I term the base good) ata low profit or even a loss in order to sellmore profitable items (which I termprofitable goods).

The term “loss leadership” is com-monly used to describe the application ofthis concept in retailing. Some productsare priced at or below cost in order toattract bargain-conscious buyers to thestore. The hope is that these buyers willpurchase other more profitable mer-chandise during their visit. Loss leaderpricing is also a way of establishing a lowprice image for the store.

The same pricing principle is at workin the so-called “razor and blade” strate-gy, which involves complementary prod-ucts. The razor is sold at or near cost inorder to promote future sales of prof-itable replacement blades. The samestrategy is also common in amateur cam-eras, aircraft engines, and elevators. . . .

Another variation of cross-subsidiza-tion is a trade-up strategy. Here product

On some prod-ucts, companies can

lose money indefi-nitely; indeed,

under certain con-ditions they may be

well advisedto do so.

17

varieties that are typically first purchasesare sold at low prices, in the hopes thatthe buyer will later purchase other moreprofitable items in the line as trade-upoccurs. This strategy is sometimesemployed, for example, in light aircraft,motorcycles, copiers, and computers.56

Selling below full unit costs may also makesense in the case of so-called coproducts or jointproducts—two or more different goods that areproduced simultaneously in the same manufac-turing process. Examples include different cutsof meat from the same animal, different oresextracted in the same mining operation, differ-ent chemicals produced by the same reaction,and products of varying quality produced in thesame manufacturing batch. For those types ofproducts, some allocation of shared manufac-turing costs among the various joint products isnecessary for cost-accounting purposes.Depending on how costs are allocated, a givencoproduct may show a profit or a loss.

Accounting results, however, are ultimatelyirrelevant to proper business decisions.Managers must decide what product mix totarget and what further processing to do after“splitoff ” of the joint products; in doing so, theyshould focus not on total unit costs but on mar-ginal costs. As a leading cost-accounting text-book explains:

No technique for allocating joint-product costs should guide managementdecisions regarding whether a productshould be sold at the splitoff point orprocessed beyond splitoff. When a prod-uct is an inevitable result of a jointprocess, the decision to further processshould not be influenced either by thesize of the total joint costs or by the por-tion of joint costs allocated to particularproducts. . . .

The decision to incur additional costsbeyond splitoff should be based on theincremental operating income attainablebeyond the splitoff point.57

Joint products are manufactured from

the same raw materials, but there are manyother ways for products to share costs.Sharing of factory overhead costs (e.g., elec-tricity, fuel, maintenance, plant and equip-ment depreciation, engineering support,research and development) and selling, gen-eral, and administrative expenses is thenorm in multiproduct firms. Indeed, econo-mists explain the very existence of multi-product firms in terms of the benefits ofcost sharing, also known as economies ofscope.58 As a leading textbook on the eco-nomics of business strategy explains,“Economies of scope are usually defined interms of the relative cost of producing avariety of goods together in one firm versusseparately in two or more firms.”59 The sametextbook goes on to clarify that “theseeconomies arise because of inputs that canbe shared to produce several products.”60

The ubiquitousness of cost sharing suggeststhat a focus on product-specific total unit costs(which include overhead and selling, general,and administrative expenses) can be deceptive.A particular product that is never profitablewhen viewed in isolation may nonetheless con-tribute to fixed costs that would be incurredanyway on other, profitable products.Paradoxically, then, a perennially money-losingproduct can help to maximize firmwide profits.

In sum, sales below cost can mean manythings other than the presence of government-caused market distortions. The antidumpinglaw, however, completely ignores this possibili-ty. When below-cost sales do end up gettingcaught in the wide net thrown in constructed-value and NME cases, the CommerceDepartment’s calculation methodologies fail todistinguish between normal commercial lossesand those that point to the existence of govern-ment interventionism. As a result, antidump-ing law too often penalizes normal commercialpractices having nothing to do with anyone’sdefinition of “unfair trade.”

For the existence of below-cost sales toraise any serious question of government inter-ventionism, the losses must either be acute(i.e., sales must be below variable costs) orchronic (i.e., losses must persist for a period of

The ubiquitousnessof cost sharing sug-gests that a focuson product-specifictotal unit costs canbe deceptive.

18

years). The current antidumping law makes noattempt to identify either acute or chroniclosses.

An examination of specific cases reveals theimpact of this omission. With respect to acutelosses, the dumping margins of Liaoning(respondent in the NME investigation of cut-to-length steel plate from China) andDieng/Surya Jaya (respondent in the con-structed-value investigation of preservedmushrooms from Indonesia) were recalculatedfor purposes of this study by comparing U.S.prices to an estimate of variable costs (asopposed to full unit costs plus profit).61 AsTable 3 shows, dumping margins for bothcompanies disappeared completely: Liaoning’smargin fell from 17.33 percent to zero, andDieng/Surya Jaya’s fell from 7.94 percent to0.04 percent (de minimis). These results showthat the Commerce Department’s affirmativedumping determinations cannot be taken asreliable indicators of acute below-cost sales.

For chronic losses, the period investigatedby the Commerce Department in antidumpingcases is only 12 months.62 Consequently,Commerce lacks the evidentiary record todetermine whether a company’s losses areabnormally persistent. Because Commercedoes not take a longer view, it cannot deter-mine whether losses reflect a temporary marketdownturn or business reversal, or whether theyflow from a conscious growth-oriented strate-gy for a new company or a new product. TheCommerce Department does make someadjustment for losses incurred on new prod-ucts, but the adjustment is restricted to situa-tions in which technical factors during thestart-up phase limit production levels.63 Thereis no adjustment for losses incurred to takeadvantage of learning-curve effects, or forinvestments in growth at the expense of currentearnings.

Likewise, antidumping investigationsdevelop no evidentiary record for determiningwhether acute or chronic losses, to the extentthey exist, have a reasonable commercial expla-nation. Commerce does not examine com-bined profitability in joint product situations,or possible reasons for cross-subsidization

when goods are complementary or sharecosts.64

Most fundamental, Commerce does notexamine whether the supposedly below-costU.S. sales it identifies are in any way connectedwith government interventionism in the homemarket. There is no investigation of whethertrade barriers or other restrictions on competi-tion create a domestic sanctuary market thatbankrolls losses abroad; nor of whether the for-eign producer receives government grants, softloans, special tax breaks, preferential access tocredit on noncommercial terms, or any otherform of assistance that supports its loss-makingoperations; nor of whether there are basicstructural flaws in a country’s economic policythat impede normal market responses to losses.

On this point, the case record reviewed inthis study argues against any reliable connec-tion between constructed-value cases andunderlying market distortions. In as many as 33of the 53 possible determinations in whichCommerce relied fully or partially on con-structed value, Commerce resorted to con-structed value only because of an absence ofabove-cost home-market sales. Such anabsence is flatly inconsistent with the supranor-mal profits supposedly associated with sanctu-ary markets. Meanwhile, in another 13 con-structed-value-based determinations, Com-merce found that there was no viable homemarket at all.

In addition, there is reason to doubt thatconstructed-value cases point with any regular-ity to the existence of foreign government sub-sidies. Here it is instructive to examine theinterplay between constructed-value-basedantidumping actions and investigations underthe countervailing duty (CVD) law. The CVDlaw directly targets foreign government subsi-dies, while the antidumping law allegedly doesso indirectly by targeting pricing practices (e.g.,U.S. sales at prices below constructed value)that supposedly reflect underlying subsidies.65

Consequently, if indeed constructed-valuecases are addressing the effects of foreign gov-ernment subsidies, one would expect to findaffirmative CVD determinations with respectto the same imported products. After all,

Commerce doesnot examine

whether the sup-posedly below-costU.S. sales it identi-fies are in any way

connected withgovernment inter-ventionism in the

home market.

19

simultaneous pursuit of antidumping andCVD remedies offers the prospect of doublerelief for the same underlying market distor-tions—surely an attractive outcome from thepetitioning U.S. industry’s perspective.66

Yet only 4 of the 26 antidumping investiga-tions in which constructed value was used cov-ered products with respect to which Commercealso made affirmative CVD findings.67 Theabsence of any associated affirmative CVDfindings with respect to any of the other 22investigations calls into serious questionwhether in fact there were any market-distort-ing government policies in those cases thatwould have accounted for any below-costsales.68

Conclusion

The antidumping law is defended as a rem-edy for market distortions caused by foreigngovernment interventionism. Yet in actualpractice, the methods of determining dumpingunder the law fail, repeatedly and at multiplelevels, to distinguish between normal commer-cial pricing practices and those that reflect gov-ernment-caused market distortions.

As a result, the antidumping law as it cur-rently exists routinely punishes normal com-petitive business practices—practices com-monly engaged in by American companies athome and abroad. It is therefore not the casethat the law guarantees a “level playing field”for American companies and their foreigncompetitors. On the contrary, it actively dis-criminates against foreign goods by subjectingthem to requirements not applicable toAmerican products.

An antidumping law that actually did targetgovernment-caused market distortions wouldlook very different from the law in its presentform. Bringing the reality of antidumping prac-tice into line with the rhetoric of antidumpingsupporters would require dramatic reforms.

Price-Discrimination DumpingAn affirmative determination of price-dis-

crimination dumping would have to include allof the following findings:

1. Properly measured international pricediscrimination. A methodologically defensi-ble comparison of prices would have toreveal a significant and stable differentialbetween a foreign producer’s U.S. prices andits home-market prices. Comparisons ofU.S. prices to above-cost home-marketprices only, or to third-country prices, or toconstructed value, would have no placewhatsoever in a proper analysis. Further-more, in comparing U.S. and home-marketprices, Commerce should abandon its cur-rent practice of comparing all U.S. sales tosome benchmark of normal value; instead, itshould select representative U.S. sales forwhich there are home-market sales madeunder nearly identical circumstances.

2. Government policies that insulate theforeign producer’s domestic market from outsidecompetition. These policies could includehigh tariffs, nontariff import barriers, andgovernment support of a domestic cartel.The U.S. industry seeking relief would haveto show that the level of protectionism washigher in the foreign market than in theU.S. market—in other words, that there wasan “unlevel playing field.”

3. High profits enjoyed by the foreign pro-ducer on its domestic sales. The existence of aprotected home market alone does notguarantee a foreign producer any kind ofcompetitive advantage; after all, its costsmay be inflated because of similar competi-tive restrictions in upstream input markets.Accordingly, it must be established—withrespect to an appropriately defined productline over an appropriately defined time peri-od—that the foreign producer’s rate of prof-it is higher than the U.S. industry average.

4. A relatively large domestic market. Evenif a foreign producer is earning high profitsin a protected home market, it derives nosignificant “unfair” advantage if the sanctu-ary market is small with respect to its exportmarkets. A small domestic market offers nopotential for significant cross-subsidizationof export sales. Accordingly, the size of theforeign producer’s domestic market in valueterms should at least equal the size of its

An antidumpinglaw that actuallydid target govern-ment-caused mar-ket distortionswould look verydifferent from thelaw in its presentform.

20

U.S. market before an affirmative dumpingfinding can be made.

For legitimate NME cases (assuming thereare any such things), the absence of market-based home-market prices can be overcomewithout resort to the crude approximations ofsurrogate-country values. In lieu of price-to-price comparisons, Commerce could comparethe profit rate on U.S. sales to the profit rate inthe home market. In this context, whether ornot home-market prices and costs reflect mar-ket transactions is irrelevant. All that matters isthe company’s relative profitability in thedomestic and U.S. markets; profitability is anobjective fact regardless of how the constituentprices and costs are derived.

Below-Cost DumpingTo support an affirmative determination of

below-cost dumping, all of the following find-ings would be needed:

1. Acute or chronic below-cost sales.Commerce would have to determine that aforeign producer was engaging in an unusu-al pattern of below-cost selling. Specifically,it could find that the company was sellingbelow average variable costs over some sig-nificant period of time (e.g., one year).Alternatively, it could find that the compa-ny was selling below full unit costs for a pro-tracted period of time (e.g., at least threeyears).

2. Absence of commercial explanation forlosses. Foreign producers should be allowedto demonstrate that any acute or chroniclosses have a commercial justification.Possible explanations would include theexistence of joint products, products thatshare overhead costs, and complementaryproducts; learning-curve effects; and invest-ment in rapid growth to establish a strongmarket position. An affirmative determina-tion could not be made if the foreign pro-ducer proved that losses on the productunder investigation were part of a consciousstrategy to maximize long-term firmwideprofits.

3. Government policies that inhibit the nor-mal market consequences of acute or chronic loss-es. Such policies would include protectionistimport barriers, explicit or implicit subsidies,and basic structural measures that block nor-mal market responses to losses (e.g., poorlydeveloped bankruptcy law, severe monetarydisorder). To prevent double counting,antidumping liability should be reduced bythe full amount of any CVD duties paidwith respect to the same product.

There is no need for a special NMEmethodology in below-cost dumping cases.The issue in such cases is whether the compa-ny is losing money on its U.S. sales as a resultof market distortions back home. In decidingthat issue, it should not matter that the compa-ny’s costs do not result from market transac-tions. Regardless of how its costs are deter-mined, if a company is selling above them, itcannot be engaging in below-cost dumping.

If the antidumping law were overhauledalong the lines suggested here, claims that itserves as a remedy against international marketdistortions would be on much firmer ground.Nevertheless, fundamental questions about thelaw’s propriety would remain. Is closing ourmarkets the correct response to policy flawsabroad? Aren’t negotiations to eliminate for-eign market distortions a better approach? Andeven if international policy differences areintractable, is it wise to sacrifice the economicbenefits of open markets in the name of “fair-ness”? And on the subject of fairness, is it real-ly fair to defend a level playing field for partic-ular U.S. industries if doing so harms down-stream U.S. industries and consumers?

The point of this paper, though, is that suchfundamental questions about free trade versus“fair trade” are irrelevant to an evaluation of theantidumping law as it currently stands. Whenthe law is analyzed on the basis of what it does,as opposed to what its supporters say it does, itis clear that the law cannot be justified as a “fairtrade” measure. Free traders who attack it assuch are giving their opponents too muchcredit.

If the antidump-ing law were over-

hauled along thelines suggested

here, claims that itserves as a remedy

against interna-tional market dis-tortions would be

on much firmerground.

21

China Polyvinyl alcohol A-570-842 4/4/95 3/29/96 Guangxi 116.75% NMEChina Polyvinyl alcohol A-570-842 4/4/95 3/29/96 Sichuan 0.00% NMETaiwan Polyvinyl alcohol A-583-824 4/4/95 3/29/96 Chang Chun 19.21% HM mixedJapan Polyvinyl alcohol A-588-836 4/4/95 3/29/96 Kuraray 77.49% FAJapan Polyvinyl alcohol A-588-836 4/4/95 3/29/96 Nippon Goshei 77.49% FAJapan Polyvinyl alcohol A-588-836 4/4/95 3/29/96 Unitika 77.49% FAJapan Polyvinyl alcohol A-588-836 4/4/95 3/29/96 Shin-Etsu 77.49% FAChina Bicycles A-570-843 5/1/95 4/30/96 Bo An 0.00% NMEChina Bicycles A-570-843 5/1/95 4/30/96 CBC 2.95% NMEChina Bicycles A-570-843 5/1/95 4/30/96 CATIC 2.02% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Giant 0.67% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Hua Chin 0.00% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Merida 0.37% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Overlord 0.00% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Chitech 1.83% NMEChina Bicycles A-570-843 5/1/95 4/30/96 Universal 2.27% NMEJapan Clad steel plate A-588-838 10/25/95 5/9/96 Japan Steel Works 118.53% FARomania Circular welded nonalloy steel pipe A-485-804 5/22/95 5/14/96 Metagrimex 85.12% NMERomania Circular welded nonalloy steel pipe A-485-804 5/22/95 5/14/96 Metalexportimport 77.61% NMESouth Africa Circular welded nonalloy steel pipe A-791-803 5/22/95 5/14/96 RIH Group 117.66% FAItaly Certain pasta A-475-818 6/8/95 6/14/96 Arrighi 19.09% HM mixedItaly Certain pasta A-475-818 6/8/95 6/14/96 De Cecco 46.67% FAItaly Certain pasta A-475-818 6/8/95 6/14/96 Delverde 1.68% HM mixedItaly Certain pasta A-475-818 6/8/95 6/14/96 De Matteis 0.00% HM mixedItaly Certain pasta A-475-818 6/8/95 6/14/96 La Molisana 14.73% HM mixedItaly Certain pasta A-475-818 6/8/95 6/14/96 Liguori 11.58% HM mixedItaly Certain pasta A-475-818 6/8/95 6/14/96 Pagani 17.47% HM mixedTurkey Certain pasta A-489-805 6/8/95 6/14/96 Filiz 63.29% FATurkey Certain pasta A-489-805 6/8/95 6/14/96 Maktas 60.87% FAGermany Large newspaper printing presses A-428-821 7/27/95 7/23/96 MRD 30.72% CVGermany Large newspaper printing presses A-428-821 7/27/95 7/23/96 KBA 46.40% FAJapan Large newspaper printing presses A-588-837 7/27/95 7/23/96 Mitsubishi 62.26% CVJapan Large newspaper printing presses A-588-837 7/27/95 7/23/96 Tokyo Kikai 56.28% CV

SeisakushoUnited Kingdom Foam extruded PVC & polystyrene A-412-817 10/6/95 10/2/96 Ecoframe 20.01% HM mixed

framing stockUnited Kingdom Foam extruded PVC & polystyrene A-412-817 10/6/95 10/2/96 Robobond 0.00% HM mixed

framing stockUnited Kingdom Foam extruded PVC & polystyrene A-412-817 10/6/95 10/2/96 Magnolia 84.82% FA

framing stockIndonesia Melamine institutional dinnerware A-560-801 3/1/96 1/13/97 Mayer Crocodile 12.90% FA

productsIndonesia Melamine institutional dinnerware A-560-801 3/1/96 1/13/97 Multiraya 8.95% HM mixed

productsChina Melamine institutional dinnerware A-570-844 3/1/96 1/13/97 Chen Hao Xiamen 0.46% NME

productsChina Melamine institutional dinnerware A-570-844 3/1/96 1/13/97 Gin Harvest 0.47% NME

productsChina Melamine institutional dinnerware A-570-844 3/1/96 1/13/97 Sam Choan 0.04% NME

productsChina Melamine institutional dinnerware A-570-844 3/1/96 1/13/97 Tar Hong Xiamen 2.74% NME

products

Appendix:U.S. Antidumping Investigations, 1995-98

Inv. Initiation Final Country Product No. Date Date Respondent Rate Methodology

continued

22

Taiwan Melamine institutional dinnerware A-583-825 3/1/96 1/13/97 Chen Hao Taiwan 3.25% HM mixedproducts

Taiwan Melamine institutional dinnerware A-583-825 3/1/96 1/13/97 Yu Cheer 0.00% HMproducts

Taiwan Melamine institutional dinnerware A-583-825 3/1/96 1/13/97 IKEA 53.13% FAproducts

Taiwan Melamine institutional dinnerware A-583-825 3/1/96 1/13/97 Gallant 53.13% FAproducts

Kazakstan Beryllium metal A-834-805 4/9/96 1/17/97 Ulba 16.56% NME& high beryllium alloys

China Brake drums A-570-845 4/3/96 2/28/97 CMC 0.00% NMEChina Brake drums A-570-845 4/3/96 2/28/97 Qingdao 0.00% NMEChina Brake drums A-570-845 4/3/96 2/28/97 Xinchangyuan 0.00% NMEChina Brake drums A-570-845 4/3/96 2/28/97 Yantai 0.00% NMEChina Brake rotors A-570-846 4/3/96 2/28/97 CAIEC & Laizhou 0.00% NME

CAPCOChina Brake rotors A-570-846 4/3/96 2/28/97 Shenyang and 0.00% NME

LaizhouChina Brake rotors A-570-846 4/3/96 2/28/97 Xinjiang 0.00% NMEChina Brake rotors A-570-846 4/3/96 2/28/97 Yantai 3.56% NMEChina Brake rotors A-570-846 4/3/96 2/28/97 Southwest 16.07% NMETurkey Certain steel concrete A-489-807 4/4/96 3/4/97 Colakoglu 9.84% HM mixed

reinforcing barsTurkey Certain steel concrete A-489-807 4/4/96 3/4/97 Ekinciler 18.68% HM mixed

reinforcing barsTurkey Certain steel concrete A-489-807 4/4/96 3/4/97 Habas 18.54% CV

reinforcing barsTurkey Certain steel concrete A-489-807 4/4/96 3/4/97 IDC 41.80% FA

reinforcing barsTurkey Certain steel concrete A-489-807 4/4/96 3/4/97 Metas 30.16% HM/CV

reinforcing barsAustria Open-end spun rayon singles yarn A-433-807 9/13/96 3/26/97 Linz 12.36% HMAustria Open-end spun rayon singles yarn A-433-807 9/13/96 3/26/97 Borckenstein 2.36% HMJapan Engineered process gas turbo- A-588-840 6/4/96 5/5/97 Mitsubishi Heavy 38.32% CV

compressor systems IndustriesChina Persulfates A-570-847 8/6/96 5/19/97 Wuxi 34.41% NMEChina Persulfates A-570-847 8/6/96 5/19/97 AJ 32.22% NMEChina Persulfates A-570-847 8/6/96 5/19/97 Guangdong 34.97% NMEChina Freshwater crawfish tail meat A-570-848 10/17/96 8/1/97 China Everbright 156.77% NMEChina Freshwater crawfish tail meat A-570-848 10/17/96 8/1/97 Binzhou 119.39% NMEChina Freshwater crawfish tail meat A-570-848 10/17/96 8/1/97 Huaiyin FTC 91.50% NMEChina Freshwater crawfish tail meat A-570-848 10/17/96 8/1/97 Yangchen FTC 108.05% NMEJapan Vector supercomputers A-588-841 8/23/96 8/28/97 Fujitsu 173.08% FAJapan Vector supercomputers A-588-841 8/23/96 8/28/97 NEC 454.00% FAChina Collated roofing nails A-570-850 12/20/96 10/1/97 Top United 0.00% NMEChina Collated roofing nails A-570-850 12/20/96 10/1/97 Qingdao Zongxun 0.00% NMEKorea Collated roofing nails A-580-827 12/20/96 10/1/97 Senco 0.00% 3C Korea Collated roofing nails A-580-827 12/20/96 10/1/97 Kabool 0.00% CVTaiwan Collated roofing nails A-583-826 12/20/96 10/1/97 Unicatch 0.07% CVTaiwan Collated roofing nails A-583-826 12/20/96 10/1/97 Lei Chu 0.00% CVTaiwan Collated roofing nails A-583-826 12/20/96 10/1/97 S&J 2.98% CVTaiwan Collated roofing nails A-583-826 12/20/96 10/1/97 Romp 40.28% FATaiwan Collated roofing nails A-583-826 12/20/96 10/1/97 K.Ticho 40.28% FA

Appendix—continued

Inv. Initiation Final Country Product No. Date Date Respondent Rate Methodology

23

China Cut-to-length carbon steel plate A-570-849 12/3/96 11/20/97 Anshan 30.68% NMEChina Cut-to-length carbon steel plate A-570-849 12/3/96 11/20/97 Baoshan 34.44% NMEChina Cut-to-length carbon steel plate A-570-849 12/3/96 11/20/97 Liaoning 17.33% NMEChina Cut-to-length carbon steel plate A-570-849 12/3/96 11/20/97 Shanghai Pudong 38.16% NMEChina Cut-to-length carbon steel plate A-570-849 12/3/96 11/20/97 WISCO 128.59% NMESouth Africa Cut-to-length carbon steel plate A-791-804 12/3/96 11/20/97 Highveld 26.01% HM mixedSouth Africa Cut-to-length carbon steel plate A-791-804 12/3/96 11/20/97 Iscor 50.87% HM mixedRussian Fed. Cut-to-length carbon steel plate A-821-808 12/3/96 11/20/97 Severstal 53.81% NMEUkraine Cut-to-length carbon steel plate A-823-808 12/3/96 11/20/97 Azovstal 81.43% NMEUkraine Cut-to-length carbon steel plate A-823-808 12/3/96 11/20/97 Ilyich 155.00% NMEKorea Static random access memory A-580-828 3/21/97 2/23/98 Samsung 1.00% HM mixed

semiconductorsKorea Static random access memory A-580-828 3/21/97 2/23/98 Hyundai 5.08% HM mixed

semiconductorsKorea Static random access memory A-580-828 3/21/97 2/23/98 LG Semicon 55.36% FA

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 Advanced 113.85% FA

semiconductors MicroelectronicsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 Alliance 50.15% HM/CV

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 BIT 113.85% FA

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 ISSI 7.56% HM mixed

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 TI-Acer 113.85% FA

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 UMC 93.71% HM mixed

semiconductorsTaiwan Static random access memory A-583-827 3/21/97 2/23/98 Winbond 101.53% FACanada Steel wire rod A-122-826 3/24/97 2/24/98 Ispat-Sidbec 11.94% HM mixedCanada Steel wire rod A-122-826 3/24/97 2/24/98 Ivaco 6.95% HM mixedCanada Steel wire rod A-122-826 3/24/97 2/24/98 Stelco 0.91% HM mixedGermany Steel wire rod A-428-822 3/24/97 2/24/98 Brandenburg 153.10% FAGermany Steel wire rod A-428-822 3/24/97 2/24/98 IHSW 72.51% FAGermany Steel wire rod A-428-822 3/24/97 2/24/98 Saarstahl 153.10% FAGermany Steel wire rod A-428-822 3/24/97 2/24/98 Thyssen 153.10% FATrinidad &Tobago Steel wire rod A-274-802 3/24/97 2/24/98 CIL 11.85% HM mixedVenezuela Steel wire rod A-307-813 3/24/97 2/24/98 Sidor 66.75% FAChile Fresh Atlantic salmon A-337-803 7/10/97 6/9/98 Aguas Claras 5.44% CVChile Fresh Atlantic salmon A-337-803 7/10/97 6/9/98 Camanchaca 0.16% CVChile Fresh Atlantic salmon A-337-803 7/10/97 6/9/98 Eicosal 10.69% CVChile Fresh Atlantic salmon A-337-803 7/10/97 6/9/98 Mares Australes 2.23% 3C mixedChile Fresh Atlantic salmon A-337-803 7/10/97 6/9/98 Marine Harvest 1.36% CVGermany Stainless steel wire rod A-428-824 8/26/97 7/29/98 Krupp 21.28% FAGermany Stainless steel wire rod A-428-824 8/26/97 7/29/98 BGH Edelstahl 21.28% FAItaly Stainless steel wire rod A-475-820 8/26/97 7/29/98 Valbruna 1.27% HMItaly Stainless steel wire rod A-475-820 8/26/97 7/29/98 CAS 12.73% HM mixedJapan Stainless steel wire rod A-588-843 8/26/97 7/29/98 Daido 34.21% HM mixedJapan Stainless steel wire rod A-588-843 8/26/97 7/29/98 Nippon Steel 21.18% HM mixedJapan Stainless steel wire rod A-588-843 8/26/97 7/29/98 Hitachi Metals 0.00% CVJapan Stainless steel wire rod A-588-843 8/26/97 7/29/98 Sanyo Special Steel 34.21% FAJapan Stainless steel wire rod A-588-843 8/26/97 7/29/98 Sumitomo 34.21% FAKorea Stainless steel wire rod A-580-829 8/26/97 7/29/98 Dongbang/Changwon 5.19% HM mixed

/POSCO

Inv. Initiation Final Country Product No. Date Date Respondent Rate Methodology

continued

24

Korea Stainless steel wire rod A-580-829 8/26/97 7/29/98 Sammi Steel 28.44% FASpain Stainless steel wire rod A-469-807 8/26/97 7/29/98 Roldan 4.73% HM mixedSweden Stainless steel wire rod A-401-806 8/26/97 7/29/98 Fagersta 5.71% HM mixedTaiwan Stainless steel wire rod A-583-828 8/26/97 7/29/98 Walsin Cartech 8.29% HM mixedTaiwan Stainless steel wire rod A-583-828 8/26/97 7/29/98 Yieh Hsing 0.02% HM mixedChile Certain preserved mushrooms A-337-804 2/2/98 10/22/98 Nature's Farm 148.51% 3C/CV

Products China Certain preserved mushrooms A-570-851 2/2/98 12/31/98 China Processed 121.47% NMEChina Certain preserved mushrooms A-570-851 2/2/98 12/31/98 Tak Fat 162.47% NMEChina Certain preserved mushrooms A-570-851 2/2/98 12/31/98 Shenzen Cofry 151.15% NMEIndia Certain preserved mushrooms A-533-813 2/2/98 12/31/98 Agro Dutch 6.28% 3C mixedIndia Certain preserved mushrooms A-533-813 2/2/98 12/31/98 Ponds 14.91% 3C/CVIndia Certain preserved mushrooms A-533-813 2/2/98 12/31/98 Alpine Biotech 243.87% FAIndia Certain preserved mushrooms A-533-813 2/2/98 12/31/98 Mandeep 243.87% FAIndonesia Certain preserved mushrooms A-560-802 2/2/98 12/31/98 Dieng /Surya Jaya 7.94% CVIndonesia Certain preserved mushrooms A-560-802 2/2/98 12/31/98 Zeta 22.84% CV

Appendix—continued

Inv. Initiation Final Country Product No. Date Date Respondent Rate Methodology

25

Notes

1. Although the U.S. antidumping law dates back to1921, dozens of other countries now have similarlaws. The authority of national governments toimpose antidumping duties is recognized underArticle VI of the General Agreement on Tariffs andTrade and the World Trade OrganizationAgreement on Implementation of Article VI.

2. This paper focuses exclusively on what constitutes“dumping”; it does not address the effect of“dumped” imports on competing U.S. industries orthe broader U.S. economy. Accordingly, this paper isconcerned solely with the phase of antidumpinginvestigations administered by the CommerceDepartment; it does not cover issues pertaining tothe International Trade Commission’s injuryinquiry.

3. Greg Mastel, Antidumping Laws and the U.S.Economy (Armonk, N.Y.: M. E. Sharpe, 1998), p. 43.

4. Ibid., p. 41.

5. Ibid., p. 40.

6. “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,p. 2. Cited hereafter as U.S. Submission.

7. Ibid., p. 15.

8. Ronald A. Cass and Richard D. Boltuck,“Antidumping and Countervailing-Duty Law: TheMirage of Equitable International Competition,” inFair Trade and Harmonization: Prerequisites for FreeTrade? ed. Jagdish N. Bhagwati and Robert E.Hudec (Cambridge, Mass.: MIT Press, 1996), vol.2, p. 351.

9. Section 773(a)(1)(B)(i) of the Tariff Act of 1930,as amended, codified at 19 U.S.C. § 1677b(a)(1)(B)(i).

10. Section 773(a)(1)(C) of the Tariff Act of 1930,

as amended, codified at 19 U.S.C. § 1677b(a)(1)(C).

11. Section 773(a)(1)(B)(ii) of the Tariff Act of1930, as amended, codified at 19 U.S.C. § 1677b(a)(1)(B)(ii).

12. 19 C.F.R. § 351.405(a) (1999).

13. Ibid. According to Commerce Departmentpractice, a product sold in the comparison marketwill normally not be considered “comparable” (inother words, sufficiently similar) to a product sold inthe United States if the difference in variable man-ufacturing costs between the two products is greaterthan 20 percent of the total manufacturing cost ofthe comparison-market product.

14. Section 773(b)(1) of the Tariff Act of 1930, asamended, codified at 19 U.S.C. § 1677b(b)(1).Under a recent court decision, Commerce may notresort to constructed value in such situations unlessthere are no contemporaneous above-cost sales ofany similar models of merchandise. Formerly,Commerce would use constructed value if therewere no above-cost sales of the particular identicalor similar model chosen by Commerce under itsproduct-matching criteria. In other words,Commerce must now do its product matching afterapplying the sales-below-cost test, not before. Thismethodological change increases the likelihood thatnormal value will be based on above-cost salesrather than constructed value.

15. While the distinction between nonmarket andmarket economies was sensible during the ColdWar, it is difficult to draw a clear line at presentbetween the “transition” economies of post-com-munist countries and other developing countries.Nevertheless, to date the Commerce Departmenthas revoked NME status only for the former EastGermany, Poland, and the former Yugoslavia andhas explicitly refused to do so for China, Russia, andUkraine.

16. Section 773(c) of the Tariff Act of 1930, asamended, codified at 19 U.S.C. § 1677b(c).

17. “Facts available” were formerly known as “best

26

information available.”

18. “Facts available” determinations are routine inNME cases. In those cases Commerce will calculateproducer-specific dumping margins for qualifyingcompanies, as well as a countrywide rate for otherproducers and exporters. That countrywide rate isusually based on “facts available.” By contrast, inmarket-economy cases, all uninvestigated compa-nies receive an “all others” rate equal to a weightedaverage of the investigated companies’ margins(“facts available” and de minimis margins, however,are excluded from the average).

19. 19 C.F.R. § 351.411(b) (1999).

20. Biases in antidumping methodologies arereviewed extensively in Richard Boltuck and RobertLitan, eds., Down in the Dumps: Administration ofthe Unfair Trade Laws (Washington: BrookingsInstitution, 1991).

21. To take just one obvious example, in “construct-ed export price” cases (normally, cases in which theimporter of the investigated merchandise is relatedto the foreign producer), Commerce deducts theprofit allocated to U.S. selling expenses from theU.S. price but makes no equivalent profit deductionfrom the comparison-market price. Section772(d)(3) of the Tariff Act of 1930, as amended,codified at 19 U.S.C. § 1677a(d)(3).

22. The antidumping statute authorizes theCommerce Department, in making a “facts avail-able” determination because a foreign producer hasfailed to cooperate, to “use an inference that isadverse to the interests of that party.” Section 776(b)of the Tariff Act of 1930, as amended, codified at 19U.S.C. § 1677e(b).

23. In constructed-value cases the costs used are theforeign producer’s own, whereas in NME casescosts are calculated by valuing the foreign producer’s“factors of production” according to price data froma surrogate market-economy country.

24. It can be argued that inclusion of an amount forprofit is appropriate on the ground that a “normal”profit is part of a company’s cost of capital. In other

words, a company earning a subnormal return isselling below its full economic costs, if above its fullaccounting costs. First of all, it should be noted thatantidumping supporters clearly convey the impres-sion that the law targets sales at a loss, not insuffi-cient profitability. More fundamentally, the claimthat low profitability is evidence of market distor-tions is much weaker than is the case with respect tooutright losses. Determining exactly what consti-tutes a normal profit for a given company in a givenindustry at a given time is significantly more diffi-cult than determining whether or not that companyis losing money. Moreover, low profits are generallysustainable over a much longer period than are out-right losses. Persistent failure to earn competitivereturns can undermine a company’s ability to makenecessary investments and thereby may lead eventu-ally to outright losses; it may also threaten theemployment security of the company’s manage-ment. Unlike sustained losses, though, low prof-itability in and of itself does not imperil a company’ssolvency and future as a going concern. Accordingly,even chronically low profits are much less suggestiveof “artificial” market conditions caused by govern-ment interventionism than are either acute orchronic losses.

25. Interestingly, among the critics of the NMEmethodology is the current U.S. TradeRepresentative, although she expressed her criticismbefore holding public office. Charlene Barshefsky,“Non-Market Economies in Transition and theU.S. Antidumping Law: Remarks on the Need forReevaluation,” Boston University International LawJournal 8, no. 2 (Fall 1990): 373-80.

26. The U.S. antidumping statute was revised innumerous respects by the Uruguay RoundAgreements Act but has not been amended since. Anumber of proposals to revise the antidumping laware currently under consideration in Congress.

27. Commerce Department investigations are spe-cific to a particular product from a particular coun-try. In a single investigation, though, Commercemay calculate separate dumping margins fornumerous different foreign producers.

28. In original antidumping investigations, any

27

dumping margin of less than 2 percent is consideredde minimis and effectively equal to zero.

29. “Dumping typically involves international pricediscrimination.” U.S. Submission, p. 8.

30. For 34 of the 36 determinations, Commercerelied entirely on “facts available”; in other words, itmade no use whatsoever of information provided bythe foreign producers. For 2 of the determinations(Maktas in the investigation of pasta from Turkeyand Winbond in the investigation of SRAMs fromTaiwan), Commmerce relied on a combination offoreign producers’ information and facts available.These two determinations were included because inboth cases the partial use of facts available was bothextensive and punitive (that is, Commerce purpose-fully selected adverse facts).

31. This figure is a subset of the 37 determinationsmentioned above in which Commerce partially ortotally rejected home-market or third-country com-parison product sales. This smaller figure excludesthose determinations in which any third-countrysales were used and those in which all comparison-market sales were rejected.

32. Interestingly, ISSI is a U.S.-based “fabless pro-ducer” that designs chips in the United States butrelies on a semiconductor “foundry” in Taiwan forproduction. In this case, then, the antidumping lawis being used by one American company (MicronInc., the petitioner) against another.

33. Specifically, the program was run without the“cost test” element, so that below-cost home-mar-ket sales were not excluded from the calculation.The recalculation was performed by ISSI’s counselat the law firm of White & Case. The revised pro-gram was reviewed by the author of this study.

34. This total consists of 16 determinations inwhich Commerce used constructed value because ofan absence of viable comparison markets or similarcomparison products, and another 4 determinationsin which publicly available information makes clearthat Commerce rejected all comparison-marketsales as below cost and therefore relied exclusivelyon constructed value.

35. The total consists of the 31 “mixed” cases dis-cussed above, plus an additional 2 determinations inwhich Commerce rejected at least some third-coun-try sales as below cost and instead relied either onabove-cost third-country sales or constructed value.In those 33 determinations, it cannot be determinedwith certainty from the public record whetherCommerce in fact used constructed value in everyinstance (such use would be unnecessary if therewere any above-cost sales of the comparison prod-ucts), nor can the extent to which constructed valuewas used be determined.

36. This respondent consists of two companies withcommon ownership whose operations are partiallyintegrated. They were treated as a single companyfor purposes of the antidumping investigation.

37. These recalculations were performed by thecompanies’ counsel at the law firm of White &Case. The revised dumping margin calculation pro-grams were reviewed by the author of this study.

38. Profit rates were derived from the followingsources: Chen Hao Taiwan, Commerce Depart-ment disclosure documents for correction of minis-terial errors, January 31, 1997; brake drums androtors from China, Commerce Department finalfactors memorandum, February 21, 1997; cut-to-length steel plate from China, CommerceDepartment final factors memorandum, October24, 1997; PT Multi Raya, Commerce Departmentdisclosure documents for correction of ministerialerrors, January 31, 1997; collated roofing nails fromChina, Commerce Department final dumping mar-gin calculation memorandum, September 23, 1997.These documents were made available by the com-panies’ counsel at the law firm of White & Case.Commerce usually calculates the profit rate as a per-centage of cost of production. To express the profitrate as a percentage of sales, the Commerce figureswere divided by one plus the Commerce profit per-centage.

39. The products under investigation and theirequivalent U.S. industries are as follows: melamineinstitutional dinnerware, rubber and miscellaneousplastic products; brake drums and rotors, motorvehicles and equipment; cut-to-length steel plate,

28

iron and steel; collated roofing nails, fabricatedmetal products. Profit rates for U.S. industries werethe average rate of pre-tax profits per dollar of salesduring the year the investigation was initiated, asreported in Bureau of the Census, “QuarterlyFinancial Report for Manufacturing, Mining, andTrade Corporations,” Table B, available atwww.census.gov/prod/www/abs/qfr-mm.html.

40. This recalculation was performed byDieng/Surya Jaya’s counsel at the law firm of White& Case. The revised dumping margin calculationprogram was reviewed by the author of this study.

41. The scope of antidumping investigations andduty orders is defined by a verbal description of the“subject merchandise,” not by U.S. HarmonizedTariff System numbers. The HTS numbers select-ed here were the ones that corresponded mostclosely with the product description; if multipleHTS numbers corresponded equally well with theproduct description, the one with the highestimports in the period immediately preceding theinitiation of the investigation was selected.

42. Since the U.S. tariff system and foreign systemsare not harmonized all the way to the 10-digit level,the foreign tariff items most closely correspondingto the relevant U.S. HTS numbers were selected.Foreign tariff rates are from the relevant country’stariff schedule for the year that the antidumpinginvestigation was initiated, with the followingexceptions: steel concrete-reinforcing bars fromTurkey (1997), SRAMs from Korea (1996),SRAMs from Taiwan (1995), stainless steel wirerod from Trinidad and Tobago (1998), and stainlesssteel wire rod from Korea (1996). U.S. tariff ratesare as of the year that the relevant antidumpinginvestigation was initiated.

43. The dumping margins included in Table 5exclude determinations calculated on the basis of“facts available,” as well as those calculated purely onthe basis of constructed value.

44. The two investigations in question are SRAMsfrom Korea and stainless steel wire rod from Japan.With respect to the former, the 1996 and 1998NTE reports charge that import licensing and

product pre-approvals impede foreign sales of elec-tronics and high-tech products to Korea, but nospecific mention of SRAMs is made. As to the lat-ter, all NTE reports reviewed allege anticompetitivepractices and restriction of distribution channels inthe Japanese steel sector, but the clear focus of thoseallegations is on carbon steel. The distinctive stain-less steel industry is not specifically mentioned.

45. ISSI’s profit rate was derived from theCommerce Department’s final computer printout,February 19, 1998. This document was made avail-able by ISSI’s counsel at the law firm of White &Case. Since Commerce calculates profit as a per-centage of cost of production, it was necessary todivide the Commerce figure by one plus theCommerce profit rate to arrive at profit as a per-centage of sales. The profit rate for the U.S. electri-cal and electronics products industry was takenfrom Bureau of the Census, “Quarterly FinancialReport for Manufacturing, Mining, and TradeCorporations,” Fourth Quarter 1997, Table B,available at www.census.gov/ prod/ www/abs / qfr-mm.html.

46. ISSI’s Taiwan and U.S. sales figures were takenfrom its August 6, 1997, supplemental response tothe Commerce Department’s antidumping ques-tionnaire. This document was made available byISSI’s counsel at the law firm of White & Case.

47. Michael Porter, Competitive Strategy: Techniquesfor Analyzing Industries and Competitors (New York:Free Press, 1980), p. 9.

48. The U.S. antidumping law does attempt tomatch U.S. and comparison-market sales made atthe same “level of trade” and make price adjust-ments when such matching is impossible. The vari-ations in competitive intensity referred to here,however, can occur within the same level of trade asdefined by Commerce and thus elude any adjust-ment.

49. An estimate now over a decade old found thatgray-market imports total $10 billion a year. S.Tamer Cavusgil and Ed Sikora, “HowMultinationals Can Counter Gray MarketImports,” Columbia Journal of World Business 23, no.

29

4 (Winter 1988): 76.

50. Lawrence M. Friedman, “Business and LegalStrategies for Combating Grey-Market Imports,”International Lawyer 32, no. 1 (Spring 1998): 28.

51. U.S. Bureau of the Census, Statistical Abstract ofthe United States: 1998, Table 863, p. 544.

52. Financial information taken from companies’10-K reports, available at www.sec.gov.

53. Financial information taken from company’s1998 10-K report, available at www.sec.gov.

54. Kara Swisher, “At Amazon, the CFO SellsInvestors on the Merits of Losses,” Wall StreetJournal, March 25, 1999.

55. See, for example, Pankaj Ghemawat, “BuildingStrategy on the Experience Curve,” HarvardBusiness Review 85, no. 2 (March-April 1985): 143-49.

56. Michael Porter, Competitive Advantage: Creatingand Sustaining Superior Performance (New York:Free Press, 1985), pp. 436-37. Emphasis in original.

57. Charles T. Horngren, George Foster, andSrikant M. Datar, Cost Accounting: A ManagerialEmphasis, 9th ed. (Upper Saddle River, N.J.:Prentice Hall, 1997), p. 555.

58. See, for example, John C. Panzar and Robert D.Willig, “Economies of Scope,” American EconomicReview 71, no. 2 (May 1981): 268. Panzar andWillig coined the term “economies of scope.”

59. David Besanko, David Dranove, and MarkShanley, The Economics of Strategy (New York: JohnWiley & Sons, 1996), p. 178.

60. Ibid., p. 184.

61. Specifically, profit, selling expenses, general andadministrative expenses, and interest expenses werestripped out of the calculation of normal value.What remained was total manufacturing costs—raw materials, direct labor, and factory overhead.

This remainder still overstates variable costs, since itincludes fixed overhead costs (for these respondentsit was not possible to separate out fixed and variableoverhead). The recalculations were performed bythe companies’ counsel at the law firm of White &Case. The revised dumping margin calculation pro-grams were reviewed by the author of this study.

62. Section 773(b)(2)(B) of the Tariff Act of 1930, asamended, codified at 19 U.S.C. § 1677b(b)(2)(B).

63. Section 773(f )(1)(C) of the Tariff Act of 1930,as amended, codified at 19 U.S.C. § 1677b(f )(1)(C).

64. Some antidumping supporters argue that cross-subsidization is itself an unfair trading practice.Thus, Terence Stewart, a prominent attorney whorepresents complaining U.S. industries inantidumping cases, has argued that the antidump-ing law is designed to “offset any artificial advantagethat flows from closed foreign markets, cross-subsi-dization by multiproduct firms, government largesse,or other factors that have nothing to do with com-parative advantage.” Terence P. Stewart,“Administration of the Antidumping Law: ADifferent Perspective,” in Boltuck and Litan, p. 288(emphasis added). This position, however, is unten-able. Cross-subsidization is endemic among multi-product firms, both foreign and American; indeed,the potential for cross-subsidization is one of themain reasons that multiproduct firms exist. It makesno sense to condemn cross-subsidization by foreigncompanies as “unfair” when identical business prac-tices are routinely pursued by their American rivals.

65. Admittedly, the scope of interventionist policiesthat the antidumping law claims to address isbroader than that covered under the CVD law: first,the CVD law does not apply to NME countries;second, it does not purport to address sanctuary-market situations or broad structural distortions likeinsufficiently developed commercial law.Nevertheless, for constructed-value cases—whichare limited to market economies and (as seen above)generally occur in situations in which sanctuarymarkets are highly unlikely—there is a significantoverlap in the ostensible targets of CVD andantidumping investigations.

30

66. There are limits on obtaining double reliefthrough simultaneous antidumping and CVD peti-tions. Specifically, the CVD law distinguishesbetween export subsidies (subsidies tied to exports)and domestic subsidies (subsidies targeted to specif-ic industries). For export subsidies, the antidumpinglaw provides an offset for any CVD duties paid; fordomestic subsidies, though, there is no offset, andtherefore it is possible that simultaneous antidump-ing and CVD actions could double count the mar-ket-distorting effects of such subsidies.

67. Those four investigations are pasta from Italy,carbon steel wire rod from Canada, carbon steelwire rod from Trinidad and Tobago, and stainlesssteel wire rod from Italy. In each of those casesCommerce conducted contemporaneous anti-dumping and CVD investigations. In addition,another constructed value-based antidumpinginvestigation included in the sample—stainless steelwire rod from Spain—covers a product that is sub-ject to an outstanding CVD order. That CVD casehas been inactive, however, since an administrativereview determination in 1990 found de minimis sub-sidies. Furthermore, three other CVD investiga-tions overlapped with antidumping investigationsincluded in the sample reviewed in this study: pastafrom Turkey, carbon steel wire rod from Germany,and carbon steel wire rod from Venezuela. In thosethree cases, however, the Commerce Department’sdeterminations were based on “facts available,” notconstructed value.

68. Indeed, in 1 of those 22 investigations, freshAtlantic salmon from Chile, Commerce did con-duct a CVD investigation but made a negative find-ing. Yet for three of the five respondents in theantidumping case, Commerce made affirmativedeterminations on the basis of comparing U.S. salesto constructed value (the other two respondentsreceived negative determinations). It is hard tosquare an affirmative constructed-value-baseddumping finding—which supposedly points tomarket-distorting government subsidies—and anegative subsidy determination in a correspondingCVD case.

Board of Advisers

James K. GlassmanAmerican EnterpriseInstitute

Douglas A. IrwinDartmouth College

Lawrence KudlowAmerican Skandia LifeAssurance, Inc.

William H. Lash IIIGeorge Mason UniversitySchool of Law

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 are $6 each ($3 for five or more). To order, contact the CatoInstitute, 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

“Trade and the Transformation of China: The Case for Normal Trade Relations” by DanielT. Griswold, Ned Graham, Robert Kapp, and Nicholas Lardy, Trade Briefing Paper no. 5( July 19, 1999)

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

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

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

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

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

“America’s Maligned and Misunderstood Trade Deficit” by Daniel T. Griswold, TradePolicy Analysis no. 2 (April 24, 1998)

Freedom to Trade: Refuting the New Protectionism, edited by Edward L. Hudgins ( July 1997)

CENTER FOR TRADE POLICY STUDIES