Shell Games and Fortune Tellers: The Sun Doesn’t Set at the Antidumping Circus

  • Downloads


Much has been written about the failure of the U.S. antidumpinglaw to live up to its rationale, even as articulated by the law'ssupporters.1Likewise, many of the methodological quirks and institutionalbiases inherent in the law's administration have been documentedand described in previous studies.2

Despite antidumping's egregious flaws, the law garners broadbipartisan support in the U.S. Congress. But only the most cynicalof policymakers could continue to support the application ofantidumping measures when their intended purpose -- to remedyinjurious dumping -- ceases to exist.

Under the World Trade Organization's Antidumping Agreement andunder U.S. law, a "sunset review" (also known as an "expiry review"or "five-year review") to determine whether an antidumping measureis still needed to redress injurious dumping must be initiatedwithin five years of the measure's originalimposition.3 But arule that was intended to prevent indefinite application ofsuperfluous import restraints has proven ineffective. During sevenyears of rendering so-called sunset reviews, the U.S. antidumpingauthorities have revoked fewer than one in every four measureswhere a domestic interest favored continuation.

In April, the U.S. International Trade Commission voted by amargin of four to two against revoking antidumping (andcountervailing duty) restraints on imports of hot-rolled steel fromBrazil, Japan, and Russia, despite compelling evidence that thedomestic industry is thriving according to every conceivablemeasure of business performance. The facts of that case areaccessible enough to illuminate some of the problems with thesunset process. It is apparent that reform of the process will berequired to rein in the abuse that interminable antidumpingrestrictions represent.

Now You See the Sunset...

Before the conclusion of the Uruguay Round of trade negotiations(1986-1994), which produced the World Trade Organization and, amongother agreements, the Antidumping Agreement (ADA), some countrieswith antidumping laws had no formal procedures for terminatingrestraints imposed pursuant to those laws. There were norequirements that existing measures be reviewed to determinewhether they were still warranted.

The problem was particularly acute in the United States. At theend of 1994, 32 U.S. antidumping measures originally imposed before1980 were still in effect without any intervening assessments ofthe conditions of the domestic industries involved.4 What were originally justified astemporary measures to counter the injurious effects of dumpingmorphed into permanent fixtures that distorted trade and subsidizedprotected producers.

The ADA set out to fix this problem. Article 11.1 of the ADAstipulates: "An anti-dumping duty shall remain in force only aslong as and to the extent necessary to counteract dumping which iscausing injury."5To ensure that national authorities honor this commitment, Article11.3 of the ADA further stipulates: "any definitive anti-dumpingduty shall be terminated on a date not later than five years fromits imposition . . . unless the authorities determine . . . thatthe expiry of the duty would be likely to lead to continuation orrecurrence of dumping and injury."6

...Now You Don't

While the sunset provisions of the ADA were hailed as aliberalizing achievement, the fact is that they have done little toensure termination of superfluous antidumping measures in theUnited States. The U.S. track record of revocation is especiallybleak.

Between July 1998 (the month of the first U.S. sunset reviewdetermination) and May 2005, 335 antidumping sunset reviewdeterminations were issued. In 80 of those 335 cases, the measureswere revoked because there was no domestic interest in theircontinuation.7

In the remaining 255 cases, where a domestic industrydemonstrated interest in preserving the antidumping measures, theDepartment of Commerce found that revocation would be likely tolead to a continuation or recurrence of dumping in everycase. And in 194 of those 255 cases, the ITC found thatrevocation would be likely to lead to a continuation or recurrenceof material injury.8 Thus, based on the accumulated record to date, thelikelihood of revocation pursuant to the U.S. sunset provisions isnonexistent at the DOC and only 23 percent at the ITC. Thesestatistics seem to contravene the intended reform negotiated duringthe Uruguay Round.

To put those numbers in perspective, consider the revocationstatistics in other large, antidumping jurisdictions. Over acomparable time frame -- cases initiated between 1998 and 2004 --the Canadian International Trade Tribunal has revoked 40 of 85antidumping orders pursuant to expiry reviews, which is arevocation rate of 47 percent, more than twice as high as that inthe United States. 9 Likewise, during the same period the European Unionrevoked 36 of 82 cases subject to expiry reviews, a revocation rateof 44 percent.10

Although these comparative data show that U.S. antidumpingauthorities are much less likely to revoke an antidumping measurethan their Canadian and European counterparts, they do not help toexplain why that is the case. A brief review of the laws,regulations, and other administrative procedures surrounding U.S.sunset reviews might help with that question.

Shell Games and Fortune Tellers

It does not require a great deal of investigative prowess tocomprehend why the DOC always finds that revocation is likely tolead to a continuation or recurrence of dumping. According to itsestablished procedures for conducting sunset reviews, the DOCnormally will determine that dumping is likely to continue or recurif

  1. dumping continued at any level above de minimis[0.5 percent] after the issuance of the order or the suspensionagreement, as applicable;
  2. imports of the subject merchandise ceased afterissuance of the order or the suspension agreement, as applicable;or
  3. dumping was eliminated after the issuance of theorder or the suspension agreement, as applicable, and importvolumes for the subject merchandise declinedsignificantly.11

Meanwhile the conditions that would normally lead to adetermination of no likelihood of continuation or recurrence ofdumping are limited to the case where "dumping was eliminated afterthe issuance of the order or the suspension agreement, asapplicable, and import volumes remained steady or increased . . .In analyzing whether import volumes remained steady or increased,the Department normally will consider companies' relative marketshare."12

Thus, the only conditions that normally would warrant a "nolikelihood" finding are conditions that are extremely unlikelygiven rational economic behavior. The foreign exporter would haveto increase his U.S. price -- so that it is no longer dumpingaccording to the DOC's slanted methodologies -- and then findcustomers willing to purchase the same or greater volume than itsold previously at lower prices. This chain of events plainlyviolates established properties of a normal demand curve.

Many of the conditions necessary for a "no likelihood" findingare simply beyond the control of the foreign exporter. Given theprospective nature of U.S. antidumping duty liabilityassessment,13importers are the entities that assume the risk when purchasingfrom a foreign producer that is subject to an antidumping order.Importers can never be sure of their ultimate financial obligationsto the U.S. Customs Service. Thus, the decisions of risk-averseimporters to forego purchases from subject foreign producers areultimately what determine those producers' sales volumes in theUnited States.

In many cases, foreign producers have a difficult timeconvincing importers to continue purchasing precisely because ofthe uncertainty presented by the antidumping measure. Expecting theimporter to discount the risk and continue purchasing, but at ahigher price, is akin to a jobless person with no assets or credithistory asking for a prime rate loan. Success at either is highlydubious.

Furthermore, conditions that normally lead to a "likelihood"determination can be false positives. The cessation or decline ofimports after the issuance of an order can reflect many factors,such as difficulty finding U.S. importers willing to continuepurchasing, the relative attractiveness of other markets, orprohibitively high antidumping rates.

Accordingly, the bright line test specified in the policybulletin is necessarily distorted by decisions beyond the controlof the producer subject to antidumping restraints and cannot beconsidered a useful barometer of what might happen if therestraints were revoked. It is a veritable shell game.

Ultimately though, a sunset review can end in revocationregardless of the DOC's determination. In fact, 23 percent of allcontested sunset reviews ended in revocation despite the DOC'sfindings.

Is it a reasonable expectation that 77 percent of domesticindustries protected by antidumping measures likely would beinjured within a foreseeable time if those measures were revoked?The more likely explanation for the paucity of revocations lies inthe fact that the ITC is obligated statutorily to gaze into acrystal ball and project what might transpire should the measuresbe revoked. Prospective assessments are problematic enough --financial experts lose money in the stock market every day -- butto mandate that officials speculate, opens the door unnecessarilyto subjective interpretations preconceptions, and politicalinfluence.

The recent decision in the sunset review of hot-rolled steelfrom Brazil, Japan, and Russia provides some rich insights into theproblems created by this prospective statutory construct.

The Greatest Hot-Rolled Steel Case on Earth

The hot-rolled steel case is instructive because it features asplit vote -- four to two in favor of continuation -- and thusavails differing opinions from the same factual record. It alsotells a story of just how difficult it is to get an antidumpingorder revoked.

The original measures were imposed in 1999, when U.S. steelproduction was relatively fragmented; many producers were havingdifficulty remaining viable; bloated labor contracts, pension andhealthcare obligations, and other inefficiencies were dragging downprofitability; and a flood of imports had been diverted to thestrong U.S. market because Asian steel demand had nearly completelycollapsed in the wake of its widespread financial crisis in 1997and 1998.

Today the situation is dramatically different. The U.S. steelindustry has undergone structural consolidation, sheddinginefficient, uncompetitive mills; unloading most of its legacycosts on the taxpayer-backed Pension Benefit Guarantee Corporation;negotiating new flexible labor contracts; consolidating operations;and achieving a degree of market power that has enabled it todictate terms to its customers. The year 2004 was characterized byrecord industry profits, persistent steel shortages, perpetualprice hikes, raw materials surcharges, and occasional brokencontracts by steel producers who reckoned they could get higherprices on the spot market by reneging on their commitments.Meanwhile, worldwide demand has experienced a massive outward shiftover the past few years, primarily because of the growing appetitefor steel in underdeveloped countries like China, India, and thoseof Eastern Europe. A strong global steel demand for many years tocome is almost certain.

The majority and dissenting opinions of the ITC in the sunsetreview do not diverge over this now-and-then characterization ofthe hot-rolled steel industry. It is supported by ample evidence onthe record and acknowledged throughout the determination. But thereis a considerable divergence of opinion with respect to what thesedata mean. For example, the majority opinion found that the"industry's greatly improved performance in 2004 . . . resultedmainly from highly favorable market conditions, includingunprecedented increases in global demand, tight global supply, andsharply rising prices."14 It went on to conclude, "In short, the factorsthat enabled the domestic industry to achieve strong operatingresults in 2004 are not likely to continue. Thus, while we do notconsider the domestic industry to be currently in a weakenedcondition, it is susceptible to the continuation or recurrence ofmaterial injury."15

By contrast, the dissenting opinion considers the industrystronger and fundamentally changed. "The depth and breadth of thesechanges all indicate that the condition of the domestic industry ismuch changed, and much improved, from the period of the originalinvestigations. The industry's excellent performance in 2004further supports this conclusion. It is unlikely, then, thatrevocation would affect the industry in the same way and to thesame extent that subject imports affected the domestic industryduring the original investigations.16

The majority opinion reflects the view that 2004 wasserendipitous; that the structural changes in U.S. and world supplyand demand will have no or little bearing on the long-term healthof the industry; and that if restrictions were lifted, imports fromthe subject countries, attracted by relatively higher U.S. prices,would flood back into the U.S. market in high volumes, suppressingor depressing prices, and negatively affecting the domesticindustry. The working assumption in the majority opinion, anargument put forth by the domestic industry, is that increasedChinese demand, which drove the conditions of the world market in2004, would be satisfied by Chinese production going forward. As aresult, the world's exporters -- including the subject producers --would again look to the United States to absorb their collectiveoutput.

Pointing to some excess capacity in the subject countries, themajority concludes that "producers in the subject countries havethe ability and incentive to send significant quantities ofhot-rolled steel to the United States."17 Furthermore, the majority finds that"underselling would likely be significant in the event ofrevocation/termination given subject producers' pricing behaviorduring the original investigations, the importance of price, thesubstitutability of the products, and the fact that increasedvolumes for this product would likely be achieved through lowerprices."18 Butthe dissenting opinion sees these data differently -- and quitefrankly, more objectively.

"Although China continues to build new steel capacity, demand isprojected to increase by 8 percent in 2005, in part because ofpreparations for the 2008 Olympics and the Expo 2010. The recordalso indicates that there is increasing price parity between theU.S. prices and other major world markets. In short, the U.S. isnot the safe haven for steel that it was during the originalinvestigation.

"Overall, the home market demand and export trends in thecumulated countries have changed considerably since the originalinvestigations. While there are existing inventories and in somecases, available capacity, given the worldwide changes in demand,we cannot conclude that it is more likely than not thatimports will increase to such an extent as to cause material injuryto the domestic hot-rolled industry."19

Exiting the Hall of Mirrors

The facts of the hot-rolled steel case underscore an inherentshortcoming of the sunset process: expecting objectivity from theDOC and forcing predisposed ITC Commissioners into the rolls ofsoothsayers make for bad policy. It is difficult enough for theITC, on a retrospective basis, when a factual record exists, toaccurately attribute domestic material injury to imports in thepresence of other causal factors. To require the ITC to speculatewhat those data might be and then speculate the effects of thosespeculated data on the domestic industry is unworthy of theformality of the process. There can be little doubt that under thecurrent system, antidumping measures that should be revoked arecontinued routinely because of vastly differing interpretations ofthe statutory criteria and because of the inherent fallibility ofdecisions that are based on speculation.

One sensible, even-handed way around the current morass would beto combine automatic termination of antidumping measures withexpedited provisions to reinstate them should conditions ofinjurious dumping recur within a reasonable time frame -- say oneyear. Under this proposal, all antidumping measures would beterminated after five years without exception or deliberation.However, domestic industries would be allowed to file newpetitions, as early as the date of revocation, which would requireevidence of material injury or threat of material injury by reasonof less than fair value imports, just as is required in originalpetitions.

For petitions filed within one year of the previous order'sexpiration, special provisions designed to expedite the processwould prevail. An affirmative preliminary injury finding renderedby the ITC within 45 days in accordance with its normal criteriawould suffice to reinstate the dumping order with rates prevailingat the time of revocation. In other words, a DOC investigation,which normally takes six to twelve months to produce a preliminarydumping rate, would be unnecessary for reinstituting the dumpingorder. Under this proposal, relief could be available to domesticindustries found to be injured within 45 days of revocation of theoriginal order. After the duties were reimposed, a traditional DOCinvestigation would take place and would entail a recalculation ofthe duty rates. This idea was originally proposed in a Cato paperpublished in 2002.20


The sunset review process in the United States is broken. As aresult, antidumping measures that distort trade and impose realcosts on interests throughout the economy are allowed to continuewell beyond their legal justification. It is imperative that U.S.policymakers fix this problem. Automatically terminatingantidumping measures after five years with provisions for expeditedre-imposition if the legal thresholds are met could go a long waytoward improving the system.

1. See Brink Lindsey,"The U.S. Antidumping Law: Rhetoric versusReality," Cato Trade Policy Analysis no. 7, August 16,1999.

2. See Brink Lindsey andDan Ikenson, "Antidumping 101: The DevilishDetails of 'Unfair Trade' Law," Cato Trade Policy Analysis no.20, November 26, 2002.

3. Actually, there is adifference between the WTO language and the U.S. law on this point.The WTO agreement reads that an antidumping measure should beterminated within five years of imposition unless the authoritiesfind that revocation would likely to lead to a continuation orrecurrence of dumping and injury. But the U.S. law mandates that anotice of initiation of an expiry review be published no later than30 days before the fifth anniversary of the antidumping measure.Accordingly, most U.S. decisions to continue or revoke an order arenot rendered until the sixth year, which would seem to beinconsistent with U.S. WTO obligations.

4. Michael O. Moore,"Commerce Department Antidumping Sunset Reviews: A MajorDisappointment," George Washington University Occasional PaperSeries, October 7, 2002, p. 1,

5. World TradeOrganization, Agreement on Implementation of Article VI of theGeneral Agreement on Tariffs and Trade 1994 (the "AntidumpingAgreement"), Article 11.1,

6. Ibid. Article 11.3.

7. Data compiled from twosources: Import Injury Investigation Case Statistics (FY1980-2003), U.S. International Trade Commission, Office ofInvestigations, November 2004,,and other sunset review decisions and data available at

8. Ibid.

9. Figures compiled fromexpiry review decisions and data available on the website ofCanadian International Trade Tribunal at

10. Figures compiledfrom antidumping expiry review data appearing in annual reports ofthe European Commission to the European Parliament titled:"Anti-dumping, Anti-subsidy, Safeguard Statistics Covering the Year[1998, 1999, 2000, 2001, 2002, 2003, 2004],"Annex F (for 2001through 2004 data) and Annex G (for 1998 through 2000 data).Reports are available on the website of the European Union,

11. Department ofCommerce, International Trade Administration, "Policies Regardingthe Conduct of Five-year ("Sunset") Reviews of Antidumping andCountervailing Duty Orders; Policy Bulletin," Federal Register,April 16, 1998, p. 18873. The "Policy Bulletin" has been thesubject of WTO dispute settlement in a few cases. Most recently,the Appellate Body upheld a dispute panel's finding that thebulletin is a "measure" subject to WTO dispute settlement, butreversed the panel's finding that Section II.A.3 of the bulletin isinconsistent with Article 11.3 of the ADA due to what is considered"insufficient analysis" by the panel. (See World TradeOrganization, Appellate Body Report, United States-Sunset Reviewsof Anti- Dumping Measures on Oil Country Tubular Goods fromArgentina, November 29, 2004.)

12. Ibid.

13. Under the U.S.system, antidumping duties are estimates of duty liability paid bythe importer of record to U.S. customs authorities. Those estimatesare based on calculations of dumping from entries during a previousperiod. Only after the merchandise has cleared customs, normally aspart of an annual "administrative review," will the actualliability be calculated. So, if the antidumping deposit rate was 10percent at the time of entry, but the actual liability wascalculated at 20 percent, the importer would receive a bill (likelymonths or even years after the entry) for the difference between 20and 10 percent, plus interest. Thus, there is a lot of uncertainty(i.e., business risk) in continuing to do business with foreignexporters that are subject to antidumping measures.

14. U.S. InternationalTrade Commission, Certain Hot-Rolled Flat-Rolled Carbon-QualitySteel Products from Brazil, Japan, and Russia: Investigations Nos.701-TA-384 and 731-TA-806- 808 (Review), April 2005, p. 28.

15. Ibid., p. 41.

16. Ibid., p. 43.

17. Ibid., p. 36.

18. Ibid., p. 38.

19. Ibid., p. 58(emphasis added) .

20. See Brink Lindseyand Dan Ikenson, "Reforming the AntidumpingAgreement: A Road Map for WTO Negotiations," Cato Trade PolicyAnalysis no. 21, December 11, 2002.