Policy, Research, and Extemal Affairs | ) f o 3 9 7 WOR,ING PAPERS L International Trade International Economics Department The World Bank April 1990 WPS 397 Antidumping Regulations or Procartel Law? The EC Chemical Cases Patrick A. Messerlin Antidumping regulations allow domestic firms to participate in V. cartels to capture domestic markets. , The Policy, Research, and External Affairs Complex distributes PRE Working Papers to disseminate the findings of wotk in progress and to encourage the exchange of ideas among Bank staffand all others intermsted in development issues. These papess carry the names of the authos, reflect only their views, and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own. They should not be attributed to the World Bank. its Board of Directon, its tnanagement, or any of its member countries. Plc,Research, and External Affairs International Trade This paper- a productof the Intemational Trade Division, Intemational Economics Department- is part of a larger effort in PRE to assist developed and developing countries evaluate the impact of protectionist instruments. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Salome Torrijos, room S8-033, extension 33709 (51 pages with figures and tables). The paper considers the intricate relations anticartel fines they had to pay. So cartels are between antidumping actions, cartelization, and likely to continue lodging antidumping actions, anticartel actions. even at the cost of antitrust fines. In 1982 and 1983, antidumping actions for Cartels can capture existing antidumping two products - polyethylene and polyvinyl procedures so easily that they make disciplines chloride - were terminated by measures that on the antidumping actions a mockery. One- doubled the protection granted to the European fourth of the antidumping cases initiated by the Community's chemical industry. Six years later, European Community involve the same products anticartel investigations revealed two active and firms as one-fourth of the EC anticartel cartels in thcse products in 1983-85. cases. Messerlin argues several points in this The relief from foreign competition that paper: antidumping laws are supposed to provide is likely to entrench domestic cartels, slow the i Antidumping actions made it possible for adjustment of the cartel's members, and generate these cartels to exist. a powerful demand for more protection in the future. * These cartels have imposed huge extra costs on EC users - roughly DM 600 million, for Messerlin concludes that the EC antidump- markets with annual sales of DM 6 billion. ing regulations should be fully subordinated to the anticartel laws embodied in Articles 85 and * Gains the EC firms got from the antidump- 86 of the Treaty of Rome. ing protection were ten times larger than the ! The PRE Working Paper Series disseminates the findings of work under way in the Bank's Policy, Research, and External Affairs Complex. An objective of the series is to get these findings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. -- _ _ _ _ _ _ _-- - _ _- __-Po.- _-c- at t _he-PRE_Dsseminaton_Ccnte Produced at the PRE Dissemination Center Foreword Using the rich information avai2able in the European Community, this paper shows that antidumping regulations are a device that domestic firms -- which would not be able to collude if there were no protection -- can easily capture for cartelizing domestic markets. Two corollaries follow. Protection under antidumping regulations is particularly costly for the domestic consumers since it is often accompanied by cartelization. The relief from foreign competition that antidumping laws are supposed to provide is likely to entrench the domestic cartels, to slow the adjustment of the cartel's members, and to generate a powerful demand for more protection in the future. Antidumping Regulations or Procartel Law? The EC Chemical Cases* by Patrick A. Messerlin** Table of Contents Introduction 1 Section 1. The aftermath of antidumping measures: the 2 anticartel cases Section 2. Two major cartels 8 Section 3. The capture of the antidumping procedures by 19 the cartels Section 4. The anticartel actions: no deterrent effect 39 Conclusion 47 References 49 Annex A 50 * I would Like to thank B. Balassa, D. Henderson, B. Hoekman, D. Tarr, and J. Temple Lang for their comments on earlier drafts. ** The World Bank, S8-035 1818 H Street, N.W. Washington, D.C. 20433 Phone: (202) 473-3834 Fax: (202) 477-0661 Introduction Do antidumping actions under the European Community (EC) regulations lead to decisions antithetical to the competition rules embodied in Articles 85 and 86, the two corner stones of the Treaty of Rome? Existing analysis of extra-EC and intra-EC trade flows suggests that the vast major'ty of antidumping measures has led to price agreements prejudicial to the EC consumers' interests [Messerlin, 1989]. However, there is a need to go beyond such studies based on aggregate import data: answering the question raised requires some detailed knowledge of market structures and firm behavior. This paper addresses the issue by doing a very simple task. It connects the official information given by specific antidumping cases issued by the Commission's antidumping office (Directorate General I) and the information provided by the "twin" --i.e.. involving the same product and firms-- anticartel cases issued by the Commission's competition office (Directorate General IV). It focuses on chemical cases because the chemical industry has been the most intensive user of the EC antidumping regulations, representing roughly 40% of all the cases initiated (about 46% of those terminated with a restrictive outcome) between 1980 and 1987. Connecting these two sources of official information gives three main results. First, cartels have used antidumping cases as de facto substitutes to exemptions from competition rules in markets difficult to be cartelized because of foreign competition. Second, cartels monitoring foreign markets as well as the domestic market have easily captured the existing antidumping procedures. Last, the costs of antidumping actions for the EC consumers have been high because antidumping protection has been high and has allowed the cartels to survive. 2 These results raise the crucial problem of "the reconciliation of anticartel and antidumping laws" [Temple Lang, 1989].1/ First, is it sound to have two procedures of exemption from competition, one embodied in Article 85:3 of the Treaty of Rome and another one --difficult to control-- indirectly provided by antidumping regulations? Second, is it compatible with the "general interest" to have antidumping procedures which are both costly in terms of welfare and so easily captured by the import-competing firms? Last, is it acceptable that domestic firms gain so much from antidumping actions that anticartel actions are unlikely to be an efficient deterrent against lodging new antidumping cases? This paper is organized as follows. Section 1 surveys all the anticartel cases preceded or followed by related antidumping cases --the twin cases-- existing in the recent EC experience. Section 2 focuses on the major twin cases which have occurred in two predominant petrochemical products. Section 3 assesses the extent of the capture of the antidumping procedures by the cartels in these two main twin cases. Section 4 estimates the costs of the antidumping protection and the deterrent power of the anticartel actions. Last, the conclusion suggests the principle of subordination of antidumping regulations to competition rules. Section 1. The aftermath of antidumping measures: the anticartel cases Economic analysis suggests that antidumping regulations should have the same content and outcome as anticartel laws because they both deal with price discrimination based on market segmentation and monopolistic behavior. However, casual observation shows this is not the case. 1/ Van Bael and Bellis [19871 note "the conflicting aims pursued by antitrust and antidumping policies. Thus far, no antitrust case has arisen concerning the "negotiation" of a price revision undertaking" [page 103]. 3 The EC anticartel laws and antidumping regulations are nurtured by two different legal environments. Articles 85 and 86 open the part of the Treaty devoted to the "common rules." They are thus restricted to anti-competitive behavior which may affect trade between Member States.2/ Article 85 deals with concerted actions (price or quantity agreements) between firms. Article 86 prohibits abuses of a dominant position. Antidumping regulations refer to Article 113 (pertaining to the part of the Treaty devoted to "economic policy"), to GATT Article VI and the Antidumping Code, which are both only concerned with the injury done to tne domestic industry by dumped imports. Anticartel and antidumping outcomes are not similar either. There are few anticartel cases on price discrimination or predatory pricing, and few of them are terminated by judgments favorable to plaintiffs [Salop & White, 1986]. By contrast, there are many antidumping cases (more than 300 in the EC since 1980) involving price discrimination, and most of them (70%) have ended with protective measures. The number of 'twin" cases: substantial Nearly thirty EC anticartel cases initiated since 1980 deal with products also involved in almost a hundred antidumping cases.3/ These figures represent 23% of all the anticartel cases after 1980 and 25% of all the antidumping cases 2/ For a detailed analysis of the competition law of the Community, see van Bael and Bellis [19871. The "common rules" include Article 91 on dumping practices in the intra-EC trade. 3/ This figure excludes several other antitrust cases which have more tenuous links with antidumping cases, such as a case on intellectual property rights in production of lamps. In what follows, antitrust cases are defined as proceedings under Articles 85 and 86 of the Treaty of Rome and under Regulation 17/62. Steel products under the ECSC rules of the Treaty of Paris are excluded because they are covered by an official crisis cartel [van Bael, 1979]. Antidumping actions concetning the "ECSC steel products" are to be seen as an integral part of the monitoring of the steel cartel. 4 initiated since 1980. Table capitulates che twin cases classified in four parts [(i) to (iv)] according to the Article involved in Xne anticartel case. Insert Table 1 Table 1:A shows there were many more antidumping cases associated with anticartel cases under Article 85:1 --which hans all agreements "which have as their object or effect the prevention, restriction or distortion of competition within the common market"-- than cases under Article 86 --the adequate provision for price discrimination and predatory pricing because it deals with dominant position and its abuse.4/ In other words, dumping appears onlv loosely associated with Rredatory intent, contrary to the alleged goal of the antidumping regulations and the claims of petitioners. Table 1:A provides three more insights. Many anticartel cases initiated under Article 85:1 --52% of all the cases initiated under this Article since 1980-- have twin antidumping cases. The number of the twin antidumping cases involved is substantial: 12.5% of all the antidumping cases initiated between 1970 and 1987. Non-market economies are the most frequently involved countries with 27 cases, i.e., 18% of all the antidtLmping cases against their exporters. Table 1:B confirms the strong correlation between antidumping cases and anticartel cases dealing with agreements between firms. Inhibiting parallel) imports witp'-. the Community (re-export bans) through marketing agreements is covered by Article 85:1. Article 85:3 defines the conditions for exemption from 4/ The Court of Justice's Advocate General has expressed some disagreements on the existence of the predatory pricing in one of these two cases (peroxide) (Financial Times, May 19, 1989). 5 Table U. IC antitrust cases (1980-1989) and related antidumping cases Provisional results Year of Antidusping cases Ad-valorem Products the Antidusping equivalents antitrust year ICs NICs Phis Total aeasures of the decisions [a] [a] [a] measures [b] (i) Cases under Article 85:1 glass, cast 1980 1983 6 6 undertaking (g] 3.7 plyuood 1980 1978 1 1 no dumping 0.0 soda ash 1981 1978 4 4 duty & undertaking glass, float 1981 1984 1985 2 4 6 undertaking [g] 21.2 steel, cast iron [c] 1983 1980 1 1 undertaking 9.6 1985 1 3 4 no dumping 0.0 peroxide, hydrogen 1984 alusinium 1984 1982 1 1 2 4 no dumping 0.0 1983 1 3 1 .5 no dumping 0.0 vood pulp 1984 1978 4 4 no injury 0.0 polypropylene 1986 1981 1 1 undertaking (g] 11.9 PVC (d] 1988 1981 4 4 undertaking [1] 13.1 LdPI [(d 1988 1982 4 4 undertaking [g] 13.3 soda ash 1989 1982 1 1 duty & undertaking [g] 16.5 cement 1989 1985 2 2 4 no dusping 0.0 TOTAL 14 9 13 27 49 S [el 51.9 8.5 9.6 18.1 12.6 (ii) Cases under Article 86 tyres, trucks 1981 1979 1 3 4 undertaking peroxide, organic [f] 1983 1983 1 1 duty & undertaking [g] 23.1 TOTAL 2 1 1 3 5 S [el 10.0 0.9 0.7 2.0 1.3 Sources: IC Official Journal. On compilations. ltes: [a] ICs: Industrialized countries; NICs: levly Ind'd couDtries; gls: hon-Harket economies. [b] unweighted averages in percent (see text). [c] steel products fully covered by the ICSC tales are not taken into account. [d] permitted forms of cooperation have been granted to some firms. (el in percent of all similar antitrust cases or antidumping cases of the same group of countries. [f] interim measure confirsed in the 1985 case. (g] undertaking based on the margiu of 'injury.' Ad valorem equivalents of such undertakings are estimated on the basis of one third of the corresponding sargins of injury (see text). 6 Table 1:8. IC antitrust cases (1980-1989) and related antidumping cases Provisional results Year of Antidusping cases Ad-valorem the _ Antidumping equivalents Products antitrust year ICs IICs Ilea Total seasures of the decisions [a] (a] (a] measures (b] (iii) Cases dealing with "re-export bans' plywood 1983 1981 2 2 no dumping 0.0 titanium, all forms 1982 1984 2 2 no injury 0.0 TVs,K7s,VCRs,CDs [cJ 1981 1979 1 1 no dumping C O 1982 1981 1 1 no dumping 0.0 1983 1982 1 1 arrangesent' (e] I984 1987 1 1 2 duty 20.5 1987 1 1 2 duty & undertaking (f] 18.3 TOTAL 6 9 2 0 11 I [d] 20.0 8.5 1.5 0.0 2.8 (iv) Cases under Article 85:3 titanium, alloys 1982 1977 1 1 no du ping 0.0 synthetic fibers 1984 1972 2 2 undertaking (f] 1978 1 1 undertaking [f] 1979 2 3 5 duty & undertaking lf] 1980 1 1 no dumping & duty 11.3 1983 1 1 no dumping & undertaking If] 4.7 1985 3 1 4 no duping & undertaking (fil 15.4 1986 2 2 4 no injury 0.0 1987 1 4 1 6 duty (g] 20.1 1987 4 4 duty (g] 20.5 colour films 1987 1983 1 1 no dumping & undertaking lf] 21.4 photocopiers 1987 1985 1 1 duty 11.5 printers 1987 1987 1 1 duty (g] 21.5 1987 1 1 duty (g] 3.5 TOTAL 5 10 19 4 33 5 (dl 11.6 9.4 14.1 2.7 8.5 All cases (fro (i) to (iv)J MAL 27 29 35 34 98 S Id] 22.5 27.4 25.9 22.8 25.1 Sources: IC Official Journal. Own cospilations. Iotes: (a] ICs: Industrialized countries, NICs; lewly Ind'd countries; HIks: Non-Market economies. [bl unweighted averages in percent (see text). lc] antitrust actions cover the chole range of consumer electronics. [d] in percent of all similar antitrust cases or antidusping cases of the same group of countries. Le] the arrangement is a voluntary export restraint. ' [f] undertaking based on the margin of 'injury.' Ad valorem equivalents of such undertakings are estimated on the basis of one third of the corresponding margins of injury. [gi provisional measures. 7 the prohibitions under Article 85:1 for a limited duration of time.5/ It can be used as a legal basis for "crisis cartels" needed to "rationalize" the EC industry. There is one interesting difference between Table 1:A and Table 1:B. The share of industrialized and newly industrialized countries is domina X the antidumping cases associated with re-export bans and exemption from competition rules. The outcomes of the antidumping actions (rather than the initiations) give similar results. Antidumping actions associated with the enforcement of Article 85:1 are of three types: duties, undertakings and no dumping (or no injury). Undertakings --i.e. , commitments by foreign firms to limit their sales or increase their prices-- are the type of protectionist tools which can be expected in a non-competitive environment. Rates of protection granted by ant_dumping actions are high: around 20% for duties and 12% for undertakings. Low rates for undertakings are likely to be due to the conservative method of estimation. The antidumping office gives little information on the ad valorem equivalents of undertakings which are supposed "to eliminate the injurious impact of the dumped imports." Estimates given in Table 1 are based on the hyr., hesis that the ad valorem equivalents are one third of the dumping marg5 (this figure is drawn from the few cases for which there is some information about the undertakings in the antidumping proceedings). The relative importance of "no dumping" actions may be a surprise. However, such terminations may cover measures fitting well -i.th a cartel situation, i.e., private alignments to the cartel conditions as a 5/ Article 85:2 declares such agreements automatically void, if they do not benefit from the provisions in Article 85:3. 8 consequence of the successful antidumping threat of the cartel. The likelihood of such outcomes is sup,orted by the fact that imports in these cases show a clear freeze [Messerlin, 1989]. Lastly, the twin cases present many configurations of firm coalitions which can be illustrat(i by extreme cases examined in Annex A. Sec=tion 2. Two major cartels This section which describes the history of the cartels, leads to a crucial question: since the antidumping actions were lodged by firms which were members of the cartel, were the antidumping actions decisive for the cartel's survival?6/ The history of the cases In November 1981, the antidumping office of the Commission initiated an antidumping proceeding against four East European producers (Czechoslovakia, East-Germany, Hungary and Romania) of polyvinyl chloride (PVC). In September 1982, the proceeding was terminated by three "undertakings" by which the foreign exporters accused of dumping "offered" to take the appropriate measures "to eliminate the injurious effect of the exports" [L274:17] and by a provisional antidumping duty of 12% on imports originating from Czechoslovakia. As the regular EC tariff on PVC is 12.5%, the antidumping action doubled the protection granted to the EC industry. In January 1983, the antidumping duty on Czechoslovakian exports was lifted and replaced by an undertaking. 6/ In what follows, quotations refer to the three issues of the EC Official Journal: C230 [3.9.1982] and L138 [27.5.1983] for the antidumping proceeding, L74 [17.3.1989] for antitrust. Figures following the issue refer to pages (antidumping iroceedings) or recitals (antitrust proceedings). 9 In September 1982 --two weeks before the publication of the antidumping measures on PVC-- the antidumping office initiated an antidumping action against four East European producers (Czechoslovakia, East Germany, Poland and Soviet Union) of polyethylene. In May 1983, the proceeding was terminated by four undertakings described in terms similar to the PVC case [L138:67]. In late 1983 --only a few months after the antidumping terminations-- the competition office of the Commission began to investigate price collusion and market sharing in the low-density polyethylene (LdPE) market. When beginning its investigations on LdPE, the competition office became aware of possible similar infringements of the competition rules of the Rome Treaty in the PVC market. In March 1988, the competition office of the Commission initiated an official anticartel proceeding --under Article 85 of the Treaty of Rome-- on the constitution of a cartel by the main European producers of LdPE during the period 1976-1985. In December 1988, the Commission terminated the anticartel case by imposing heavy fines on the EC and non-EC firms involved. The PVC anticartel case followed the same track as the LdPE one: in March 1988 after five years of investigations, an anticartel proceeding was initiated and in December 1988, the Commission terminated the anticartel case by imposing heavy fines on the EC and non-EC firms involved.7/ The nature of the cartels The anticartel investigations provide ample evidence that between 1980 and 1985, the main firms selling LdPE and PVC in the EC markets systematically set up market sharing and concerted price practices in regular meetings. 7/ In both cases, antitrust investigations have been made extremely difficult by the concerned firms, particularly in 1987 and after. 10 Tables 2 and 3 present the EC and non-EC firms which were members of the LdPE and PVC cartels. A common characteristic of the two cartels is the large number of firms involved. This characteristic has had two important but opposite implications. Insert Tables 2 and 3 Large cartels. On the one hand, the cartels were large enough to control most of the EC market, both in terms of capacities of production and of consumption. According to the anticartel decision in LdPE, Exxon did not contribute to the collusion, and Shell and BP were only marginally involved. In 1984, Exxon represented 10% of the EC production capacity and the two other firms roughly 15%. As a result, the cartel controlled 75% to 90% of the production capacity. On the consumption side, a large proportion (50% to 55%) was provided by imports. Intra-EC imports (75% of all imports) were under the cartel control, roughly in the same proportion as production capacities. Extra-EC imports mainly came from the rest of Western Europe, i.e., from the non-EC members of the cartel. As a result, the LdPE cartel controlled between 70% to 80% of the EC apparent consumption (production plus imports minus exports). According to the PVC anticartel decision, Shell was the only firm to "operate on the periphery" of the cartel [L74:53]. As a result, the cartel controlled roughly 95% of EC production capacity. On the consumption side, a substantial proportion (roughly 30%) was provided by imports. Almost half of the extra-EC imports (20% of all imports) come from the four countries under antidumping investigation. As a result, the PVC cartel controlled roughly 90% of the EC apparent consumption of PVC. 11 Table 2. The LdPl auticartel case: fins, capacities of production and fines (Provisional results) Capacities of Petitioners Production Anticartel Discriui- rins in the Notes Plut locations ffou/yr fines natioa antiduping (a] ratio fb] cases 1981 1986 .lIcu S (1981] IC firm subject to anticartel fiies Atoche yes Fruce 350 350 3.60 9.7 150.1 SV yes eGemany 593 218 5.50 14.9 135.4 Bayer yes Germany 150 150 2.50 6.8 243.2 BP Chticals Yes Belgiu, Germny, 400 570 0.15 2.0 27.4 France, Italy DSt yes letherlad 430 430 3.30 8.9 112.0 Inichem-Nonte. yes (c] Italy 680 595 6.50 17.6 139.5 Hoechst yes Germany 160 145 1.00 2.7 91.2 ICI yes france 485 100 3.50 9.5 105.3 Orke yes France 570 500 5.00 13.5 128.0 Shell yes France, Gemany, U.. 575 460 0.85 2.3 21.6 Von-IC firm subject to anticartel finse leste Oy [d]fe] lialad, Sweden 165 465 1.00 2.7 88.5 lorek iydro Norway 125 125 0.50 1.4 58.4 Chesie Holding Autria 115 250 0.50 1.4 63.5 UP Spain 232 312 0.10 0.3 6.3 Dow Chemical Netherland, Spain 320 550 2.25 6.1 102.6 onuanto U.l. 50 0.15 0.4 43.8 Firm operating in the IC and not subject to aticartel fines InOn yes Belgium 480 480 huianca Italy 55 S.1.R. Italy 165 Other firm operating in Western Imiupe and not subject to anticartel fines [PSI Portugal 120 SMAI Spain 150 Unifos Sweden 165 Cartel mbers 5400 5220 37.00 100 All firm 5880 5820 Sorces: IC Official Journal (L74), List (1986], Cheical kconowics Handbook. lotes: (a) Plant locations ad capacities in Western lurope only. (bi firm share in the fines divided by firm share in capacities. (el Iniche becae the unique Italian LdPI producer after 1984. (dl itcluding Unifos (Sweden) bought from Union Carbide in 1985. [el Sweden was the reference country in the antidping case. 12 Table 3. The PVC aticartel case: firs, capacities of productioa ad fines (Provisiosal results) Capacities of Petitioners production nticartel Discrimi- rim in the lotes Plant locations Hfons/yr fices natioa autidunping (a] _ ratio (bl cues 1981 1988 h Icu I (l9811 IC firms subject to aticartel fines Atoches yes (cI France 615 485 3.20 13.6 i0O.5 US! yes belgiu, Germy 280 230 1.50 6.4 103.5 DSh yes letherlad 170 180 1.35 5.7 153.4 Iniche-lonte. Ed] Italy 650 700 4.25 18.1 126.3 Hoechst yes (el Gerany 240 240 1.50 6.4 120.? Huels-Veba yes Germany 410 38o 2.20 9.4 103.7 ICI yes If] Gerny, U.K., Svit:1 603 657 2.50 10.6 80.1 SAV-ICN yes Trance 150 160 0.40 1.7 51.5 Shell yes France 175 350 0.85 3.6 93.8 Solvay yes [el Belgiu, Germay 772 768 3.50 14.9 87.6 Frace, Italy Wacker-Hoechst yes Germany 345 350 1.50 6.4 84.0 Non-IC firm subject to uticurtel fines lorek Hydro yes [g] Iorvay, Ui., Sweden 130 300 0.75 3.2 111.5 lirms operating in the IC nd not subject to aticartel fines Aluusisse-Lonsa yes Gerany, Svit:erlud 67 32 BP yes U.K. 230 0 mnon yes Greece 47 47 Id. Generali Italy 0 25 Liquichima Italy 60 0 uianlca Italy 115 0 S.I.R. Italy 140 0 Other firms operating in Western lurope ad not subject to aticartel fines Aiscondel Spain 45 90 CIBIS Portugal 50 68 leuaord (g] Soeden 140 20 honsanto Spain 45 0 leste Oy linland 60 60 Rio Rodano Spain 120 100 Cartel sembers 4540 4800 23.50 100 All firm 5659 5242 Sources: IC Official Journal (L74), List (19861, Chemical Iconomics HUndbook. lotes: [a' Plat locations and capacities in Western hurope only. [b] firs sbare in the fines divided by firm share in capacities. (el including plants from Chloe, PCUI ad Rhone-Poulenc. ldI Inicbes becue the unique Italian PVC producer after 1988. (el including 751 of Solvic plants. (el including BIP Vinyls nd 251 of Solvic plants. (gl Sveden was the reference country in the antiduping cae. 13 But unstable cartels. On the other hand, large cartels tend to be "unstable" or weak, i.e., have great difficulties in both establishing common rules and monitoring them. Anticartel proceedings substantiate this point in both cases. The PVC anticartel decision provides the most interesting information related to both the LdPE and PVC cartels. A document found at ICI compared the LdPE and PVC collusion mechanisms and raised the following question: "Can the LdPE scheme be operated with 2-3 producers outside? PVC has only one outside" [L74:11]. Indeed, the anticartel decision in the LdPE case mentions that "peace formulas were hard to reach." Although there is some evidence that the origin of the LdPE cartel can be traced back to 1976, market sharing and concerted prices practices were not organized before 1981. Unstable cartels look for public support. These two cases are no exception, as best shown by the LdPE case. In 1982, the LdPE firms tried to get from the Commission an exception from the competition rules through a "crisis cartel" policy, arguing from the large "overcapacity" of LdPE production in Western Europe when demand strongly fell and remained stagnant between 1980 and 1984. This attempt failed [European Chemical News, 19 January 1987]. As a result, antidumping actions were lodged and initiated in very specific circumstances. In the PVC case, the antidumping case was initiated the 19th of December 1981, a few months after that "the collusive arrangements originated in a proposal made in August 1980" [L74:7]. In the LdPE case, an agreement on the "quota scheme" was reached by the cartel meeting of 20 January 1982 [L74:19]: the initiation of the antidumping action occurred the 3rd of September 1982. 14 The antidumping- actions: decisive for the survival of the cartels The antidumping actions obviously were part of the cartels' strategy. But were they decisive for the survival of the cartels? The impact of antidumping actions on price evolutions. A first answer to the question is provided by the relationship between the antidumping actions and the evolutions of LdPE and PVC prices, illustrated by Figures I and 2, respectively. The Commission reported the "target" prices and the prices prevailing for both products between 1980 and 1985. Figures 1 and 2 show only the second type of prices because they incorporate the buyer reactions to the "target" prices. Insert Figures 1 and 2 Both Figures 1 and 2 show that prevailing prices strongly reacted to the two steps of the antidumping actions.8/ First, the initiation of the antidumping procedures allowed the ongoing sharp declines of prevailing prices to stop and to stabilize for the period of the antidumping investigations. Second, as soon as the antidumping measures were decided, prices rebounded sharply. Both evolutions cannot be cxplained by changes in demand for LdPE and PVC which were modest between 1982 and late 1983, according to all industry sources. As a result, price evolutions can only be related to changes in foreign and/or domestic supply. The imRact of antidumRing actions on foreign supply. In the LdPE case illustrated by Table 4, quantities exported by the four "dumping" countries began to increase in 1980 for Czechoslovakia, East Germany and Poland, and in 8/ Prevailing prices suggested by another source [European Chemical News] con'irmed the Commission's data. 15 Fiqure 1 Monthl./ LdPE Prices, 1 '980-8.5 f nde:e. 1i80 = 100 140 -_._. (provisional) 1130 -~~~~~~~~~~ -o~~~~~~~. . . .... .... .. ,, ~ .. . ...... . . ... ... BO.01 el 01 81D1l 83.01 84.01 veor, hbnfth EC preuil - opon exfk 0 * O USePO*t F-ig ure 2 . h,lonthlIy PVfC Pric,es , 198 1-85 140 - ~~~~[ndoxo, I 91811-8 =100] | @ @ ~~~(provisional) 120 r 80- Yn [. -lli 81 ~~4 U., 0~~~~~~~~~~~~~~~ Ila - ; ! IOU - 80.01 80 a.1 8.l1 84.01 4 E.op ha,pt O- EC p'Budoi - .bpon ex:et $ US homr 16 1981 for the Soviet Union, i.e., precisely when the cartel was trying to tighten up price and quantity discipline, according to the anticartel decision. The growth of the East European exports may look impressive because of the low initial level of trade. However, the increases in terms of import market shares were modest --four points for all the four exporters between 1979 and 1981-- particularly if one notes that the EC's total imports were stable during these three years. In the PVC case, illustrated by Table 5, similar observations can be made. Insert Tables 4 and 5 The undertakings which terminated the antidumping cases proved to be extremely effective for both products. In the LdPE case, one year after the undertakings in 1984, imports from Czechoslovakia and the USSR were back to their 1978-1979 level. In 1985 and 1986, imports from Poland and from East Germany, respectively, were back to their 1978-1979 level. The PVC case shows similar results. In 1987, imports from Czechoslovakia and East Germany were back to the 1979 level. What was the impact of the additional protection granted by the antidumping measures on the cartels' share in trade? In the PVC case, the intra-EC trade stabilized at around 80% of the total trade in PVC after 1981. In the LdPE case, the situation is slightly more complex because the cartel has a marked non-EC component: all the domestic producers of the five non-EC Western European countries were members of the cartel. As a result, the marked decline of the East European import market share benefited the imports from the non-EC Western European firms --mainly from Spain and Sweden--- which reached a peak in 1983-1984. Their substitution for the East European countries can be seen as the eviction of "mavericks" --obliged to lower their prices in order to 17 Table 4. luropean Couuniy trade in LdPI, 1978-1987 Countries 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 (a] [b] Import uarket sbares, in I of total (intra and extra) IC imports Duma.ng' countries 2.4 2.4 2.8 5.2 6.8 8.2 4.5 2.9 2.7 1.6 2.0 Czekoslovakia 0.1 0.6 0.7 1.9 0.8 0.9 0.6 0.3 0.6 0.5 0.4 last-Gernany 0.3 0.3 0.3 1.4 1.9 2.2 1.3 1.1 0.7 0.4 0.8 Poland 0.0 0.7 0.7 0.7 0.4 0.3 0.0 0.1 Soviet Union 2.0 1.4 1.8 1.3 3.3 4.4 2.2 1.2 1.4 0.7 0.1 Ion-IC West-lurope 7.2 9.5 9.6 11.6 13.7 12.7 14.2 15.7 13.0 9.7 10.8 Austria 1.8 2.6 1.8 3.7 6.0 4.6 6.1 6.0 3.3 0.7 1.3 lorway 0.5 2.4 2.6 2.7 2.6 2.5 2.8 2.4 2.2 1.7 Finland 2.0 1.8 1.4 1.1 1.0 0.9 1.1 1.1 0.8 0.4 0.5 Sweden 3.1 3.4 3.3 2.9 3.0 3.3 3.6 4.5 5.2 5.1 5.6 Spain 0.3 1.2 0.7 1.2 1.0 1.2 0.9 1.3 1.4 1.4 1.8 Hiddle-last countries 0.1 0.1 10.1 15.5 11.6 Intra-IC trade 88.9 86.2 84.7 79.9 75.2 73.0 76.4 75.7 66.8 68.4 70.7 Total IC imports 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 in tons 1379.9 1560.4 1745.9 1746.9 1731.2 1857.8 2065.8 2207.7 1842.5 1938.9 1741.9 litra-IC exports lxtra-KC trade in tons 817.0 842.0 745.7 674.3 633.5 606.6 767.5 800.6 409.2 382.2 407.4 Unit values [c] 1536 1227 1988 2099 1978 1744 1775 2006 1880 1425 1378 Relative unit values (in teres of intra-IC unit values) 'Duping' countries Czekoslovakia 53.5 65.6 59.8 66.2 71.1 77.5 68.8 76.2 69.7 65.4 77.3 last-Germany 67.3 66.4 68.1 40.8 71.8 70.8 76.0 76.8 71.9 65.5 64.0 Poland 161.5 33.4 75.4 72.3 76.6 90.9 86.8 74.7 Soviet Union 67.8 69.7 56.9 72.1 72.0 70.9 63.9 77.9 63.1 52.3 58.6 Ron-IC West-lurope Austria 100.9 105.6 97.5 98.8 100.0 97.0 97.1 99.6 96.1 86.6 89.7 Iorway 99.2 86.2 98.2 95.9 95.1 99.7 96.1 95.4 90.9 97.0 87.6 Finland 102.7 99.9 97.6 102.0 101.0 97.0 100.8 107.2 96.3 100.3 103.6 Sweden 117.0 123.7 115.9 116.8 118.7 125.1 121.0 114.9 114.0 121.9 120.4 Spain 87.7 85.7 89.2 91.7 95.0 103.7 89.9 95.0 89.4 91.2 91.0 Niddle-Iast countries 51.8 70.1 73.3 70.2 79.8 [ntra-IC trade 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 unit values [cI 1620 1419 1974 2204 2121 1840 1832 1942 1809 1404 1358 Total IC isports 100.3 100.8 99.3 98.9 98.8 99.2 99.6 100.4 96.9 96.2 97.4 Sources: lurostat, various issues. Own computations. lotes : [3] September 1982: initiation of the antiduping proceeding. [b] Hay 1983: undertakings vith all the 'duping' countries. [c) in 1982 Deutsche harks per ton. 18 Table 5. luropean Comuuity trade in PVC, 1978-1$87 Countries 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 (al [b] [cl lmport market shares, in t of total (intra and extra) IC isports 'Duping" countries 4.1 4.8 4.8 7.9 10.1 8.7 6.9 6.4 5.8 4.4 3.8 Czechoslovakia 0.8 1.2 1.7 2.5 3.0 3.5 2.7 2.8 2.6 2.0 1.6 last Gerany 0.5 0.9 1.2 1.6 2.3 2.8 2.7 2.4 1.7 1.4 0.9 Hungary 0.5 1.0 1.5 2.7 2.6 1.2 1.2 1.0 1.3 1.0 1.1 Rolania 2.3 1.7 0.5 1.0 2.2 1.1 0.4 0.2 0.2 0.0 0.2 Non-IC Ind. countries 3.9 3.9 3.8 3.6 3.3 3.8 0.2 4.6 3.0 3.3 3.3 Norway 1.6 1.9 1.7 1.5 1.3 1.7 na 2.2 1.8 1.7 1.6 Sweden 2.0 1.7 1.8 1.2 1.1 1.7 Da 2.1 1.1 1.5 1.3 USA 0.3 0.3 0.4 0.9 0.9 0.4 0.2 0.2 0.2 0.1 0.3 Hiddle-last countries 0.0 0.5 0.6 1.0 2.3 2.1 0.0 3.0 2.8 2.8 2.0 Intra-EC trade 86.6 83.9 84.0 82.2 79.2 80.2 79.0 78.7 79.6 78.4 79.9 Total IC isports 100.0 100.0 100.0 100.0 100.0 !00.0 100.0 100.0 100.0 100.0 100.0 in tons 996.8 1045.3 1197.6 1142.8 1222.6 1069.4 1184.9 1234.8 1361.4 1551.1 1697.4 lxtra-IC exports tons 469.4 530.6 590.5 508.0 521.4 481.5 625.8 714.2 739.7 636.5 605.0 unit values [td 1898 1596 1908 2019 1873 1709 1612 1746 1619 1445 1450 Belative unit values (in term of intra-IC unit values) Duuping countries Czechoslovakia 69.2 64.2 68.9 66.8 70.3 63.8 73.7 75.3 71.7 73.3 73.8 last Germany 78.8 77.1 76.8 69.5 79.5 67.2 70.9 73.7 72.3 71.0 72.3 hungary 77.8 73.2 69.4 74.3 74.8 72.4 79.5 77.8 76.0 75.3 76.6 Romania 69.9 67.1 62.6 70.8 66.7 64.0 68.0 66.8 74.6 0.0 66.8 Ion-IC countries lorway 114.8 115.6 106.8 113.5 129.3 117.8 na 107.4 109.8 98.3 102.7 Sweden 101.5 98.0 105.8 117.9 130.2 113.9 na 108.5 121.0 110.6- 118.7 USA 186.0 186.1 139.1 116.6 139.3 207.9 268.8 283.6 328.0 294.1 198.0 Biddle-last countries 0.0 97.8 102.7 113.9 100.9 85.4 70.0 88.7 85.9 81.2 83.7 Intra-IC trade 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 unit values (d] 1754 1565 1824 1795 1540 1509 1558 1689 1561 1405 1390 Total IC imports 99.2 98.2 98.3 98.4 98.4 97.3 98.3 98.6 98.9 98.5 99.4 Sources: Iurostat, varioas issues. On computations. Iotes : [a] Decesber 1981: initiation of the antidumping proceeding. (b] September 1982: undertakings for last Germany, Hungary & Romania; provisional duty for Czechoslovakia. [c] January 1983: undertaking for Czechoslovakia. [d] in 1982 Deutsche larks per ton. 19 enter the EC market-- by well-established firms more willing to follow cartel policy. That the sum of the intra-EC trade and the imports from the five other countries was stable over the 1980-1984 period leads to the conclusion that the cartel has been successful. Antidumping measures had such an impact because they doubled the already high tariffs (12.5%). In the PVC case, the ad valorem equivalents of the undertakings are easily computed because the estimates of the margins of injury (i.e., margins by which prices of the foreign producers should ba raised in order to allow Community producerc to cover their production costs and earn a "reasonable" rate of profit) are mentioned by the antidumping proceeding. The ad valorem equivalent of the global protection can thus be estimated at 20%- 25%, roughly. In the LdPE case, there are no official estimates of the margins of injury. However, the fact that the dumping margins and the price undercuttings were similar to those in the PVC case suggests the same order of magnitude for the tariff equivalent of the antidumping measures and the corresponding global protection.9/ Section 3. The capture of the antidumping procedures by the cartels Section 2 leaves little doubt about the crucial role of the antidumping actions in the survival of effective cartels. This conclusion leads to the next question: were the antidumping procedures captured by the cartels? Though they never mention the antidumping actions, the anticartel proceedings provide ample evidence showing such a capture at the four major steps of an antidumping proceeding: the definition of the market, the 9/ This estimate fits with the conservative hypothesis of one third of the dumping margin, which is based on the few observations available. 20 existence of dumping, the determination of injury (the injury tests) and the duration of the measures (the sunset clause). The capture of the definitions of the markets The definition of the industry consists of two elements: the petitioners and the definition of the product. Both antidumping complaints were lodged by the European Council of Chemical Manufacturers' Federations (CEFIC) "on behalf of the EC producers accounting for the Community entire production" i tr-he polyethylene case [C230:2] and "on behalf of 24 EEC producers repre'senting almost the total Community industry" in the PVC case [C332:2]. Tables 2 and 3 provide a first crucial result: all the petitioners mentioned in the two antidumping complaints are cartel members. Exxon (a significant European producer of LdPE but a negligible one of PVC), and to a lesser extent, BP and Shell in the polyethylene case are the exceptions. However, the fact that Exxon, BP and Shell supported the antidumping actions suggests that these firms were not totally immune to the cartellization process, as it makes good economic sense. One should also note that foreign firms operating plants in the EC (as Dow or Exxon in the LdPE case and Norsk Hydro through Vinatex in the PVC case) were considered as EC producers in the antidumping procedures. This is in sharp contrast with the approach used in a case (Bisphenol) --initiated five months before the polyethylene case and terminated four months after-- where subsidiaries of US firms operating plants in the Community were not considered as part of the Community industry. Products in antidumping actions are defined by the complaining firms whereas products in anticartel cases are defined by the authorities in charge of competition rules. This difference has a strong impact in the coverage of the cases. Complaining firms in antidumping actions are induced to impose the 21 widest definition of the product in order to get a more complete protection. Authorities in charge of competition rules tend to take narrow definitions in order to make their proofs stronger. The polyethylene cases illustrate the point. The antidumping case covers the two types of polyethylenes existing in the early 1980s (low and high density). However, almost 90% of the East European exports during the period examined consist of LdPE. By contrast, the anticartel procedure was limited to low density polyethylene. Definition issues are crucial when the initial regular tariffs are high. In the EC, both the polyethylenes and PVC benefit from the highest rate of tariff (12.5%) for plastics. To impose undertakings (or some equivalent measures of antidumping duties) on the exports of high density polyethylene from East Europe when, by far, LdPE constituted the predominant exports is a disguised unilateral increase of the tariff levied on high density polyethylene. The capture of the dumini determination Margins of dumping are the differences between the overseas and export prices of the exporters. The notices of initiation mention that the EC petitioners provided data revealing "considerable" dumping margins. In the PVC investigation, the antidumping office of the Commission concluded that the margins varied between 53-68%, 58-74%, 26-37% and 25-45% for imports from Czechoslovakia, Romania, East Germany and Hungary, respectively. In the LdPE case, the antidumping office concluded that the weighted average margins of dumping were roughly 47%, 40%, 38% and 30% for imports from the Soviet Union, Poland, East Germany and Czechoslovakia, respectively. The crucial question is: were these huge margins of dumping resulting from computations captured by the petitioners? 22 All the exporting countries under antidumping actions were non-market economies at the time of the investigations. In such a case, GATT and EC rules allow the investigating authorities of the importing country not to rely on the actual overseas prices for determining the dumping margins but to find prices in a "country of reference" which could be considered as a reasonable proxy of the exporting countries. The choice of the "reference country" is thus crucial. In both LdPE and PVC cases, Sweden was chosen.l/ The Commission justified the use of the prices on the Swedish domestic market for estimating the overseas prices in the "dumping" countries on two main grounds. First, according to the antidumping office, basic conditions for the production were comparable."/ At a first glance, this argument may seem surprising when one looks at factor markets. Swedish "actor prices (for instance, wages) have little in common with the corresponding prices in the "dumping" countries. However, factor costs (such as labor costs) are minor in the plastic productions. According to various issues of the Chemical Economics Handbook, the wage bill represented 3% to 5% of the total production costs of LdPE in the early 1980s and a similar figure seems reasonable for PVC. In sum, lo/ In both cases, the choice of Sweden was suggested by the petitioners whereas other (not specified) parties advocated a comparison with the US or Canadian prices. In the PVC case, the antidumping office rejected the comparison with the US on unspecified "certain structural factors." In the LdPE case, the office rejected it on the main ground that the US LdPE production was based on natural gas --contrary to the EC production-- the price of which was regulated [L138:66]. Both antidumping procedures gave no specific explanations about the refusal to compare with Canadian prices. ll/ In both the LdPE and PVC cases, the antidumping office specified that conditions of production are broadly comparable in Sweden, the exporting countries named in the complaint and the Community [our emphasis]. The mention of the Community is puzzling. 23 this first argument on factor markets --though an economic nonsense-- is without substantial consequences. The second argument of the antidumping office concerns the markets of the products examined. In the LdPE case, the fact that "well over half of Swedish demand is covered by imports, mainly from other Scandinavian countries and the Commu-ity" led the antidumping office of the Commission "to the conclusion that the price level on the Swedish market is reasonable in view of the conditions of competition prevailing there" [L138:66, our emphasis]. In the PVC case, a very similar sentence led the antidumping office to the conclusion that "the Swedish market was an appropriate reflection of the conditions of competition for Western European PVC manufactures and that the level of prices was reasonable" [L274:16, our emphasis]. However, these assessments on the level of competition nrevailing in the Swedish markets are contradicted by the two anticartel decisions which both suggest that the LdPE and PVC cartels were more powerful in the Swedish market than in the EC markets. In the early 1980s, there is a unique LdPE producer in Sweden: Unifos Kemie, a joint subsidiary of Union Carbide and KemaNobel bought in 1984 by Neste Oy, the Finnish LdPE producer. Both Unifos and Neste Oy are mentioned by the EC competition office as founders and active members of the cartel. For 1981, the Unifos market share of the Swedish consumption is estimated at 30%. The rest of the Swedish demand is provided by imports, more than half (51.2%) coming from EC producers and more than one third (37.2%) from the non-EC producers members of the cartel. As a result, the cartel was --directly or indirectly--able to control 100% of Swedish production and 90% of Swedish demand, i.e., more than the EC production and consumption. 24 Looking at the PVC case, there are two producers operating a plant in Sweden: KemaNord and Norsk Hydro, both mentioned by the EC competition office as founders and active members of the cartel. In 1981, the two producers covered 42% of Swedish consumption. Swedish imports were mainly from EC (48%) and Norwegian (37%) sources. In sum, the cartel was able to control 100% of Swedish production and 93% of Swedish demand, i.e., again more than the EC production and consumption. One might argue that firms can belong to a cartel in one market and be under competition in another one. However, there is strong evidence that this argument is not valid for both the LdPE and PVC cases. If valid, competition in the Swedish market between the firms' members of the EC cartel would have generated highly profitable arbitrage opportunities by re-importing from Sweden to the EC. Tables 4 and 5 show that it was not the case: the market shares of the imports from Sweden by the EC were decreasing slightly in the case of LdPE and substantially falling in the case of PVC, at least up to the year when antidumping measures were taken. The crucial conclusion that Swedish and EC prices were aligned is supported by an additional observation. The ratios of the unit values of the EC imports from the East European countries to the unit values of the EC imports from Sweden provide very close estimates of the dumping margins computed by the EC Commission. In the LdPE case, the average ratios for 1982 and 1983, weighted by the imports of these two years, led to dumping margins of 39%, 40%, 41% and 45% for Czechoslovakia, Poland, East Germany and Soviet Union, compared to the margins found by the Commission of 34%, 38%, 40% and 47%, respectively. In the PVC, the same computations for 1981 give dumping margins of 46%, 38%, 42% and 49% for Czechoslovakia, East Germany, Hungary and 25 Romania, compared to the average margins of respectively 60%, 31%, 35% and 66% found by the Commission.12/ In such circumstances, to use the LdPE and PVC Swedish prices as proxies for the overseas price of the East European exporters can be reasonably seen as a direct comparison between the EC cartel prices and the export prices of the fore.gn exporters. Such comparisons have nothing to do with dumping, as defined by GATT principles. They lead with absolute certainty to the conclusion of the existence of dumping. To summarize, both LdPE and PVC cartels have been able to capture the procedure determining the existence of dumping.13! The existing evidence might even raise the aggravating possibility that Swedish prices were indeed the highest possible prices of reference. First, Table 4 shows that the unit value of the EC imports from Sweden was 20% higher than the unit value of intra-EC trade (indeed they had by far the highest unit value). Second, the unit value of the Swedish LdPE imports was also higher than the unit value of the intra-EC trade (by 6% in 1983). This characteristic of the Swedish prices may be related to the firmer grip of the cartel on this market (and supports the assumption of a differentiated product). 12/ The comparison is more difficult to do because the antidumping office has only given ranges for margins of dumping (the averages given in the text are unweighted averages of the maximum and minimum). Nevertheless, the order of magnitude and the ranking are similar. 13/ In the PVC case, the antidumping office has an additional argument in favor of the choice of Sweden which has a strange resonance when linked to the evidence given by the anticartel decisions: the exporting countries named in the complaint take a "negligible" share of the Swedish market [L274:16]. Economists would argue that such a feature is a sign of lack of competition and would induce not to consider Sweden as the country of reference. 26 The caRture of the iniury tests With regard to injury, both antidumping investigations based their affirmative conclusions on three arguments. The shares of the Community market held by imports from East Europe had risen. The prices of the Community producers had been undercut by the selling prices in the Community of LdPE and PVC originating in the "dumping" count.ies. There was a "causal relationship" between allegedly dumped imports, falling prices and losses: "the dumped imports constituted a significant factor in the inability to contain an accelerating fall in (LdPE) prices [and] the losses sustained by the Community producers were increasing to an extent that was no longer tolerable" [L138:67] and "the dumped imports have been an important factor in depressing (PVC) prices [and] Community producers are making unacceptable losses" [L274:17].14/ Were the injury tests in the antidumping procedures captured by the cartels, directly or indirectly, in the sense that these elements were straightforwardly seen as proving injury, or were they were merely mirroring the impact of the cartel? Risina imort market shares in cartelized markets. The straightforward interpretation of rising import market shares as a sign of injurious competition is a good example of "indirect" capture. To consider rising import market shares as an injury in the case of cartelized markets is a major extension of the definition of injury. It means that domestic cartels are entitled to restrict sales and raise prices, that 14/ Any economist will look at these evolutions as signs of healthy competition. He will portray the first argument as changing the antidumping procedure in a selective safeguard procedure based on import surge conditional to the "existence" of dumping. He will consider the restrictions to the two last ones as likely infringements to Articles 85 and/or 86. 27 foreign competition is not allowed to erode their market power, and that antidumping actions should protect the rents of the domestic oligopolies. Moreover, to base the injury test on the existence of high market penetration ratios in some EC states ignores the fact that increases of market share depend on market structures which are shaped by domestic firms more than by foreign ones. Two hypotheses can be suggested. The more homogeneous the cartel is, the more similar the increases in import market shares in the various EC states should be. And the harder the cartelized firms try to enforce "price umbrellas," the more the mavericks have both the incentives and the opportunities to increase their sales. The penetration ratios by EC states given by the antidumping proceedings support these two hypothesis and provide an interesting "mirror" image of the cartels. The first hypothesis is supported by both the stable dispersion of the market shares of East European imports in the PVC case (seen as a strong cartel by the firms involved and by the authorities in charge of competition rules) during the period of "dumping" and the increased dispersion in the LdPE case (seen as a weaker cartel) which means that some national markets were less cartelized than others.15/ The second hypothesis fits well with the fact that in the LdPE case (where dispersion has changed), the market with the lowest increase in the East European import market share (Britain) is also the market where the two local producers (BP and Shell) were only marginally involved in the cartel [L74:32]. 15/ Considering the evolution of the market shares of East European LdPE and PVC in Belgium, Britain, France, Germany and Netherlands, the dispersion (measured by the Herfindhal coefficients) varies from 0.23 to 0.26 in the PVC case, and from 0.34 to 0.28 in the LdPE case. 28 Price undercuttings. The second element on which the injury tests are based concerns the existence of price undercuttings to different degrees, the maximum mentioned by the antidumping office being 20% to 25% in both products. Tables 4 and 5 confirm this information by computing price undercuttings from the import unit values of the four East European countries and the unit values of the intra-EC trade. They show that LdPE and PVC East European import unit values were 25% to 35% lower than intra-EC unit values in 1981-1983. With the 12.5% tariffs, these price differentials are reduced to roughly 15% to 25%, i.e., the range of price undercuttings. In other words, East European LdPE and PVC price undercuttings in the EC markets simplv mirror the relatively lower export prices of these countries.16/ The "causal relationship" between imports and prices. Evidence presented in the two anticartel investigations does not support the first part of the causal relationship, i.e., the link between "dumped" imports and falling prices. Figure 3 shows that over the ten last years, the LdPE and PVC prices (in current Deutsche Mark) were subject to price cycle. The price drops occurring before the antidumping actions appear to be of normal magnitude. Indeed, the main falling trend of both prices occurs after the antidumping measures. (This apparent paradox will be explained later). That price cycles occurred before and after antidumping measures implies that imports (once authorized, then prohibited) have little to do with such cycles.17/ 16 This conclusion is important in terms of the capture of the procedures: it means that the validity of the argument on price undercutting depends on the proof of the existence of dumping which --as shown above-- has not been done. 17/ If some evolutions could be qualified as "abnormal," there are the price increases of 1979 (not shown because of no comparable data). 29 Insert Figure 3 This conclusion is reinforced by the converging evidence provided by Figures 1 and 2 and the anticartel proceedings. Figures 1 and 2 show that the downsides of the price cycles began only five (LdPE) and two (PVC) months before the initiations of the antidumping actions. Such short delays suggest that complaints were loc,gea with little (if any) information on the ongoing evolution of the market prices. Moreover, "target" prices --manipulated by the cartel members-- did not reflect accurately the evolution of market prices. In the LdPE case, target prices in the few months before and during the investigations gave a more pessimistic view of the situation than prevailing prices (they declined by more). Another piece of evidence is given by the comparison of the unit values of the "dumped" imports in relative terms (i.e., in terms of the unit values of the intra-EC trade taken as a proxy for the EC prices) and in absolute terms (i.e., in constant Deutsche Marks). Absolute prices of East European exports have fallen in half of the cases. However, as shown in Tables 4 and 5, the unit values of the East European LdPE and PVC exports in relative terms have been stable, except in very few situations.18/ In other words, the East European and intra-EC prices have fallen in the same proportions. That leads to the crucial question the antidumping proceedings did not address: who triggered the price decreases? In the case of unstable cartels, the answer is likely to be firms belonging to the cartels. The two anticartel decisions support this conclusion in quesi-identical terms: "A number of 18/ If the years of investigations are excluded (East European prices can then be a reaction to the antidumping actions), there are only two (both in PVC) where relative prices decrease by more than 5%. 30 Fi'ure . E-' cherT,iI:'M I priCezi, 1 1 .'-89 [Sourm: European CFomimi Hem.] 2.4 - 1.2 > 1.2 Ethylene I 0.4 A,.. 0.+- Naphtha ', 0.2 11'. HIM _1 * 8. MM _ ,_ 8O.D1 81.01 52.D1 53,.81 54.01 FS.01 SF8.01 57.01 55.01 59;.81 31 producers who took part in the meetings were named as aggressive in certain markets by other producers who considered themselves as strong supporters of price initiatives and were prepared to lose volume in order to force through an increase" [L74:26] and [L74:22). In sum, the East European exporters were essentially meeting the competition on the fringe of the cartels, i.e., aligiing their prices in the fringe EC markets where competition was still going on. Why did the cartels pay so much attention to trade flows, which represented a small proportion of the imports and a minute proportion of the EC consumption and the relative prices of which were not dropping, but at most slowly eroding? Economic theory says that limited variations of supply can generate substantial price changes because markets are cleared at the margin. However, the European chemical markets are known for larger "captive" markets and more extensive price discrimination between large and small buyers than other chemical markets, such as the US one. In such sticky markets, lower prices can exert a pressure only if suppliers are credible, i.e, if they can offer good prices for a long period. As a result, it was important for the cartels that the foreign suppliers not be viewed as credible in the long run. The anticartel decision in the PVC case illustrates this point by mentioning a 1980 note by a cartel member specifying --among other measures-- that the "deltas" for importers should be limited to 2% and "tourism" reduced [L74:17].19/ The antidumping measures made "tourism" very expensive. 19/ A "delta" is a permitted discount off price. "Tourism" is the phenomenon whereby customers faced with a price increase from their regular supplier seek lower quotes from other producers. 32 A last piece of evidence showing the capture of the antidumping procedures by the cartels concerns the imports of LdPE and PVC into the Community from the countries not mentioned in the complaints. The antidumping office looked at these imports in order to assess whether they contributed --or not-- to the injury done to the EC firms. In the LdPE case, the antidumping office noted that these imports "were imported into the Community at prices in line with the market" [L138:67, our emphasis]. In the PVC case, the office noted that "such imports have declined with the level of demand in the Community. changing their market share little" [L274:17, our emphasis]. Thus, the antidumping office came to the conclusion that these imports should not be cumulated with the "dumped" imports. However, the bulk of the extra-EC imports to which the antidumping office referred came from the rest of Western Europe, i.e., from firms, a vast majority of which are shown as members of the cartels by the anticartel investigations. In other words, the considerations about prices in line and constant market shares quoted above represent an implicit exemption to competition rules. The "causal relationshiR" between imRorts and losses The second part of the "causal relationship" --the links between prices and losses (profits)-- can be analyzed with Figure 3 which relates the LdPE and PVC prices to naphtha and ethylene prices. Naphtha is produced from oil and is the essential input of ethylene. Ethylene is the main input for LdPE and PVC. For instance, in 1981, ethylene represents roughly 60% of the LdPE and 50% of the PVC production costs. The output (LdPE and PVC) price falls after 1985 largely reflect the large falls of the naphtha (oil and dollar) and ethylene prices. 33 The "spread" between the input (ethylene) and output (LdPE and PVC) prices is the major determinant of profits, for a given level of efficiency. During the early 1980s, the spread was too small to allow the EC firms to be profitable at the existing level of efficiency of the EC industry.20/ The larger spread achieved in 1984 --mainly by the increase of the LdPE and PVC priccs through cartelization-- allowed EC firms to reduce their losses or become profitable, as shown by Figure 4. In sum, the cartel strategy was to absorb the ethylene price increases by increasing LdPE and PVC prices as much as possible in order to restore a profitable spread between input and output prices. The antidumping actions were a decisive part of this strategy. Insert Figure 4 However, this strategy of raising prices in a attempt to cover the costs of the existing equipment maintained a massive excess supply, or the so-called phenomenon of "overcapacity." As this effect becomes more visible in the long run, it leads to the last aspect of the antidumping procedures: duration. The capture of the sunset clause In the EC law, the duration of antidumping measures is five years. The measures can be repealed earlier if there are reviews proving that dumping has ceased. There were no reviews in both the LdPE and PVC cases and the expiry of the measures was announced in December 1987 for the three countries that immediately offered undertakings in the PVC case. The measure for the fourth country caught in the PVC case and all the measures in the LdPE case expired in May 1988. As of November 1989, no new complaint has been lodged. As a result, 20/ The fall in oil, dollar and ethylene prices led to significant reductions in the upstream profits which were offset by a strong performance in the downstream earnings. 34 FigtJre 4. LdPFE mnarlgins, 1 978-85 Coah ood .:Sour: Chem S,tmiwF 0.- 0..5 - L.2 L.1 -0.5 _ - ', 6 -0.1 ZO-.1 3f.1 01.1 0z.1 B3.1 84.1 Z5.1 Y r. quv l 35 the sunset clause seems to have played its role. However, a correct assessment of the full impact of the sunset clause cannot be limited to the observation that the antidumping measures were transitory. It should consider two other points. Did the protection granted through antidumping measures open a period of spreading protection? Did the relief granted lead to economically soundI adjustments or did it offer an opportunity to reinforce anticompetitive forces? Two years after the imposition of the antidumping measures, additional protection was granted against imports of LdPE from Saudi Arabia. In 1985, Saudi Arabia began to export a variety of LdPE (linear LdPE) that the EC industry was very slow to introduce because it was trying to amortize existing LdPE plants in successfully cartelized markets. As shown by Table 4, Saudi LdPE and linear LdPE exports increased tremendously.21/ In 1985 and 1986, the EC General System of Preferences (GSP) classified Saudi exports of LdPE and linear LdPE as subject to a "tariff ceiling" procedure which makes the imposition of full tariffs (once a ceiling is reached) conditioned on the request of an EC state. The full tariff (12.5%) was imposed in July 1985, but in 1986, it took only five working days for the Commission to impose the full tariff on Saudi LdPE and linear LdPE exports.22/ In 1987, the GSP status of Saudi exports of linear LdPE became subject to the procedure of "fixed duty free amounts" which makes the imposition of the full tariff automatic, once the 21/ The two partners of Sabic --the Saudi producer and exporter-- are Mobil and Exxon. 22/ According to the CEFIC, Saudi Arabia sold polyethylene to the EC in volumes totalling an equivalent of four months' sales in 1985 [European Chemical News, 27 January 1986]. 36 duty-free quota is reached.23/ In those circumstances, antidumping procedures were a first step in a move by the industry to get more protection. Has the high level of protection from the antidumping measures --roughly 20% to 25%-- facilitated adjustment? Adjustment ("rationalization") generates benefits when inefficient plants are eliminated or improved. It also generates costs when implemented through means which may reinforce anticompetitive forces, such as mergers. Since 1981, there have been a substantial number of partial mergers, i.e., mergers of LdPE and PVC activities of two different firms. The number of firms actively involved in the LdPE sector has been reduced from 23 (1981) to 17 (1984) and since then has remained stable (1988). Jn the PVC sector, the respective figures are 28, 19 and 18, with the additional circumstance that two firms (Solvay and the common subsidiary of ICI and Enichem) represent 40% of the total production capacity. Profits accruing to new (and old) firms surged. For instance, European Vinyls Corporation (EVC) --the joint subsidiary of ICI and Enichem producing PVC-- was making profits estimated at $85 million in 1987 and was mentioned as the largest single contributor to Enichem performance for that year. However, these profit surges can mirror ethylene price decreases or enhanced capacity of the remaining firms to collude more than rationalization. 23/ The quota granted to Saudi linear LdPE under the new procedure has been extended from Ecus 6.6 million (1986) to Ecus 10 million (1987). It is worthwhile to note that it is widely accepted that transport and marketing costs (more than $ 200 per ton with the duties) for Saudi exports are high enough to almost eliminate Saudi comparative advantages when ethylene is cheap (Saudi production is based on ethane). Figure 3 show that ethylene prices are up after 1987 and Table 5 shows a substantial increase in the Saudi unit values in 1987. It is also worth mentioning that two studies done in 1985 suggested that duties of 13% were enough to eliminate the Saudi comparative advantage [European Chemical News, 15 and 22 April 1985]. 37 Additional considerations are thus necessary. They raise doubts about the extent of the rationalization process. First, Tables 2 and 3 show no substantial reductions in the total production caRacity of the Community, a worrisome fact when one remembers that time to reduce "overcapacity" was the rationale of the antidumping measures. Second, the concentration ratios of Lhe share (in terms of production capacity) by firms increased up to 1984 and then decreased when the antidumping measures took effect. In LdPE, the ratio increased by 20% between 1981 and 1984 and decreased by 2% between 1984 and 1988. The figures for the PVC sector are 53% and -8%, respectively. Third, the concentration ratios of the share (in terms of production capacity) by Rlants (independently from the ownership) show an even worse evolution. In the LdPE sector, the concentration ratio decreased by 5% between 1981 and 1984 and increased by a mere 0.8% between 1984 and 1988. The ratios for the PVC sector increased by 5% and then decreased by 5% again. In sum, there is some evidence of a rational*.zation process only for the years, i.e., before the antidumping measures took effect.24/ One may argue that the dramatic change in the prices after 1985 may explain the. relaxation of the rationalization effects. However, one may also argue that the antidumping measures were adding another motive to relax adjustment efforts. The most worrisome consequence of tt-is lack of adjustment is that "overcapacity" will still be an argument for protection available during the next business cycles in LdPE and PVC. Politically powerful, the argument is 24/ Industry sources suggested an "overcapacity" of 1.7 millior. tons in 1984 in the LdPE sector. By 1987, mothballing reactors had cut capacity by 0.9 million tons. [European Chemical News, 198.]. 38 based on an economic mistake. "Overcapacity" can be defined as an excess supply at a given output price. By closing the EC markets behind antidumping measures, by making the LdPE and PVC markets more cartelized so that they were able to increase their prices and get the appropriate spread (at given ethylene prices and conditions of production and efficiency) for generating profits or eliminating losses, the EC firms transformed a problem of low-cost imports into a problem of too high-priced domestic production, i.e., into a problem of "overcapacity." As soon as the ethylene price decrease led to decreased output prices in the late 1980s, demand increased and "overcapacity" has been replaced by "undercapacity." The danger of the renewed use of the "overcapacity" argument for more protection is increased by the evolution of EC exports. EC exports of both LdPE and PVC are declining, another sign of the increased inefficiency of the EC industry and its insufficient rationalization. One can expect this decline to continue because of the increased production capacities in the Middle East (LdPE and linear LdPE) and in the Far East (roughly 1.330 million tons until 1992 in vinyl chloride monomer, an intermediate stage of production between ethylene and PVC). In sum, the limited duration of the antidumping measures did not prevent the negative effect (a increased concentration through mergers) while it did not generate the positive effect (a drastic adjustment of production capacity). The costs ir. terms of potentially stronger cartellization power do not seem to be compensated by benefits in terms of "rationalization." 39 Section 4. The anticartel actions: no deterrent effect How large were the costs of the antidumping actions from the "general interest" perspective of the EC consumers? Large costs could be expected since the two products alone represent sales of over Ecus 6 billion annually, i.e., more than 3% of all Western European chemical industry sales. There is a corollary question. Were the anticartel measures severe enough to inhibit future antidumping actions by recurrent cartels (as the previous section has shown possible)? The answer to this question focuses on the costs and benefits from protection and competition from the point of view of the firms. The costs of antidxuping measures in cartelized markets Section 2 shows that the LdPE and PVC sectors were dominated by domestic producers having a joint monopoly power while "dumping" foreign firms had a more competitive behavior. Section 3 suggests that imports were substantial enough to allow the assumption of a horizontal import supply curve for the price range to be considered. It also shows that the intra-industry imports and exports were large. As a result, LdPE from different sources were likely to be considered as differentiated (due to quality or timing) products by the users. A similar trade pattern leads to the same conclusion for PVC. The "tourism" concept used by the cartel confirms this differentiated product approach. The costs of the antidumping measures can be calculated within a range of minimum and maximum estimates. These estimates are based on the assumption that the antidumping measures imposed --the undertakings-- consisted of minima 40 prices or quotas.25/ The minimum estimate assumes that the LdPE and PVC cartels would have been able to emerge and survive even in the absence of import restrictions. Antidumping measures might have helped tr. cartels but were not necessary. Both the evolution of the prevailing prices and the documents quoted by the anticartel decisions suggest it is not the most likely situation. However, it is useful to have an idea of the minimal costs of protection. The maximum estimate assumes that the LdPE and PVC cartels would have collapsed in the absence of the antidumping measures and that drastic changes due to stiffer competition would have prevailed. This situation seems the most likely one for both the LdPE and PVC cases. These two estimates also refer to the degree of substitutability between the EC LdPE and PVC and the corresponding East European products.26/ Low substitutability is what allows cartels to impose the monopoly surcharge (measured by the markup of price over marginal revenue) without having to worry too much about the ability of users to buy cheaper (since traded on more competitive markets) imports. In such a case, antidumping 25/ One may argue that this hypothesis --the most likely in an environment of cartelization and trade with East Europe-- does not fit with the fact that the decrease in East European exports was not accompanied by noticeable changes in the unit values of these exports, as shown by Tables S and 6. This price stability might be explained by two reasons. First, prices have not been adjusted to the new constrained equilibrium. Non-clearing markets have led to some kind of rationing and the rents associated with the import quotas or VERs were captured by the purchasers of East European LdPE. Alternatively, prices may have been adjusted. However, East European firms might have wanted to keep their share of the rents in Deutsche Marks or other EC currencies because of exchange controls. As a result, the rents granted to East European firms have been captured by the trading correspondents or affiliates of the East European Exporters. 26/ LdPE and PVC are not substitutes. They have different technological characteristics and uses. 41 measures are more a clear signal sent to the users about the cartel's willingness to impose price discipline than a tool for dominating markets. By contrast, high substitutability restricts cartels in their ability to surcharge their buyers because these buyers can buy cheaper imports. In such a case, antidumping measures are necessary for the enforcement of the cartels. Morecver, as underlined by Morkre [1986], they generate higher costs in the domestic and the import markets --even if the level of protection imposed on the imports is initially small-- than expected for two reasons. They allow the cartel to impose full monopoly prices instead of monopolistic prices "limited" by the potential shift of consumers from the domestic products to the foreign ones. And the price increases of the domestic products make it profitable for some users to shift from the domestic to foreign suppliers, i.e., make the constraints imposed on the imports more binding than they initially were.27/ The cost estimates for the LdPE consumers The minimum estimate can be calculated by comparing two sets of prices: the market prices observed by the competition office from June 1983 to September 1984, i.e., during the period the cartel was operating under the protection of antidumping measures; or the prices observed from January 1981 to May 1982, which can be used as the benchmark for the prices which would have prevailed if the cartel would not have been protected by antidumping measures (and yet would have survived). These two sets of prices are comparable because 27/ In other words, the lower the substitutability is, the closer from the usual marginal revenue curve is the "constrained" marginal revenue curve (i.e., the one taking into account that the demand of the domestic product is a function of domestic price and the import price. In case of linear demands [Morkre, 1986], the difference between these two marginal curves is a function of the cross-price elasticities and an inverse function of the own-price elasticity of the demand of the imported products. 42 they prevailed under similar conditions of excess supply. The average price differential between these two periods is roughly 3%. With an average consumption of LdPE in the EC in 1983-84 estimated at 3.4 million tons, this price increase represented an additional expense of DM 204 million per year (at the given quantity sold) for the EC users due to the antidumping measures. The welfare losses are calculated on a year basis (according to the usual formula) and on a compounded basis (using an 8% discount rate) in order to take into account the fact that antidumping measures last five years.28/ Table 6 summarizes the various results, depending upon the value of the demand elasticity adopted. Insert Table 6 The maximum estimate is based on the conservative hypothesis that the cartel survived for only two years (1984 and 1985) owing to the antidumping measures and that the market conditions after 1985 were again competitive. To calculate this estimate requires evaluating the prices which would have prevailed if the EC markets would have been competitive. Two price series were used to compute such a benchmark: japanese LdPE export prices (based on indexes reported by the Bank of Japan) and US LdPE export prices (based on unit values). Both series (in indexes based on 1980=100) are shown in Figure 1. Applying the evolution observed for both the Japanese and US indexes to EC prices would have made EC market prices lower by 11.6% than the observed prevailing prices for the period June 1983-September 1984. This differential is interpreted as the 28/ The formula used is: 0.5. p2 .e.V, where P is the price increase, e the price elasticity of demand and V the demand value. 43 Table 6. The costs of antiduping easures in LdPI and PVC cases Lou density Polyvinyl polyethylene chloride (i) the minim estimate: the cartel would have survived without antidmping action Increase in prices (%] 3.0 10.0 Increase in expenses 1Dm millionl 204.0 496.0 Anul welfare loS6 1I million] low demand elasticity (1.3) 3.9 31.4 high demnd elasticity (1.8) 5.4 43.5 Compounded welfare loss (DI sillioni low desand elasticity (1.3) 23.0 184.1 high dmand elasticity (1.8) 31.8 255.0 (ii) the aximus estiate: the cartel would not have survived without antiduping action Increase in prices (t] 11.6 14.0 Increase in expenses [M million] 737.2 653.9 innual welfare loss (IM million] lov deand elasticity (1.3) 55.6 59.5 high deand elasticity (1.8) 77.0 82.4 Compounded welfare 1o8s (DM Billion] lov deand elasticity (1.3) 168.4 257.8 high demad elasticity (1.81 219.3 337.0 Sources: Previous tables. 44 contribution the antidumping measures made to cartelize the EC markets.29/ Based on this differential, the annual increase of expenses paid by EC LdPE users is DM 737 million in 1984 and 1985. Table 6 reports the corresponding annual and compounded welfare losses for the EC consumers. The cost estimates for the PVC consumers The same approach is used for the PVC sector. The minimum estimate is based on the comparison between the prices prevailing from November 1980 to November 1981 (the period when the cartel is supposed to have been able to operate without additional protection) and the prices prevailing from February 1983 to October 1984, when the cartel was functioning under antidumping measures. The price differential between these two periods is roughly 10%, much larger than in the LdPE case. With such a price differential, the EC consumers would have paid DM 496 million in excess. Table 6 reports the corresponding annual and compounded welfare losses for the EC consumers. The maximum estimate is again based on two price benchmarks: Japanese export prices (Bank of Japan) and US domestic market prices (Petrochemical Pricing News). Applying the observed evolution of these two price series (illustrated in Figure 2) to EC domestic prices indicates EC market prices would have been lower by 15.9% (on the basis of the Japanese prices) or 14% (on the basis of the US prices) than the EC observed prevailing prices. 29/ One may note the close level of the EC and Japanese indexes during the "dumping" period (from 1982 to mid-1983). The lowest point of the EC prices relative to the Japanese prices occurred in June 1986, one month after the antidumping measures. 45 Using the price differential based on the US prices, in order to take the more conservative approach, EC consumers would have paid DM 654 million in excess because of the antidumping measures. Table 6 reports the corresponding annual and compounded welfare losses for the EC consumers. There is an interesting difference between the LdPE and the PVC cases. The minimum and maximum estimates in the LdPE case are very different, whereas there are not in the PVC case. That suggests that the PVC industry has been more able to extract its monopoly power without antidumping measures than the LdPE sector. This conclusion fits well with the anticartel decisions (and the judgments of the firms themselves) on the relative weakness of the LdPE cartel compared to the PVC cartel. The cartel perspective: antidumping benefits vs anticartel costs The cartels look at the antidumping and anticartel actions in a totally opposite way than consumers and the "general interest." For EC firms, antidumping measures grant benefits because they bar foreign competitors and anticartel enforcement imposes costs (fines). What is the net result for the cartel members? If benefits exceed costs, one should expect new antidumping actions followed by anticartel decisions, a rather strange version of "managed trade and domestic competition." The LdPE anticartel case ended with the imposition of fines amounting to Ecu 37 million (DM 77 million) in 1989, a very severe decision by EC standards. For the member firms of the cartel, the antidumping measures provide more revenues (at the given quantity sold), DM 737 million according to the most likely (maximum) estimate. These increased revenues are gross of the fines. However, the fines still leave a positive net revenue of the antidumping measures for the producers of roughly DM 660 million. 46 The PVC anticartel case ended with the imposition of fines amounting to Ecu 23.5 million (DM 49 million) in 1989, also a severe decision by EC standards. The cartel members got DM 654 million in additional revenue (at the given quantity sold) from the antidumping measures, according to the maximum estimate hypothesis. The net revenues are thus over DM 600 million. These crucial results leave few doubts about the future behavior of the producers. Despite the severity of the competition office of the Commission, the antidumping complaints were a profitable operation from the producers' point of view. Moreover, one should consider two more points. First, the benefits from antidumping measures go to the firms with the largest market shares which will thus be the most tempted to lodge new complaints. For instance, EVC which represents 25% of the EC PVC production capacity has paid a fine of Ecu 6.75 million (the sum of the fines pai, by ICI and Enichem) or DM 14 million, for additional revenues of DM 163 million (one fourth of increase in expenses, if one assumes that capacities of production mirror sales as one should in the cortext of a market-sharing cartel according to the anticartel decision). Second, there is the possibility that the firms do not pay the anticartel fines out of their past profits but consider them as a cost of future production. For the EC consumers, this would mean that the antidumping actions would have indeed generated additional costs under the form of anticartel fines.30/ The 30/ These results are based on two implicit assumptions. Firms did not expect any antitrust action. The possibility of antitrust proceeding has no deterrence impact on the firms' behavior and does not induce the firms to take more competitive output and pricing decisions. The impact of ex ante deterrent antitrust rules has been examined by the literature on the antitrust enforcement [Block, Nold & Sidack, 1981; Smith & Vaughan, 1986] and leads to some harsh implications in terms of types of fines to impose. 47 more EC markets are protected from foreign protection (for instance, under antidumping measures), the more such a passthrough of anticartel fines is likely. Conclusion The paper provides four crucial conclusions based on Commission's documents and data. First, it shows that antidumping procedures have been easy to capture by the same few firms which lodged antidumping complaints in order to enforce cartel agreements. As a result, there are two procedures in the Community available for granting an exemption from competition rules, one under the Treaty of Rome and another one as a far-reaching consequence of a regulation only vaguely linked to the Treaty. Second, antidumping measures impose a high level of protection and are a crucial factor of cartelization. The paper estimates these measures have imposed a huge burden on EC users, DM 600 to 650 million, with compounded welfare (deadweight) costs ranging from DM 100 to 300 millions for annual sales of DM 6 billion. Third, antidumping protection may be transitory but it gives the time necessary to enhance the industry capacity to generate more stable and stronger cartels, whereas it may reduce the incentive to adjust. Lastly, the paper shows that firms have obtained large net benefits from antidumping measures, despite the anticartel fines imposed later. As a result, antidumping actions may be expected to recur. These results should be confirmed by studies on other cases: soda ash, synthetic fibers, ball bearings and consumer electronics. However, they are robust enough to send a very powerful message in terms of policy-making to the Community. 48 As mentioned by many economists, the first-best solution from the point of view of the Community and EC consumers is to unilaterally repeal the antidumping regulations. If the political will for such a drastic solution does not exist, it is necessary to subordinate the antidumping procedures to the competition rules, with all the consequences that implies. When countries are so integrated into world trade, it makes little sense to have two types of rules, one for "domestic" competition, one for "external" competition. Articles 85 and 86 should be applied to extra-EC trade when the distinction between intra-EC and extra-EC trade flows becomes meaningless. To deny this point leads to the "confusion des genres": the antidumping office will give de facto exemptions of competition rules when applying antidumping regulations while the comnetition office will impose --often too late because firms can put up considerable resistance to anticartel investigations-- trade liberalization when applying competition rules. 49 References Bank of Japan, Price indexes monthly (Wholesale price indexes), various issues, Tokyo. Block, Michael Kent, Frederick Carl Nold and Joseph Gregory Sidak, 1981, "The Deterrent Effect of Antitrust Enforcement," Journal of Political Economy, vol. 89, no. 3. European Chemical News, various issues. Lang, John Temple, 1989, "Reconciling European Community Antitrust and Antidumping, Transport and Trade Safeguard Policies--Practical Problems," Annual Proceedings of the Fordham Corporate Law Institute, Matthew Bender & Company, Inc., New York. Messerlin, Patrick A., 1989, "The EC Antidumping Regulations: A First Economic Appraisal, 1980-85," Weltwirtschaftliches Archiv, Band 125, Heft 3. Morkre, Morris E., "An Economic Framework to Analyze the Cost to Consumers and to the U.S. Economy of a Global Quota on Imports of Electric Shavers and Costs for Alternative Policies to Assist the Domestic Industry," Federal Trade Commission, mimeo. Official Journal of the EC, various issues. Salop, Steven C. and Lawrence J. White, 1986, Economic Analysis of Private Antitrust Litigation, Georgetown Law Journal, 1001. Smith, W. James and Michael B. Vaughan, 1986, "Economic Welfare, Price and Profit: The Deterrent Effect of Alternative Antitrust Regimes," Economic Inguiry, Vol. XXIV, October. van Bael, Ivo, 1979, "Ten Years of EEC Antidumping Enforcement," The Journal of World Trade and Law, September-October, p. 395-408. van Bael, Ivo and Jean-Francois Bellis, 1986, International Trade Law and Practice of the European Community: EEC Anti-Dumping and other Trade Protection Laws, CCH Editions Limited, Oxfordshire. van Bael, Ivo and Jean-Frangois Bellis, 1987, Competition Law of the EEC, CCH Editions Limited, Oxfordshire. 50 Annex A: Few firms in complex coalitions The simplest configuration of firm coalitions is shown by the low density polyethylene case. All the firms subject to the anticartel proceedings were petitioners in the antidumping case. All the defendants in the antidumping case were not members of the cartel. In sum, the antidumping action looks as a tool used by a weak and endangered cartel for eliminating foreign "mavericks," a point the following Sections will demonstrate. The aluminum case suggests a more complex configuration illust-rated by Table A:1. There are 40 firms involved as petitioners or as defendanLc in the antidumping and anticartel cases. However, only six parent companies are Insert Table Annex A:1 present in all three cases: Alcan, Alusuisse, EFIM, Norsk Hydro, Pechiney and VIAG. Interestingly, half of these firms are non-EC firms: this suggests the possibility of capture of a "national" instrument (antidumping regulations) by foreign firms. Two more firms --Elkem and Hoogovens-- are involved in the anticartel case and in one antidumping case. Together, the eight parent companies controlled two-thirds of the 40 firms involved in these three cases. The history of the cases suggests an ongoing coalition process. Alcan, Alusuisse, EFIM, Pechiney and VIAG are petitioners in the two antidumping cases. Norsk Hydro was a petitioner in the first antidumping case but a defendant --with Elkem-- in the second one. Hoogovens was only a petitioner. The first antidumping case (1982) covers an intermediate good (aluminum foils). The second antidumping case (1983-1984) covers unwrought aluminum which involves traditional suppliers of the cartel (USSR, Yugoslavia). The two cases together suggests that the antidumping measures may have had a strong impact on the vertical integration dimension of the industry. 51 Anex A: Table 1. The aluminiu antiduping and antitrust cases: the players Provisional Action Year irm roles lirms involved Country Parent cospanies Country ID 11 1982 Defendent not specified Israel affiliation unknown AD S1 19a2 Petitioner ILVAL Greece affiliation unknow AD 81 1982 Defendent not specified GDR affiliation unknwn AD 11 1982 Petitioner Cositel Italy affiliation unknown AD 11 1982 Defendent not specified Hungary affiliation unknown AD 82 1983 Defendent Raznoimport uSSR affiliation unkhow AD 12 1983 Defendent lombinat Aluminiji a Yugoslavia affiliation unkown ID 12 1983 Defendent Eaergoinvest Yugoslavia affiliation ukhown AD 81 1982 Petitioner Bacofoil dl Alcan Canada AD 82 1983 Petitioner Alcan Aluminium UK UK Alcan Cuaad Acartel 1984 Alcan Aluminium U1 UK AlcaD Cada AD 12 1983 Defendent Aluiniu Cy of Egypt gypt Iluiniu Cy of Egypt gypt AD 81 1982 Petitioner Aluminium-Waltzverke Singen Gerany Alusuisse Suitzerl'd AD 81 1982 Petitioner Star Aluminiu UK Alusuisee Suitzerl'd AD 12 1983 Petitioner Alusuisse Deutschland Germany Alusuisse Svitzerl'd Acartel 1984 Alusuiase Deutechland Gersany Alusuiuse Suitzerld Acartel 1984 BICC UK BICC Ul AD 11 1982 Petitioner Sava-Alluinio Veneto Italy fill Italy AD 12 1983 Petitioner Alluiinio Itaiia Italy RIll Italy Acartel 1984 Alluminio Italia Itaij Eill Italy Icartel 1984 Grages Aluminiu Sueden Electrolux Sveden AD 12 1983 Defendent llkes Mosal Aluminium Yorway Ilkle lorwa Acartel 1984 llket Spigerverket lorway Elkem Norway AD 11 1982 Defendent Teich Aluiniu Austria Iolienvalzverk Briiden Teich Austria Acartel 1984 lisen und etall Germany Hoesch Gemn AD 12 1983 Petitioner Estel Aluminiu Netherl'd Hoogovens letherl'd Acartel 1984 Hollad Aluminium letherl'd Hoogovens letherl'd Acartel 1984 Kaiser Aominium Europe IC Kaiser Aluiinn & Chemicals USA Acartel 1984 Hetallgesaellachaft Gerany Nstallgesellachaft Germany Acartel 1984 Noatedison Italy dontediaon Italy AD 11 1982 Petitioner ASV lolies Danemark Norsk Hydro lorway AD 82 1983 Defendent Sor gorge Alainium lorway Norsk Hydro lorway AD 12 1983 Defendent Ardal og Sunadal lorway Norsk Hydro lorway Acartel 1984 loruk Hydro Norway lorsk Hydro lorway Acartel 1984 Ardal og Sunndal Norway lorsk ilydro orvwa AD 81 1982 Petiticner SCAL Prance Pecbiney fruce AD 12 1983 Petitioner Aluminiu Pechiney fruce Pechiney franee Acartel 1984 Aluminiu Pechiney france Pechiney Ftarce Acartel 1984 Brandeis Goldschmidt A Co UK Pechiney Irance AD 11 1982 Petitioner Unidare Ireland Philips letherl'd Acartel 1984 Reynolds Aluminium GE Germny Reynolds Metals USA Acartel 1984 Rio Tinto Zinc U1 Rio Tinto Zinc UK AD 12 1983 DefeDdent Surina Aluminium Surina Surina Alumiini Surinu AD 11 1982 Petitioner VAN Leichtmetall Germay VIAG German AD 82 1983 Petitioner Vereinigte Aluinium Werke Germany VIAG Germany Acartel 1984 Vereignite iluinium Werke Germany VIAG Germny Acartel 1984 Vereinigte letallverke Austria Voest-Alpine Austria Sources: IC Official Journal, Who Owns Whom, Hopensted. OSn compilations. PRE Working Paper Series Contact LiCj Author Date forpaUer WPS373 Are Better-nff Households More Lawrence Haddad March 1990 J. Sweeney Unequal or Less Unequal? Ravi Kanbur 31021 WPS374 Two Sources of Bias in Standard Samuel Laird February 1990 J. Epps Partial Equilibrium Trade Models Alexander J. Yeats 33710 WPS375 Regional Disparities, Targeting, and Gaurav Datt March 1990 C. Spooner Poverty in India Martin Ravallion 30464 WPS376 The World Economy in the Colin I. Bradford, Jr. March 1990 C. Evangelista Mid-1990s: Alternative Patterns of 32645 Trade and Growth WPS377 After the Cold War: Security for John Stremlau April 1990 C. Evangelista Development 32645 WPS378 How Does the Debt Crisis Affect Patricio Arrau April 1990 S. King-Watson Investment and Growth? A Neoclassi- 31047 cal Growth Model Applied to Mexico WPS379 Some Implications of Policy Games Miguel A. Kiguel March 1990 R. Luz for High Inflation Economies Nissan Liviatan 39059 WPS380 Techniques for Railway Lee W. Huff March 1990 S. Shive Restructuring Louis S. Thompson 33761 WPS381 Trade in Banking Services: Alan Gelb March 1990 W. Pitayatona- Issues for Multilateral Negotiations Silvia Sagari karn 37666 WPS382 The Indonesian Vegetable Oils Donald F. Larsor March 1990 D. Gustafson Sector: Modeling the Impact of 33714 Policy Changes WPS383 On the Relevance of World Yair Mundlak March 1990 D. Gustafson Agricultural Prices Donald F. Larson 33714 WPS384 The Rational Expectations Christopher L. Gilbert April 1990 A. Daruwala Hypothesis in Models of Primary 33713 Commodity Prices WPS385 The Principles of Targeting Timothy Besley March 1990 J. Sweeney Ravi Kanbur 31021 WPS386 Argentina's Labor Markets in an Era Luis A. Riveros March 1990 R. Luz of Adjustment Carlos E. Sanchez 39059 WPS387 Productivity and Externalities: Jaime de Melo March 1990 M. Ameal Models of Export-Led Growth Sherman Robinson 37947 PRE Working Paper Series Contact ItilE Author Dte for a WPS388 The Distortionary Effects of Tariff Faezeh Foroutan March '1990 S. Fallon Exemptions in Argentina 38009 WPS389 Monetary Cooperation in the CFA Patrick Honohan March 1990 Wilai Pitayato- Zone nakarn 37666 WPS390 Price and Monetary Convergence Patrick Honohan March 1990 Wilai Pitayato- in Currency Unions: The Franc and nakarr. Rand Zones 37666 WPS391 Wealth Effects of Voluntary Debt Daniel Oks April 1990 S. King-Watson Reduction in Latin America 31047 WPS392 Institutional Development in World Samuel Paul April 1990 E. Madrona Bank Projects: A Cross-Sectional 37489 Review WPS393 Debt-for-Nature Swaps Michael Occhiolini March 1990 S. King-Watson 31047 WPS394 Threshold Effects in International Mark M. Spiegel April 1990 S. King-Watson Lending 31047 WPS395 How Gambians Save - and What Parker Shipton April 1990 C. Spooner Their Strategies Imply for International 30464 Aid WPS396 Strategic Trade Policy: How New? Max Corden How Sensible? WPS397 Antidumping Regulations or Patrick A. Messerlin April 1990 S. Torrijos Procartel Law? The EC Chemical S8-033 Cases WPS398 Agricultural Extension for Women Katrine A. Saito Farmers in Africa C. Jean Weidemann WPS399 Macroeconomic Adjustment, Andres Solimano April 1990 E. Khine Stabilization, and Growth in 30361 Reforming Socialist Economies: Analytical and Policy Issues WPS400 Macroeconomic Constraints for Andr6s Solimano April 1990 E. Khine Medium-Term Growth and Distribution: 39361 A Model for Chile WPS401 Policing Unfair Imports: The U.S. J. Michael Finger March 1990 N. Artis Example Tracy Murray 38010