1

MinimumEfficientScale:TowardsaMoreCoherentUnderstandingofSpecialResponsibilityandCompetitionontheMeritsunderArticle102TFEU

XingyuYanandHansVedder*

AbstractAstwokeyconceptsintheArticle102framework,competitiononthemeritsandspecialresponsibility suffer from a lack of concretization. As a result, the competition on themeritsconceptmayleadtheabuseassessmentintoanillusionofinherentimpropriety,whereasthespecialresponsibilityconceptissusceptibletoexpansiveinterpretationsthatunderminethepivotalroleofdominanceintheArticle102paradigm.Theseissuesareshownincasesconcerningexclusivedealingandexclusivityrebates.Thisarticlesuggestsusingtheasefficientcompetitorrationale—complementedbythenoeconomic sense rationale—to distinguish competition on the merits and abuse withrespect toexclusionary conduct.Thedistinction canbeexpressedas “efficiency-basedversusnon-efficiency-basedexclusions”.Nonetheless, the limitationsof theasefficientcompetitor test necessitate some calibrations. To that end and to give content to thespecialresponsibilityconcept,thisarticlespotlightsthenotionofminimumefficientscale,whichhasbeensomewhatoverlookedandmischaracterizedinthelawenforcement.ThisconceptdeservesmoreappreciationinproducingmorecoherenttheoriesofharmunderArticle102.

1. IntroductionThe“abuseofdominance”paradigmistheEU’schoiceofapproachtotackleunilateral(mainlyexclusionary)conduct thatharmscompetition in the internalmarket. It is laiddowninArticle102oftheTreatyontheFunctioningoftheEuropeanUnion(TFEU).1ThisArticlehasbeeninplacesincetheEuropeanEconomicCommunityTreaty(EECTreaty)enteredintoforcein1958.2Butwithitcamethedifficultyofformulatingacleardoctrine

*Yan:postdoctoralresearcheratGuanghuaLawSchool,ZhejiangUniversity,Hangzhou,China(email:xingyu.yan@outlook.com);Vedder:professorofeconomiclawatFacultyofLaw,UniversityofGroningen,Groningen,theNetherlands(email:h.h.b.vedder@rug.nl).1Article102oftheConsolidatedversionoftheTreatyontheFunctioningoftheEuropeanUnion,[October26,2012]OJC326,47–390(hereinafter,“theTFEU”).2Article86oftheTreatyestablishingtheEuropeanEconomicCommunity,effectiveon1January1958(hereinafter, “theEECTreaty”).This is theearliestversionof theTreatyprovisionthat isnowArticle102TFEU. ItwasrenumberedasArticle82 intheTreatyestablishingtheEuropeanCommunity(Consolidatedversion2002),[2002]OJC325,33–184(hereinafter,“theECTreaty”),andrenumberedagainasArticle102oftheTFEU.ThecontentofthisTreatyprovisionremainsthesame.Toavoidconfusion,thisarticleuses“Article86”(of theEECTreaty), “Article82”(of theECTreaty),andArticle102of theTFEUtorefertothesameTreatyprovisioninthedifferenttimeperiods.

2

onaccountofthevaguelystipulatedconceptstherein.3ThisexplainswhythisArticlewasnotactivateduntilContinentalCanin1971.4Supposedly,Article102condemnsabusiveconduct bydominantundertakings, rather than theirdominantstatusesas such.5Thisaccordswiththeeconomicideathatthepursuitortheholdingofadominantpositionassuchisacceptable,foritistheexpectedoutcomeofthecompetitiveprocess.6Nonetheless,thewordingofArticle102wasunclearaboutwhatqualifiesasabusiveandsounlawfulconduct.Theconceptofcompetitiononthemeritsisparadigmaticofthisambiguity.IntroducedbytheCommissionandadoptedbytheCourtofJusticeinannulmentcases,7thisconceptisthe antithesis of abuse, and together they form a dichotomy for explicating what isunlawfulunderArticle102.Thevirtueof thisdichotomylies in that itrespondstothequestionofhowtoprotectcompetitioninsteadofcompetitors,8orputdifferently,howshouldcompetitionlawdealwithmonopolizationbehaviorwhen“toachievemonopoly”iswhat thatdrives competition.9In that regard, thisdichotomyconcedes that,despitebeing prohibited from abusing its position, a dominant undertaking is still entitled toequalopportunitiesofcompetition,namelycompetitiononthemerits.Unfortunately, however, the line between abuse and competition on themerits is notentirelyclear,makingtheconceptofcompetitiononthemeritsostensiblyvacuous.10The

3SigfridoM.RamírezPérezandSebastianvandeScheur,“TheEvolutionoftheLawonArticles85and86EEC[Articles101and102TFEU],”inTheHistoricalFoundationsofEUCompetitionLaw,ed.KiranKlausPatelandHeikeSchweitzer(Oxford:OxfordUniversityPress, 2013), 35. See also, I. Samkalden and I. E. Druker, “Legal ProblemsRelating toArticle86oftheRomeTreaty,”CommonMarketLawReview3,no.2(1966):169.4TheCommissiondecidedinthiscasethatamergerbetweenContinentalandTDVhadinfringed Art 86 of the EEC Treaty, on the ground that such amerger eliminated thecompetitionstillremaininginthemarket.SeeCase6/72EuropemballageCorporationandContinentalCanCompanyInc.vCommissionoftheEuropeanCommunities[1973]ECR215,228(hereinafter,“theECJjudgmentonContinentalCan”).5Case C-209/10Post Danmark A/S v Konkurrencerådet, ECLI:EU:C:2012:172, para 21(hereinafter,“theECJrulingonPostDanmarkI”).6RichardA.Posner,AntitrustLaw,2nded.(Chicago:UniversityofChicagoPress,2001)248–249(“Themoreprotectionfromcompetitionthefirmthatsucceedsinobtainingamonopolywillenjoy,themorecompetitiontherewillbetobecomethatmonopolist;andprovidedthattheonlyfeasibleorpermittedmeansofobtainingthemonopolyaresociallyproductive,thiscompetitionmaybewhollydesirable.”)7SeeSection2.2.2.below.8RyanRStones,“TheChicagoSchoolandtheFormalRuleofLaw,”JournalofCompetitionLaw&Economics14,no.4(2018):539(describingtheChicagoanargumentsthatmarketconcentrationcouldbetheresultofimprovedefficiencyandthattocondemnbigfirmonaccountoftheconcentratedmarketstructureistosuggestthatthey“shouldcompetebutshouldnotwin”).9SeeSection3.1.10 OECD, “Competition on the Merits,” 2005,http://www.oecd.org/daf/competition/abuse/35911017.pdf (accessed January 27,2020)17 (observing that this conceptwaswidelyused “inamanner thatpresumesacommonunderstandingofwhat thephrasemeans”and thusservedasa “shortcut” to

3

term“competitiononthemerits”assumesaprioriametaphysicaldistinctionbetweennormal business practices and abnormal ones. Taken at face value, this distinctionborders on the passing of normative judgments as to what is rightful and wrongfulbusiness conduct. 11 This could be problematic, as it may divert us from identifyingcompetitive harms and to an “inherent impropriety” illusion. In other words, thecompetitiononthemeritsconceptrunstheriskofcontradictingthepragmaticnatureofcompetition law,which is a legal instrument toachieve certain social andpolicygoalsratherthantoprotectnaturalrights.12Thetendencyofresortingtonormativejudgmentsappearsstrong,asweshallsee in thevariousattemptstoconstructabaselinetest fordelineatingcompetitiononthemeritsandabuse.Besides,thereisthepracticalchallengethat conduct may have both pro- and anti-competitive effects, thus making the saiddichotomynotstraightforwardlyapplicable.13Consequently,riftscouldeasilyariseintheapproachestofindingabuse,asexemplifiedbytheIntelcase.To further complicate the situation, we have another vague concept: the specialresponsibility of a dominant undertaking. It imposes on a dominant undertaking theobligationnot toallow its conduct to impairgenuineundistorted competition14in themarketwherecompetitionisalreadyweakenedbecauseofthepresenceofthedominantfirm.15 This concept treads a thin line between condemning dominance as such andpreserving competition on the part of dominant firms. It singles out dominant firms,suggestingthatcertainconductthatislawfulwhencarriedoutbynon-dominantfirmsmaybeunlawfulwhencarriedoutbydominantfirms.16Theunderlyingrationalecanbeexpressed as follows: “Article 102, therefore, is not a prohibition of socially harmful

competitive assessments); EinerElhauge, “DefiningBetterMonopolization Standards,”StanfordLawReview56,no.2(2003):255(“Vacuousstandards,incontrast,areutterlyconclusory, failing to identify a coherent norm that provides any real help indistinguishing bad behavior from good or even in know which way certain factualconclusionscut.”);DamienGeradin,“TheUncertaintiesCreatedbyRelyingontheVague‘Competition on the Merits’ Standard in the Pharmaceutical Sector: The ItalianPfizer/PharmaciaCase,”JournalofEuropeanCompetitionLaw&Practice5,no.6(2014):347;HerbertHovenkamp,“ExclusionandtheShermanAct,”TheUniversityofChicagoLawReview72,no.1(2005):149.11Elhauge, “DefiningBetterMonopolization Standards,” 265, 266, 272 (discussing theuncertaintiescausedbysuchnormativejudgments).12 Tim Wu, “Intellectual Property Experimentalism By Way of Competition Law,”CompetitionPolicyInternational9,no.2(2013):30–31.13Elhauge,“DefiningBetterMonopolizationStandards,”263(posingaseriesofquestionsrelating to the difficulty of implementing a similar dichotomy established by the USSupremeCourtinGrinnell);Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,53 (“Different forms of conduct that, in the abstract, could all produce the same anti-competitive result may in reality be more or less likely to have an anti-competitivepurposeandmoreorlesslikelytobecompetitiononthemerits.”).14Case322/81NVNederlandscheBandenIndustrieMichelinvCommissionoftheEuropeanCommunities[1983]ECR3461,para57(hereinafter,“theECJjudgmentonMichelinI”).15Case85/76Hoffmann-LaRoche&Co.AGvCommissionof theEuropeanCommunities[1979]ECR461,para91(hereinafter,“theECJjudgmentonHoffmann-LaRoche”).16EkaterinaRousseva,RethinkingExclusionaryAbuses inEUCompetitionLaw (Oxford:HartPublishing,2010)71–72.

4

conductthatappliesonlytodominantundertakingsbutisaprohibitionofconductthatissociallyundesirablebecauseitiscarriedoutbydominantundertakings.”17However,theproblem is that the lack of concretization makes the special responsibility conceptsusceptibletoexpansiveinterpretationsandpotentiallyinbreachofthesaidrationale.ThisisdiscussedinSection2.1.1.There are excellent scholarlyworks that reflect on the conceptof abuse—which is inimplicit relation to competition on the merits—against the backdrop of a moreeconomics-basedapproachtoArticle102.18Nonetheless,notenoughstudieshavebeendone to explore a way forward that accounts for the trade-off between enforcementprecisionandenforcementcosts.Thisarticleisaneffortinthatregard.Itasks(1)howtheconceptsofspecialresponsibilityandcompetitiononthemeritscanbeconcretizedmorecoherently for the sakeof enforcementprecision, and (2)along this line,whether theeconomic concept of minimum efficient scale (MES) can help ameliorate the cost-effectivenessoftheArticle102enforcement.Itsearchesforanswersthroughsystematiccaselaw-analysisandcriticalliteraturereview.The remainder of this article is structured as follows. Part II situates the concepts ofspecialresponsibilityandcompetitiononthemeritsintheArticle102legalframework.Part III examines the solidity of an efficiency-based conception of competition on themeritsandwhytheasefficientcompetitor(AEC)rationaleisindispensableinthatregard.PartIVdiscusseshowthenotionofMEScanbeofuseinresolvingthelimitationsoftheAECrationaleandconcretizingthespecialresponsibilityconcept.PartVconcludes.Twoclarifications of the scope should be made. First, this article focuses on exclusionaryconduct, which is the main subject of Article 102 enforcement. Other enforcementsubjects include exploitative conduct and (by a strict distinction) discriminatoryconduct.19Second,thisarticletermsthesaidmainsubjectmorespecificallyas“unilateralexclusionaryconduct”,whichincludesagreements(suchasexclusivedealingagreements)and unilateral practices in the strict sense (such as refusal to deal). In that sense, it

17Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,178.18Seeforexample,PinarAkman,TheConceptofAbuseinEUCompetitionLaw:LawandEconomicApproaches(Oxford:HartPublishing,2012);RenatoNazzini,TheFoundationsofEuropeanUnionCompetitionLaw:TheObjectiveandPrinciplesofArticle102 (OxfordUniversity Press, 2011); Pablo Ibáñez Colomo, “Beyond the ‘More Economics-BasedApproach’: A Legal Perspective onArticle 102TFEUCase Law,”CommonMarket LawReview53,no.3(2016):709–40;Pablo IbáñezColomo,TheShapingofEUCompetitionLaw(Cambridge:CambridgeUniversityPress,2018)152;JohnTempleLang,“ComparingMicrosoftandGoogle:TheConceptofExclusionaryAbuse,”WorldCompetition39,no.1(2016):5–28;ThomasEilmansberger,“HowtoDistinguishGoodfromBadCompetitionunder Article 82 EC: In Search of Clearer and More Coherent Standards for Anti-CompetitiveAbuses,”CommonMarketLawReview42,no.1(2005):129–77.19 Nazzini, The Foundations of European Union Competition Law, 51 (adopting “thetraditionaltaxonomywherebyabusesaredistinguishedasexclusionary,exploitative,ordiscriminatory”andexplaining that “thedifference lies in theway inwhich long-termsocialwelfaremaybeharmedbytheunilateralconductoffirms”).

5

acknowledgesbutdoesnotaddressthepossibleoverlapbetweenArticles101and102onverticalcompetitiverestraints.20

2. SituatingcompetitiononthemeritsThissectionexamineswhatcompetitiononthemeritsmeansinthecontextofArticle102.Theoretically speaking, Article 102 is a paradigm that consists of two parts: thedominancepremiseandabuseassessments.Thesetwopartsareinextricablylinked,andthatholdsthekeytounderstandingcompetitiononthemerits.

2.1. ThedominancepremiseFindingadominantpositionisaprerequisitefordeclaringconductinviolationofArticle102.Traditionally, the lawenforcersrelyheavilyonstructuralcriteriasuchasmarketsharesandconcentrationlevels.Asthedigitaleconomygrowsandourantitrustlearningadvances,theyarebecomingmorereceptiveofcriteriathatshownuanceinlightofthechanging economic reality. However, the Article 102 legal framework develops in acommon-lawmanner,andthatsuggeststheinstrumentalityofcertainlong-establishedlegalprinciples.Thissubsectionlooksattwoofthose:thespecialresponsibilityconceptandthelinkbetweendominanceandabuse.

2.1.1. ThespecialresponsibilityconceptHowthecase lawconstrueddominanceshapedthe legalapproachtotacklingabusiveconduct.21Dominanceisconstruedinawaythataccentuatesanundertaking’sabilitytoforeclosure competition and thus entails from the outset a vigilant view of theundertakinginquestion.AccordingtothedefinitionofdominanceinUnitedBrands22andthedefinitionofabuseinHoffmann-LaRoche,23theCourtofJusticeconstrueddominanceas“apositionofeconomicstrength”,whichisanindicatorthatthemarketcompetitionhasalreadybeenweakenedandisapotentialriskoffurtherweakeningthecompetition.In that sense, holding a dominant position means having the ability to foreclosecompetition,andanyconductusedtoactualizethatabilitywouldbeidentifiedasabuseandconsideredunlawful.24Alongthisline,onecansaythatdominanceisneverperceivedasneutral;rather,thelawconsidersthatholdingadominantposition,althoughnotyetreachingtothepointofbeingcondemnableassuch,entailscertainbehavioralrestraintsonthedominantundertaking.

20ForapropositionthatseparatesthesetwoandputstheformerundertheregulationofArticle101,seeRousseva,RethinkingExclusionaryAbusesinEUCompetitionLaw,478.21Colomo,TheShapingofEUCompetitionLaw,153–154.22Case27/76UnitedBrandsCompanyandUnitedBrandsContinentaalBVvCommissionoftheEuropeanCommunities[1978]ECR207,para65;theECJjudgmentonHoffmann-LaRoche,para38.23TheECJjudgmentonHoffmann-LaRoche,para91.24Colomo,TheShapingofEUCompetitionLaw,153–154.

6

The Court of Justice elaborated the behavioral restraints as the special responsibilityconceptinMichelinI.RespondingtoMichelin’scontentionthatitwasbeingpunishedforhavingsuperiorperformance,itstatedthefollowing:

A finding that an undertaking has a dominant position is not in itself arecriminationbutsimplymeansthat,irrespectiveofthereasonsforwhichithassuchadominantposition,theundertakingconcernedhasaspecialresponsibility not to allow its conduct to impair genuine undistortedcompetitiononthecommonmarket.25

Supposedly,thisspecialresponsibilityisthelegalimplicationoffindingdominance,butambiguityarisesastowhatitentailsexactly.Thisambiguitymakesthisconceptsubjecttoexpansiveinterpretations,thusservingasashortcutforfindingabuse,especiallywhenthemarketpowerofthedominantfirminquestionisnotsogreatcomparedtothoseofnon-dominant firms. 26 When this happens, the assessment focus shifts away from“whethertheconductisharmfulintheparticulardominancecontext”andto“whethertheconductcanbeharmfulingeneralandwhetherthereisdominanceinparticular”.27Thisis exemplified by the presumptive illegality treatment of exclusive dealing (includingexclusivityrebates)intheArticle102caselaw.28

2.1.2. Thenon-causallinkbetweendominanceandabuseTo this point, the law has made clear that dominance is, by definition, the ability toforeclosecompetition,whichentailsbehavioralrestraintsontheundertakinginquestion,andthatanabuseofdominanceisprohibitedforactualizingthatability.Thislineoflogicgivesrisetothequestionofwhetherapracticehastobeanexerciseofthemarketpowerunderlyingthedominantpositionforittoconstituteanabuse.WhenprescribingtheobjectiveconceptionofabuseinHoffmann-LaRoche,theCourtofJusticeheldthatnocausationisneededbetweendominanceandabuse.29TheAdvocateGeneraltothiscasegavethesameopinion.30This“nocausallinkneeded”standpointis

25Case322/81NVNederlandscheBandenIndustrieMichelinvCommissionoftheEuropeanCommunities[1983]ECR3461,para57(hereinafter,“theECJjudgmentonMichelinI”).26Rousseva,RethinkingExclusionaryAbusesinEUCompetitionLaw,72.27ibid.(“Asaresult,thespecialresponsibilityisinvokedbytheCommunityCourtsmostoftenwhentheyfearthatadominantundertaking’sconductmayharmcompetitors,butatthesametimetheylackasufficientexplanationorevidencetoestablishtrulyharmfuleffectsoncompetition.”).28AlisonJones,“DistinguishingLegitimatePriceCompetitionfromUnlawfulExclusionaryBehaviour:ReconcilingandRationalisingtheCase-Law,”inCompetitionLawChallengesintheNextDecade,ed.SofiaOliveiraPais(PeterLangPublishing,2016),150–153.29TheECJjudgmentonHoffmann-LaRoche,para91(“theinterpretationsuggestedbytheapplicant that an abuse implies that the use of the economic power bestowed by adominantposition is themeanswherebytheabusehasbeenbroughtaboutcannotbeaccepted”).30OpinionofMrAdvocateGeneralReischl (deliveredonSeptember19,1978) inCase85/76Hoffmann-LaRoche&Co.AGvCommissionoftheEuropeanCommunities[1979]ECR

7

consistent with the Continental Can ruling that activated Article 102. There, theCommission considered an acquisition by a dominant undertaking of a competitor anabuse of dominance because this acquisition would entail a scale of competitionrestriction thatnormal concentrationswouldnotentail.31TheCourtof Justiceagreed,holding that a practice strengthening the dominant position needed not to beobjectionableasamannerofexercisingmarketpowertotriggertheArticle86prohibition,if “thedegreeofdominancereachedsubstantially fetterscompetition”.32TheAdvocateGeneraltothiscaseadvisedotherwise,buttheCourtofJusticedidnotfollow.33This“nocausallinkneeded”standpointdevelopedmorenuancesinTetraPak.There,theCourt of Justice stated that generally, “application of Article 86 presupposes a linkbetween the dominant position and the alleged abusive conduct”, but in “specialcircumstances”(suchastheonesinthiscasewheretheallegedabusewascarriedoutinamarketdifferent fromthedominatedone), this linkrequirementcouldbe lifted.34Atfirstsight,thismayseemlikeacontradictiontoHoffmann-LaRoche,butarguablyitisnot:inTetraPak,the“link”thattheCourtofJusticewasreferringtocanbeunderstoodasadimensionalone(whichsupposestheabuseandthedominancetobeinthesamemarket),ratherthanacausalone.Intheabsenceoffollow-upclarificationsfromtheCourtofJustice,thereisnoreasontodoubtthevalidityofthe“nocausallinkneeded”rulinginHoffmann-LaRoche.Inthatsense,TetraPakcomplementsHoffmann-LaRochebyexplainingthatthewaivingofacausallinkdoesnotprecludetheneedforalinkofadifferentkindbetweendominanceandabuse.35This“nocausallinkneeded”standpointmayappearoddtheoretically.36Whatputsitintoperspectiveisthe“abilitytoforeclose”definitionofdominance:giventhatthepresenceof adominantposition is alreadya signofweakenedmarket competition, any furtherweakeningonaccountofthedominantundertaking—causaltothedominantpositionornot—would be unacceptable. In that sense, it is not surprising that this standpointeffectivelywidened the application scopeof Article 102 to include almost all kinds ofbusiness practices. 37 This widened scope of application ensures that the vigilance

461,583(“thecriterionisnottheexerciseofmarketpowerbutthatthereisabusewhereanundertakinginadominantpositioninfluencesthestructureofcompetitionbyitsacts”).31TheECJjudgmentonContinentalCan,229.32ibid.,paras26–27.33OpinionofMrAdvocateGeneralRoemer (deliveredonNovember21,1972) inCase6/72EuropemballageCorporationandContinentalCanCompanyInc.vCommissionoftheEuropeanCommunities[1973]ECR215,254(suggestingthat,accordingtothewordingofArticle86,“itsapplicationcanbeconsideredonlyifthepositiononthemarketisusedasaninstrumentandisusedinanobjectionablemanner”).34CaseC-333/94PTetraPakInternationalSAvCommissionoftheEuropeanCommunities[1996]ECRI5951,para27.35Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,177–178.36 Elhauge, “Defining Better Monopolization Standards,” 331–333 (arguing for thepresenceofsuchalinkintheUSantitrustlaw).37 Colomo, The Shaping of EU Competition Law, 157. See also, Rousseva, RethinkingExclusionaryAbusesinEUCompetitionLaw,73–74(understanding“negativeeffect”asaconstitutiveelementofabuseandlinkingdominancetothatelement).

8

underlyingthisArticleisuncompromised.Tobeclear,however,theabusesdoappeartobecausallylinkedtodominanceinmostcases.38Theimplicationisthatbothdominantandnon-dominantfirmsarecapableofcarryingoutaparticulartypeofpracticethatpotentiallyconstitutesabuseinindividualcases.39ButonlydominantfirmscarryingoutsuchapracticewouldinfringeArticle102.Inthatlight, itwould be nonsensical to say that theabusiveness derives inherently from theconduct type since the conduct is perfectly lawfulwhen carriedout by non-dominantundertakings. A more reliable answer would be that the abusiveness flows from themarketpowerunderlyingthedominantposition.40Buttheproblemisthatinparticularcircumstances,themarketpowerofadominantfirmmaynotbequalitativelydifferentfrom that of a non-dominant firm.41So it becomes the law enforcer’s choice to eitherresort to the assessment shortcut provided by the legal framework or devise amorecoherentbenchmark.

2.2. AbuseassessmentsAfterthefindingofadominantposition,thelawapplicationmovesontotheassessmentoftheallegedlyabusiveconductathand.Thissubsectionexaminestwoissuesessentialtoabuseassessments:theeffectsthresholdandtheconceptualdelineationofabusefromlawfulconduct.Twopremisesunderpintheexaminationinthissubsection:

- First,underthedirectlegalobjectiveunderlyingArticle102—theprotectionofthecompetitiveprocess,42whatqualifiesacontentiouspracticeasanunlawfulabuseisessentiallythecompetitiveharmsitgenerates.Tobesure,itisanopenquestionastowhatlegalintereststhecompetitiveharmconceptshouldreflect.43

- Second,althoughadivergencemaystillexistbetweena“form-based”approachand an “effects-based” approach to the case-by-case assessment of competitiveharms,44it is indisputable that competitive harms need to be in place—either

38RobertO’DonoghueandA.JorgePadilla,TheLawandEconomicsofArticle102TFEU,2nded.(Oxford,UnitedKingdom:HartPublishing,2013)264–265.39Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,177.40ibid.,178.41Rousseva,RethinkingExclusionaryAbusesinEUCompetitionLaw,72.42ThisdirectobjectiveofEUcompetitionlawcanbeidentifiedfromaliteralreadingofProtocol27oftheTFEU,theprecedingversionofwhichwasArticle3(f)oftheEECTreaty.Seealso,MauriceE.Stucke,“WhatIsCompetition?,”inTheGoalsofCompetitionLaw,ed.DanielZimmer(Cheltenham:EdwardElgar,2012),29.43 Anne C Witt, The More Economic Approach to EU Antitrust Law, Hart Studies inCompetition Law, volume 14 (Oxford, United Kingdom: Hart Publishing, 2016) 110(describingthatalthoughonegets“thegeneralideathattheinvestigatedconductneedstoharmcompetition,oratleastbelikelytodoso”,itisstillquestionableastowhatisharmtocompetition).44ViktorJ.Vanberg,“ConsumerWelfare,TotalWelfareandEconomicFreedom–OntheNormativeFoundationsofCompetitionPolicy,” inCompetitionPolicyandtheEconomicApproach:FoundationsandLimitations,ed.JosefDrexl,WolfgangKerber,andRupprechtPodszun (Cheltenham: Edward Elgar Publishing, 2011), 62–63 (pointing out that thechoicebetweena form-basedandaneffects-basedapproachisnoteasytomakewhencomplicated by other issues, such as the fact that “competition agencies and their

9

presumed or verified—in a particular case before anArticle102 prohibition istriggered.45Therefore, the concept of competitive harm can be paraphrased as“anticompetitiveeffects”—atleastforassessmentpurposes.

2.2.1. The“potentialeffects”threshold

The case law consistently held that potential effects, as opposed to actual effects, areenoughforestablishinganabuseofdominance.46Toalargeextent,expressionssuchas“thepotentialof”,“thelikelihoodof”,and“thecapabilityofhavingsucheffects”areusedinterchangeably in the case law. 47 Also, the objective or intention of causinganticompetitiveeffectsisoftenequatedtothepotentialofhavingsucheffects.48Thisismorecontroversial.49Like the “no causal link needed” standpoint, the potential effect threshold is alsoattributable to the “ability to foreclose” definition of dominance: the presence ofdominance is already a sign of weakened competitive structure and any further

economicadvisorsarenotperfectandsubject toerrorsintheirattemptstoassesstheoverallwelfareconsequencesinparticularinstances”).Seealso,LizaLovdahlGormsen,“AreAnti-CompetitiveEffectsNecessaryforanAnalysisunderArticle102TFEU?,”WorldCompetition36,no.2(2013):224(describingthedebateinEUcompetitionlawregardingthe extent of anticompetitive effects that should be demonstrated when assessingexclusionarypractices).45 Witt, More Economic Approach, 271 (“Restriction of competition by object andrestrictionofcompetitionbyeffectdonotrefertodifferenttypesofcompetitiveharm.The distinction relates to different tests and a different standard ofproof.”). See also,O’DonoghueandPadilla,TheLawandEconomicsofArticle102TFEU,269.46CaseT-228/97IrishSugarplcvCommissionoftheEuropeanCommunities[1999]ECRII2969, para 191 (hereinafter, “the CFI judgment on Irish Sugar”); Case T-203/01Manufacture française des pneumatiques Michelin v Commission of the EuropeanCommunities [2003] ECR II 4071, paras 239, 241 (hereinafter, “the CFI judgment onMichelinII”);CaseT-201/04MicrosoftCorp.vCommissionoftheEuropeanCommunities[2007] ECR II 3601, paras 561, 563; Case C-52/09 Konkurrensverket v TeliaSoneraSverige AB [2011] ECR I 527, para 64; Case C-23/14 Post Danmark A/S vKonkurrencerådet,ECLI:EU:C:2015:651,paras64–66(hereinafter,“theECJrulingonPostDanmarkII”).47See, for example, the CFI judgment on Irish Sugar, para 191; the CFI judgment onMichelin II,paras239,241; theECJrulingonPostDanmarkII,para64;CaseT-219/99BritishAirwaysplcvCommissionoftheEuropeanCommunities[2003]ECRII5917,para293(hereinafter,theCFIjudgmentonBritishAirways);CaseC-549/10PTomraSystemsASAandOthersvEuropeanCommission,ECLI:EU:C:2012:221,para79.48TheCFIjudgmentonMichelinII,para241(“Ifitisshownthattheobjectpursuedbytheconductofanundertakinginadominantpositionistolimitcompetition,thatconductwillalso be liable to have such an effect.”). See also Case T-340/03 France Télécom SA vCommissionoftheEuropeanCommunities[2007]ECRII107,para195.49 For a critique on the equation of anticompetitive intention and (potential)anticompetitiveeffects,seeO’DonoghueandPadilla,TheLawandEconomicsofArticle102TFEU,267.

10

weakening would be unacceptable, so a dominant undertaking bears the specialresponsibilityofnotlettingthathappenonitsaccount.50The“potentialeffects” thresholdbringstwopotentialproblems.First, itmay leavethecausation between the contentious practice and the identified exclusionary effectsunascertained, in thesensethat theexclusionofcompetitorscould justbethenaturaloutcome of the competitive process. 51 This point was exemplified in the Microsoftdecision.TheCommissiondismissedMicrosoft’sdefensethat thesuccessofWMP(thetiedproductintheallegedlyabusivetie-in)wastheresultofcompetitiononthemerits.52Itdidsobypresumingthecausallinkbetweenthetie-inandthedeclineofRealNetworks(a competingmedia player product).53Admittedly, the Commissionmanaged to drawfromfactualfindingsthatMicrosoft’sdefensewasunsubstantiated,54butinthesamevein,it could also be said that these facts did not conclusively support the Commission’spresumption.Second,attachedtothispotentialthresholdisacounterfactualpremisethatiscapableofarbitrarilydismissingpositiveevidenceof competition remaining.55TheGeneralCourtruled that evidence of competition remaining in the market cannot cancel theanticompetitiveness of the conduct in question since it is not for the dominantundertakingtodictatewhatlevelofcompetitionshouldexist.56Inthepresenceofsuchevidence, it ruled that the level of competition could have been even better in thealternativescenariowheretheaccusedconductdidnottakeplace.57

2.2.2. ThecompetitiononthemeritsconceptAKZOThenotionofcompetitiononthemeritsemergedinAKZOastheantithesisofabuse.58Butthe linebetweenthesetwowasunclear.Byaddingvaguelydefinedconditionssuchas“unfair”, “intended”, and “smaller competitors” to the rule of prohibition, 59 the

50Note46above.51Colomo,TheShapingofEUCompetitionLaw,154.52CommissionDecisionof24March2004relatingtoaproceedingunderArticle82oftheECTreaty(CaseCOMP/C-3/37.792Microsoft),recitals947–48.53ibid.,recital948.54ibid.,recitals949–51.55O’DonoghueandPadilla,TheLawandEconomicsofArticle102TFEU,268(pointingoutthecircularlogicinusingthecounterfactualscenariotodismissthefactthatsometimesthecompetitorsareactuallythriving).56CaseT-286/09IntelCorp.vEuropeanCommission,ECLI:EU:T:2014:547,paras117,124(hereinafter,“thereferred-backGCjudgmentonIntel”).57TheCFIjudgmentonBritishAirways,para298.58 85/609/EEC: Commission Decision of 14 December 1985 relating to a proceedingunderArticle86oftheEECTreaty(IV/30.698-ECS/AKZO),[December31,1985]OJL374,para81(“TheCommissionemphasizesthatitdoesnotconsideranintentionevenbya dominant firm to prevail over its rivals as unlawful. A dominant firm is entitled tocompeteonthemerits.”).59ibid.,para74.

11

Commissionadvancedcertainequityconcernsthatwerelatentinthelaw.60Theproblemisthattheseconcernsobscured,insteadofclarifying,howtheforeclosureofcompetition(asopposedtocompetitors)wouldtranspireandhowcompetitiononthemeritswastoberecognized.Inotherwords,itseemsthat,whenadoptingthisconcept,theCommissionwas not prepared to fully embrace the ensued implications. This is exemplified by itsleavingopenthequestionofwhetherlessefficientcompetitorsshouldbeprotected.Intheappeal,theCourtofJusticeapprovedthenotionofcompetitiononthemeritsastheantithesis of abuse. 61 But it also sidestepped the question of whether the targetedcompetitorsdeservedtobeexcludedforhavinginferiorperformance.Itonlyspeculatedthatthoseexcludedcompetitors“areperhapsasefficientasthedominantundertaking”.62Thiscontradictedthestandpointofencouraging“competitiononthebasisofquality”andcouldonlyberationalizedunderthemarketintegrationmandateoftheEEC:structuralpreservation of competition (which is likely to result in protecting less efficientcompetitors)servedasanexpedienttorectifycertaintrade-restrainingpracticesandanintermediarysteptofostercompetitioninthestageofestablishingacommonmarket.AstraZenecaAstraZenecahostedamoreextensivediscussionon this concept.63BasedonAKZO, theGeneral Court held in this case thatArticle 82 of the ECTreaty prohibits a dominantundertaking from employing non-merits-based methods to compete. 64 The Court ofJustice approved this standpoint. 65 More importantly, the General Court linked

60ibid.,para77.61Case62/86AKZOChemieBVvCommissionoftheEuropeanCommunities[1991]ECRI3359, para 70 (“Article 86 prohibits a dominant undertaking from eliminating acompetitor and thereby strengthening its position by usingmethodsother than thosewhichcomewithinthescopeofcompetitiononthebasisofquality.”)62ibid.,para72.63Thecaseinvolvestwoabusesofdominance:(1)makingmisleadingrepresentationstonational patent agencies so as to obtain Supplementary Protection Certificates that,withoutthemisleadingrepresentations,wouldhavebeenobtainedforashorterperiodofvalidityorwouldnothavebeenobtainedatall;(2)astrategyincludingtheselectivederegistration ofmarket authorizations and the switch from capsules to tablet of themedical product in question. See Commission Decision of 15 June 2005 relating to aproceedingunderArticle82oftheECTreatyandArticle54oftheEEAAgreement(CaseCOMP/A.37.507/F3–AstraZeneca),recitals144,255,788,(hereinafter,“theCommissiondecisiononAstraZeneca”).64CaseT-321/05AstraZenecaABandAstraZenecaplcvEuropeanCommission[2010]ECRII2805,para354(hereinafter,“theGCjudgmentonAstraZeneca”).65 Case C-457/10 P AstraZeneca AB and AstraZeneca plc v European Commission,ECLI:EU:C:2012:770,para75(hereinafter,“theECJjudgmentonAstraZeneca”).

12

“competitiononthemerits”tothespecialresponsibilityconcept.66TheCourtofJusticeapprovedthisaswell.67Whenitcametodeterminingwhetherapracticefallsoutsidethescopeofcompetitiononthemerits,theGeneralCourtheldthatintentcouldbearelevantcriterion,althoughnotanecessary condition for finding abuse. This follows the objective conception of abuseestablished in Hoffmann-La Roche. 68 The Court of Justice approved this and furtherclarifiedthat,whenassessingwhethermisusesofthepatentprocessareoutsidethescopeof competition on the merits, a contextualized approach should be taken, instead ofpresuming all fallible practices in regulatory proceedings condemnable under Article82.69Inthatsense,theECJseemstohavesuggestedaconsiderationoftheproportionalityprincipleforbalancing“competitionpreservation”and“respectinganundertaking’srighttodefenditslegitimatebusinessinterests”.PostDanmarkIThedistinctionbetweencompetitiononthemeritsandabusebecamemuchclearer inPostDanmarkI.TheCourtofJusticeclarifiedthatdominationormonopolization“onitsown merits” is not condemnable, because it could be the natural process andmanifestationofcompetitiononthemerits,whichistobeencouragedunderArticle82ofthe EC Treaty. 70 It distilled this point from Michelin I, Compagnie maritime belgetransports,andTeliaSonera.Accordingly,theCourtofJusticeruledthatArticle82protectscompetitioninsteadoflessefficientcompetitors,astheirbeingexcludedistheexpectedoutcomeofcompetitiononthemerits.71Tothatpoint,theCourtofJusticeadvancedtwoaspectsofconsiderationforfindingabuseinthe“allcircumstances”analyticalframework:

- The“methods”dimensionrequiresthat,totriggertheprohibitionofArticle82,themethodsemployedshouldnotbe“partofcompetitiononthemerits”;

- The“effects”dimensionrequiresthatthereshouldbeexclusionaryeffectsonas-efficientcompetitors.72

Critically, thesetwoaspectsofconsiderationshouldnotbeunderstoodrespectivelyas“the forms” and “the competitive effects” of the conduct in question and so as twocumulativeconditions for findingabuse.Rather, theyshouldbeunderstoodas thetwoaspectsdescribingthesamething:aqualifiedlevelofanticompetitiveeffectsthatleadstotheabusefinding.Thereasonisdiscussedearlier:thelawdoesnotrequireacausallink

66TheGCjudgmentonAstraZeneca,para671(“Article82ECimposesonanundertakingin a dominant position, irrespective of the reasons forwhich it has such a dominantposition,thespecialresponsibilitynottoimpair,byusingmethodsotherthanthosewhichcomewithinthescopeofcompetitiononthemerits,genuineundistortedcompetitioninthecommonmarket.”).67TheECJjudgmentonAstraZeneca,paras98,134,149.68TheGCjudgmentonAstraZeneca,para359.69TheECJjudgmentonAstraZeneca,para99.70TheECJrulingonPostDanmarkI,paras21–22.71ibid.,para22.72ibid.,paras25–26.

13

between dominance and abuse, and therefore both dominant and non-dominantundertakings can carry out a type of conduct that is potentially abusive. So, theabusivenessofconductessentiallyboilsdownto“effects”andhasnoinnateconnectionwiththeformsofconduct.Admittedly,conductformscouldberelevantinanalyses,butonly in the sense that they are proxies of anticompetitive effects. The dissociation ofabusivenessfromformsalsomeansthatthereisnoinherentimpropriety—atleastnotinthenon-effectssense—ofanabuse.IntelWhether“forms”shouldbeunderstoodasa(lessthanideal)proxyof“effects”wasatthebaseofthewholeInteldispute.Inthiscase,theGeneralCourtpresumedtheabusivenessof exclusivity rebates on the sole basis of the exclusivity element.73 To be sure, thispresumption was well supported by a string of case law starting fromHoffmann-LaRoche,74buttheGeneralCourtoverlookedthereasonwhytheprecedentssupportedthispresumption:itwasbecausetheexclusivityelementwasconsideredaworkableproxy(incombination with the dominance factor) for identifying the relevant anticompetitiveeffectsatatimewhennobetterapproacheswereavailable.Theexclusivityelementaloneisillusoryinthesensethatthereisnoinherentimproprietybehindit.Therefore,whentheCommissionperformedonitsinitiativeanAECanalysistoascertainthelevelofanticompetitiveeffectscausedbyIntel’srebateschemes,thesingle-mindedadherencetothepresumptionbytheGeneralCourtbecameratherirrelevant.TheCourtof Justice referred back the General Court’s judgment, precisely because of thisirrelevance.Thus,inthependingnewjudgment,theGeneralCourtisconfrontedbythetaskofre-examiningthesolidityofthepresumptionregardingexclusivity.GenericsGenericsisthenewestadditiontothecaselawthatclarifiestheconceptofcompetitiononthemerits.Inthispreliminaryruling,theCourtofJusticereiteratedthe“effects”thresholdand the “all circumstances” framework for findingabuse.75It implicitlyreferred to thedistinctionbetweenabuseandcompetitiononthemeritsbystatingthatadominantfirmhastheright todefend itscommercial interests,althoughthatmustbedonebytaking“reasonablesteps”.76Suchreasonablestepsinclude,interalia,theexerciseofanexclusiveintellectual property right that does not “fall outside the scope of competition on themerits”.77Whenitcametothedelineationofthescopeofcompetitiononthemeritsinthisparticularcase,theCourtofJusticeadvancedtheelementofintent,78but,tobeclear,intentisonlyoneofthemanyaspectsofconsiderationintheallcircumstancesframeworkandundertheobjectiveconceptionofabuse.79

73Thereferred-backGCjudgmentonIntel,para85.74ibid.,paras80–81.75 Case C-307/18 Generics (UK) Ltd and Others v Competition and Markets Authority,ECLI:EU:C:2020:52,para154(hereinafter,“theECJrulingonGenericsUK”).76ibid.,para149.77ibid.,paras150–52.78ibid.,paras151,162.79ibid.,paras148,154,169.

14

Inthatsense,theCourtofJusticepointedtotheprincipleofproportionality.Namely,itsuggested that, to distinguish “abuse” from “competition on the merits involving theexerciseofanintellectualpropertyright”,dueeffortsmustbemadetowardsaweighingofthepro-andanti-competitiveeffectsoftheconductinquestiontoascertainwhetherthe exercise of right went beyond its justifiable scope. 80 This corresponds to thestandpoint that “the economic and legal context” is instrumental for establishing aconcertedpracticeasa“restrictionbyobject”underArticle101.81TheCourtof Justicemadeitabundantlyclearinthiscasethatthe“context”—whichincludesbutisnotlimitedtotheconsiderationoftheallegedpro-competitiveeffects—playsanessentialpartintheestablishmentofbothrestrictionsbyobjectandrestrictionsbyeffect.82Byimplication,themere existence of pro-competitive effects does not necessarily save an agreementfrombeingqualifiedasarestrictionbyobject.83Accordingly,inthiscase,arestrictionbyobject is establishedwhen, put through a contextualizedweighing analysis, a patent-dispute settlement agreement has no plausible explanation other than the collusiverestrictionofcompetition.84To sum up this section, the EU case law points to a contextualized approach fordistinguishingcompetitiononthemeritsfromabuse.Inthetwo-tierArticle102paradigm,however,undertakingsuchacontextualizedapproachnecessitatesnotonlyconfrontingthepractical,evidentiarychallengesbutalsoreconcilingthefrictionsstemmingfromthe“ability to foreclose” conception of dominance. In that consideration, the next sectionexaminescompetitiononthemeritsinatheoreticalcontext.

3. Anefficiency-orientedconceptionof“competitiononthemerits”TheEU isnottheonlycompetition lawregimestrugglingtorationalizethedistinctionbetweencompetitiononthemeritsandabuseofdominance.Section2oftheUSShermanActdealswithasimilarissue.85AroundthetimeoftheEUcompetitionlawmodernization,therewasaheateddebateintheUSontheappropriatestandardforidentifyingunlawfulexclusionaryconductunderSection2.ThisSectionlooksatthescholarshipconcerningthisdebateforabetterunderstandingofthenormativemeaningofcompetitiononthemerits.Thethesisofthissectioncanbeoutlinedasfollows.First,underthepremisethatthelawpursuesnon-efficiency-basedexclusionaryconduct,theideaofinherentimproprietyhasnocontentandthebenchmarkingofcompetitiveharmsisinevitable.Second,weshouldforgo the idea of constructing a unifying abuse test because it is too costly from the

80ibid.,para167.81ibid.,para67.82ibid.,paras67,70,82,104.83ibid.,para106.84ibid.,paras87–90.85MarkS.Popofsky,“DefiningExclusionaryConduct:Section2,theRuleofReason,andtheUnifyingPrincipleUnderlyingAntitrustRules,”AntitrustLawJournal73,no.2(2006):438 (pointingout the lackof “a clearprincipleoroperational legal test thatseparateslawfulfromunlawfulconduct”underSection2oftheShermanAct).

15

viewpointofdecisiontheory.Whatweneedinsteadisacoherentwaytorationalizetheexistingabusetests.Inthatregard,weshouldalsoconsiderhowthepredictabilityoflawshelpsustainex-antetheincentivestoinnovate,andsoitmaybedesirabletodevisesomepresumptionsoflegalityandillegalityundertheexistingabusetests.

3.1. The“competingtomonopolize”process:thestaticanddynamicperspectivesInherent in the concept of competition on the merits is the logic that unilateralexclusionary conduct employed by a firm in the presence of business rivals can beconceptuallyclassifiedintotwokinds:

- non-merits-basedcompetitiveconduct,whichisprohibitedsupposedlybecauseitimpairs,asopposedtoenhancing,competition;

- merits-basedcompetitiveconduct,whichislawfulbecauseitisthemanifestationofefficiencyandenhancementofcompetition.86

Someclarificationsareneeded.Here, thisarticledescribesthedistinctionas“whetherconductharmscompetition”,butthisisonlyanintermediatedescription,inthesensethattherecanbemorefundamentalobjectivesatplay.87Forexample,onevalue-ladenwaytodescribe thedistinction is “whether conduct increasesefficiency andwelfare”.88Otherfundamentalconcernsincludefairnessand,mostrelevanttotheEU,marketintegration.89Underlyingthisconceptualdistinctionistheinternalfrictionofthecompetitiveprocess:“competitiontoobtainamonopoly isan important formofcompetition”.90Thiscanbeexplainedinthestaticmodelofpricetheory.Supposingthatfirmsaredrivenbyrationalprofit maximization and that the two poles of market performance are perfect

86RobertH. Bork,TheAntitrust Paradox: A Policy atWarwith Itself (NewYork: BasicBooks,1978)135(pointingoutthelaw’sdilemmaofnotbeingabletoseparatebeneficialcompetitive behavior from detrimental competitive behavior in practice); Elhauge,“DefiningBetterMonopolizationStandards,”264–65(describingtheuncertaintyinthisconceptualdistinction).87Note42above.88 Bork, The Antitrust Paradox, 116 (classifying business behavior into three typesaccording to their relationshipswith consumerwelfare: “primarilyefficiency creating,primarily output restricting, or neutral in its consumer welfare impact”), 137(distinguishing “efficiency exclusion” from “improper exclusion”, and suggesting “Theconceptualapparatusnowinuse is incapableofmakingthatcrucialdistinction.”),311(characterizingnon-merits-basedcompetitionasartificialbarriersthat“arenotformsofsuperiorefficiencyandwhichyetpreventtheforcesofthemarket—entryorthegrowthofsmallerfirmsalreadywithintheindustry—fromoperatingtoerodemarketpositionsnotbasedonefficiency.Caremustbetakentodistinguishbetweenformsofefficiencyandartificialbarriers.”).89EleanorM.Fox,“MonopolizationandDominanceintheUnitedStatesandtheEuropeanCommunity: Efficiency Opportunity and Fairness,”Notre Dame Law Review 61, no. 5(1986):1008;Claus-DieterEhlermann,“TheContributionofECCompetitionPolicytotheSingle Market,” Common Market Law Review 29, no. 2 (1992): 258–59; Witt, MoreEconomicApproach,114,144–45.90Posner,AntitrustLaw,248.

16

competition and monopoly, 91 all firms aspire to monopolize (by whatever means),becausemonopolyrewardsthemostopportunitiesforreapingprofitswhereasperfectcompetitiontheleast.92Butsincemonopolyisattainableonlybyonefirm,competitioncomes into being and intensifies as the monopoly rewards become greater. 93 This“competingtomonopolize”processinevitablyleadstotheexclusionofcertainrivals.94From a static viewpoint, rivalry-induced exclusions are inevitable. From a dynamicviewpoint, they are potentially efficiency-enhancing: when dynamic competition isfactoredin,asomewhatconcentratedmarketstructuremaybeanindicationofimprovedefficiency,95andinturn,itisperhapsmorelikelytostimulateinnovationthanahighlyfragmented one, considering the complexity of impact factors on innovation. 96 Thisdynamic viewpoint underscores the importance of preserving ex-ante dominantundertakings’ incentives to innovate by allowing them to reap their business successmanifestedasmarketpower, especially in scenarioswhere firmswouldonly invest ininnovationifthereweretheprospectofsupra-competitivereturns.97

91Bork,TheAntitrustParadox,119–21(discussingthevalidityoftheprofitmaximizationpremise in price theory); Lynne Pepall, Dan Richards, and George Norman, IndustrialOrganization: Contemporary Theory and Empirical Applications, 5th ed. (Hoboken, NJ:Wiley,2014)19.92Elhauge,“DefiningBetterMonopolizationStandards,”262–63(pointingoutthatfirmscompete in, forexample,productdevelopment “inordertoreap theprofits that comefromproducingaproductthatissufficientlysuperiortowhatrivalscanprovidethatitreapsmonopolyprofits”),298(discussinghowtheprospectofmonopolyprofitsdrivesfirmstoinnovate).93 Posner, Antitrust Law, 248. See also, Gregory J. Werden, “Network Effects andConditionsofEntry:Lessons from theMicrosoftCase,”AntitrustLaw Journal 69,no.1(2001):89–91(describingtheextremescenariowherepowerfulnetworkeffectsenablefirmstocompete“forthemarket”insteadof“inthemarket”).94A.DouglasMelamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct—Are There Unifying Principles?,” Antitrust Law Journal 73, no. 2 (2006): 382–383(describingthepowerfulintuitionthat“oneisentitledtocompeteonthemeritsandtoreapthefruitsofhissuccess”intheformofmonopolypower).95 Giovanni Dosi et al., “Learning, Market Selection and the Evolution of IndustrialStructures,”SmallBusinessEconomics7,no.6(1995):423(“moreefficientmarketstendtoyield,inevolutionaryenvironments,moreconcentratedmarketstructures,ratherthanmore‘perfect’onesinthestandardsense”).Seealso,J.G.SidakandD.J.Teece,“DynamicCompetitioninAntitrustLaw,”JournalofCompetitionLawandEconomics5,no.4(2009):612.96SidakandTeece,“DynamicCompetitioninAntitrustLaw,”587–89,592,600–01.Seealso,DavidBAudretsch,WilliamJBaumol,andAndrewEBurke,“CompetitionPolicyinDynamicMarkets,”InternationalJournalofIndustrialOrganization19,no.5(2001):626;Massimo Motta, Competition Policy: Theory and Practice (Cambridge: CambridgeUniversity Press, 2004) 55; Philippe Aghion et al., “Competition and Innovation: AnInverted-URelationship,”TheQuarterlyJournalofEconomics120,no.2(2005):720–21.97 Elhauge, “Defining Better Monopolization Standards,” 262–63, 298. See also,Hovenkamp,“ExclusionandtheShermanAct,”158(“willingnessto‘sacrifice’short-termprofitsinanticipationoflatermonopolyprofitsdoesnotdistinguishanticompetitivefromprocompetitive uses of innovation”); Janusz A. Ordover and Robert D. Willig, “An

17

Competition policy responds to these economic insightswith the following normativenotion: monopolists (and dominant firms) are entitled to equal opportunities ofcompetition and to reap the fruits of their success manifested as market power. 98Accordingly,competitionlawdoesnotconcernitselfwitheverypracticethatexcludes.Rather, it is intended to separate condemnable ones (abuse) from permissible ones(competitiononthemerits)throughabusetests.Thekeyquestionishow.

3.2. Theformalrequirementsofagoodtest

3.2.1. AdministrabilityandpredictabilityTwopointscanbedistilledfromtheprecedingsubsection:(1)exclusionaryeffectsareinherent in the competitive process, and (2) it is normatively necessary and sociallydesirabletosustaincompetitivebehaviorofdominantfirms.Theyprescribebothformaland substantive requirements on an abuse test. This subsection discusses the formalrequirements.Ontheonehand,servingasaroadmapfordecision-makers,anabusetestissetouttoreduce enforcement errors (false positives and false negatives) in individual cases. Inotherwords, it aspires to induce precise findingsof unlawful exclusionary conduct invarying case circumstances. 99 On the other hand, however, the level of precisionattainable in individual cases will be inherently limited considering the reality ofimperfect information.100 Besides, the pursuit of precision is costly and enforcementresourcesarescarce.Allocatingmoreresourcestoonecasemeansfewerresourcesforanother.Therefore,itwouldbeunsustainabletoindulgeinthepursuitofperfectprecisioninindividualcasesandgeneralizationwouldhavetobeintroducedatacertainpoint.101Thisrequiresanabusetesttobeadministrable,soastohelpthecourtsandenforcement

EconomicDefinitionofPredation:PricingandProductInnovation,”YaleLawJournal91,no.1 (1981):22 (finding that evengenuine innovations canbeanticompetitive in thesense that they may be driven solely by exclusionary objectives and do not provideconsumerbenefitsthatworththecost).98Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”379,382,394 (reiterating the “normative intuition that firmsought tobepermitted to reap thefruits of competition on the merits, even if less skillful or less lucky rivals aredisadvantaged”);Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,78.99Hovenkamp,“ExclusionandtheShermanAct,”148.100IsaacEhrlichandRichardA.Posner,“AnEconomicAnalysisofLegalRulemaking,”TheJournalofLegalStudies3,no.1(1974):268–270(discussingtheallocativeinefficienciesensuedfromthefactthatitisimpossibleforaruletofitperfectlywiththerealityitseekstoregulate).Seealso,Popofsky,“DefiningExclusionaryConduct,”450.101Anotherwaytodescribethis“generalization”elementis“categorization”,which“notonly specifies the substantive standards, but also determines (implicitly) whatinformationisrelevanttotheanalysisandinwhatsequencethatinformationshouldbeconsidered”.SeeC.FrederickBecknerandStevenC.Salop,“DecisionTheoryandAntitrustRules,”AntitrustLawJournal67,no.1(1999):42.

18

agenciesnavigatethroughtheinformationalconstraintsandpursueprecisionwithinanaffordablerangeofenforcementcosts.102A dynamic perspective advances another quality requirement: a test should bepredictableenoughforfirmstoassessexantetheirbusinesspractices.Onethatfailsthisrequirementwouldchillinnovationandunderminecompetitioninthelongrun.103Thisrequirementalsocallsforgeneralizationatacertainpoint,becausepursuingprecisionregardlessof theensuedcostwould causeunpredictabilityandundermine innovationincentives.104Onetestthatfailsthesetworequirementsisthewelfare-balancingtest.Thistestexamines“whethertheconductharmscompetitorsbyraisingtheircostsandwhetherthosehighercostsharmconsumersandcompetitionbyallowingthedefendanttoachieve,maintain,orenhancemonopolypower”.105Itwascritiquedasbeinginsufficientlypredictableandadministrable, making it difficult for firms to assess ex-ante their conduct and thusunderminingdynamicefficiencies.106Tobe fair, theconsumerwelfaretestdoesreflectsomepredictabilityconsiderations: itmayguideadecisionmaker tomake theoptimaldecision based on the limited available information, and when operationalized, thisdecision-making processwould be able to guide firms’ self-assessments.107But still, itdoesnotappearattractiveundertheadministrabilityrequirementandsoitsapplicationcanbecostly.

3.2.2. AunifyingabusetestiscostlyNotably,thereisadifferencebetween“thesomewhatgeneralizedroadmapguidancethatanabusetestoffers”and“thelevelofgeneralizationthatthetestitselfpossesses”.Theformeristhesubstanceofthetestwhereasthelatteristheform.Regardingforms,thereis thecommonlyreferenceddichotomybetweenrulesandstandards.108Anabusetest,functioningasaroadmap,ismorelikelytofallunderthelatterthantheformer.Therefore,

102 Melamed, “Exclusive Dealing Agreements and Other Exclusionary Conduct,” 383;Hovenkamp,“ExclusionandtheShermanAct,”148;BecknerandSalop,“DecisionTheoryandAntitrustRules,”45–46.103Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”381,383,393;StevenC.Salop,“ExclusionaryConduct,EffectonConsumers,andtheFlawedProfit-SacrificeStandard,”AntitrustLawJournal73(2006):349.104 Melamed, “Exclusive Dealing Agreements and Other Exclusionary Conduct,” 387(phrasingthereducedincentivesforinvestmentandinnovationas“thedynamiccostsofantitrust intervention”); Elhauge, “Defining Better Monopolization Standards,” 301(doubting from an ex-anteperspectivewhether a case-by-case analytical approach tobalancingpatentprotectionandcompetitionprotectionwouldbesociallydesirable).105Salop,“ExclusionaryConduct,”318(thequote),331.106 Gregory J. Werden, “Identifying Exclusionary Conduct under Section 2: The NoEconomicSenseTest,”AntitrustLawJournal73,no.2(2006):431–432.107Salop,“ExclusionaryConduct,”343–45,352.108EhrlichandPosner, “AnEconomicAnalysisofLegalRulemaking,”258(“Astandardindicatesthekindsofcircumstancesthatarerelevanttoadecisiononlegalityandisthusopen-ended.”;“Arulewithdrawsfromthedecisionmaker’sconsiderationoneormoreofthecircumstancesthatwouldberelevanttodecisionaccordingtoastandard.”).

19

the costs identified in decision theory for devising a precise rulemay not be entirelyapplicable to an abuse test. 109 But there is a degree of commonality. Particularly,applicable to both an abuse test and a precise rule are the two determinants offormulationcosts:howhomogeneoustheaddressedconducttypesareandhowfrequentthe test/rule encounters them.110The (marginal) costs for the pursuit of precision godownasthelevelofconducttypehomogeneityortheencounterfrequencygoesup.111In that sense, applying a unifying abuse test would be costly. First, the level ofhomogeneityacrossalltypesofunilateralexclusionaryconductislow.Somesubtypes,suchasexclusivedealingandloyaltyrebates,aresomewhathomogeneous,inthesensethattheirtheoriesofharmarearguablyanalogoustoeachother.Butothersubtypes,suchaspredatorypricingandrefusaltodeal,aremuchlessso.Second,thechanceforthelawtoencounterdifferentconducttypesvaries.Inthatregard,itisdifficulttogiveapreciseanswer regarding the chances because that would require an empirical study, but itshould be relatively easy tomake a general observation from an EU competition lawtextbookthatrecordsthecaseprecedentsunderArticle102.

3.3. Theillusorynotionofinherentimproprietyandthenoeconomicsensetest

3.3.1. ThemeaningofinherentimproprietyApopularsuggestionastowhatjustifiestheseparationofabusefromcompetitiononthemeritsistheinherentimpropriety—morespecifically,theinherentcompetitiveharms—ofabusiveconduct.Asaleadingfigureofthe“chastisedHarvardSchool”,112Areeda(1989)construed “impropriety” as “the conduct requirement—not just market power—necessary to constitute monopolization under Section 2 of the Sherman Act”.113 Theunderlying idea is that all unilateral business conduct excludes so there has to be aqualificationofsomekindfortheprohibitionofconduct.Asheexplained,

It is perfectly clear that the concept of monopolization requires someelement of impropriety; it has to bemonopoly power coupled notwithbuilding the bettermousetrap, butmonopoly power coupledwith someimproprietyinitsachievementormaintenance.114

109ibid.,267–71(identifyingthecostsofdevisingandimplementinghighlypreciserules).Seealso,Popofsky, “DefiningExclusionaryConduct,”458 (“At somepoint, the costsofcreatingamorerefinedrule(includinglitigationcostsinacommon-lawsystem)becomeprohibitivelyhigh.”).110Popofsky, “DefiningExclusionaryConduct,”458;EhrlichandPosner, “AnEconomicAnalysisofLegalRulemaking,”270.111 Popofsky, “Defining Exclusionary Conduct,” 437, 459–60; Ehrlich and Posner, “AnEconomicAnalysisofLegalRulemaking,”264,267.112HerbertHovenkamp,“TheHarvardandChicagoSchoolsandtheDominantFirm,”inHowtheChicagoSchoolOvershottheMark:TheEffectofConservativeEconomicAnalysisonUSAntitrust,ed.RobertPitofsky(Oxford:OxfordUniversityPress,2008),109–10.113 Phillip Areeda, “Essential Facilities: An Epithet in Need of Limiting Principles,”AntitrustLawJournal58(1989):845–47.114ibid.,846.

20

SomeoftheChicagoSchoolscholarsalsoembracedthe“inherentimpropriety”notion.InhisseminalTheAntitrustParadox,RobertBorkupheldChiefJusticeWhite’sstatementinAmerican Tobacco that antitrust law condemns certain practices—which were foundcontrary to “normal methods”—“either because of their inherent nature or effect orbecause of the evident purpose of the acts”. 115 He understood “normal methods” assuperiorefficiency.Heconsideredthe“inherentnature”ofabnormalitytobeatriggerofper se ruleof illegality.116He furthered suggested that such “inherence”dismissed theneedtoexamineandweightheactualeffectsofacontentiouspracticeandthatinferencefromotheraspectsoffactwouldsuffice.117WhatBorkfoundunsatisfactoryinJusticeWhite’sopinionswastheambiguityastowhat“abnormality” meant: all business practices exclude, so it would be impossible todistinguishagoodonefromabadonewithoutanexternalbenchmark.118Alongtheline,he adopted the benchmark of efficiency, “one that distinguishes exclusion throughefficiencyfromexclusionbymeansunrelatedtoefficiency”.119Bork understood “means unrelated efficiency”—which was inherent impropriety—specifically and narrowly as a practice’s misalignment with the profit maximizationpremise of price theory.120This is exemplified by his idea of adopting the predationrationale as the overarching approach to regulating all unilateral exclusionarypractices.121Forexample,Borkconsideredpredatoryprice-cuttingtobeoneofthefewpractices potentially condemnable, 122 because the short-run profit sacrifice made noapparentsenseundertheprofitmaximizationpremise;123forotherpractices,aslongastheywereprofit-maximizing, theywouldbepermissible.Even so,he concluded,uponfurtherexamination,thatpredatorypricingwasunlikelytogeneratecompetitiveharms

115U.S.v.AmericanTobaccoCo.,221U.S.106(1911),179;Bork,TheAntitrustParadox,18,36–37.Seealso,Stones,“TheChicagoSchoolandtheFormalRuleofLaw,”548.116Bork,TheAntitrustParadox,36,38(“Wheneconomicanalysisshowsthatapracticecanhavenosignificantbeneficialeffectsbutissolelyameansofrestrictingoutput,the‘inherent nature’ of the practice is injuriously to restrain trade. The practice then islabeledillegalperse,whichmeansthatthereisnodefenseandthatthecourtneednotexamineeitherintentormarketpowerbeforepronouncingthebehaviorunlawful”).117RobertH.Bork,“TheRuleofReasonandthePerSeConcept:PriceFixingandMarketDivision,”TheYaleLawJournal74,no.5(1965):804(“Iftheword‘inherent’inWhite'ssentencemodifies ‘effect,’ as seems likely, itmaybe that the test contemplatednotanexaminationofactualeffectsbutaninferenceoftheeffectfromsomeotherfact,probablyfromthemarketsizeorpowerofthepartyorparties.”).118Bork,TheAntitrustParadox,38–39.119 ibid., 38–39 (quote), 137, 311 (distinguishing “artificial barriers to entry” frombarriers thatare“formsofsuperiorefficiency”andequatingthe formerwithunlawfulexclusionarypractices).120ibid.,119–21.121Borkused“predation”tocharacterizeunilateralexclusionaryconduct,whichwastoimposelossesonrivalsbybearinginsignificantorrecoupableshort-runlossesitself.Seeibid.,137,144,147–48.Seealso,Elhauge, “DefiningBetterMonopolizationStandards,”269.122Bork,TheAntitrustParadox,39,148.123ibid.,144;Elhauge,“DefiningBetterMonopolizationStandards,”269.

21

according to the single monopoly profit theory: the lowered short-run prices wouldbenefit the consumers in away that cancels any possible competitive harms that aredirectlyensued,andtheexpectationoflong-runrecoupmentisunfeasiblebecauseanyattemptofpriceincreasewouldimmediatelyattractnewentrants.124Thesamelogicwasapplied to other practices such as exclusive dealing: a monopolist cannot coerce adistributorintoanexclusivedealingagreementwhilechargingamonopolyprice;sowhenitsignsadistributorforexclusivedealing,itmusthaveloweredthemonopolypriceinexchange.AccordingtoBork,thiswouldbe“tobuyalessprofitablemarketshare”.Underthe profit maximization premise, it would be a nonsensical strategy for excludingcompetitors,butitwouldmakesenseasastrategyofefficiency-enhancingintegration.125

3.3.2. ThenoeconomicsensetestMelamed (2006) distilled from the Alcoa precedent the normative notion that amonopolistshouldbeallowedtoreapitssuccessintheformofmonopolypowerandtoexcluderivalsinthatprocess.126Onthatbasis,heconstruedcompetitiononthemeritsas“competitionthatmeetssomenotionoffairplay—evenifrivalsareharmedbyit”,andsuggestedthatthelawfindsillegalityin“thenatureofthedefendant’sconduct”.127Alongtheline,headvocatedthenoeconomicsensetest—whichisanimprovedversionoftheprofitsacrificetestproposedbyBork—astheunifyingstandardforassessingunilateralexclusionaryconduct.128Accordingtothistest,competitiononthemeritsessentiallymeans“conductthatmakesbusinesssense”forbeingprofitable,129andsotheperselawfulnessoftheconductneedsnoweighingofitspro-andanti-competitiveeffects.130Infact,thistestseescompetitiononthemeritsasnotcapableofgeneratingcounterweighinganticompetitivenessatall.131Meanwhile, conduct thatmakes no economic sense but for its likelihood of excludingrivalsandgainingmarketpowerforchargingsupra-competitivepriceswoulddenoteabad nature and therefore should be presumed unlawful. 132 These two categories ofconductaredistinguished through the counterfactual “continuedviabilityof the rival”rationale:foracontentiousexclusionarypractice,if itappearsprofit-maximizinginthecounterfactual scenario where the rival was not excluded, it makes economic sense;

124Bork,TheAntitrustParadox,149–54.125ibid.,306–09.126Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”382.127ibid.,387–88.Thisreferencetotheintuitiveandunspecifiedideaof“fairplay”isthekeyinhisconceptualizationofunlawfulexclusionaryconductandendorsementof“thesacrificetest”.128Inessence,thenoeconomicsensetestisanimprovementoftheoriginalprofitsacrificetestbecauseittakesadynamicaccountofwhatisprofitmaximization.Thisisexemplifiedin the former’s application to predatory pricing. See ibid 391–92, 396–98; Werden,“IdentifyingExclusionaryConductunderSection2,”415–16,424.129Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”392,398.130ibid.,388.131AndrewI.Gavil,“ExclusionaryDistributionStrategiesbyDominantFirms:StrikingaBetterBalance,”AntitrustLawJournal72,no.1(2004):52–53.132Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”390–91.

22

otherwise, it does not.133 The underlying idea was that “making no economic sense”essentiallymeans “makingnoefficiency sense”and thus iswelfare-reducing.134This isshowninTable1.Table1:ThepresupposedanddeduceddistinctionsofthenoeconomicsensetestPresupposeddistinction Efficiency-based

exclusionary conduct:lawful

Non-efficiency-basedexclusionary conduct:unlawful

Deduceddistinction Conduct making economic(profit-maximization)sense

Conductmakingsenseonlyin the “but for”exclusionaryway

However,thereisaproblemwiththispresupposeddistinctionof“good(efficiency-based)”and“bad(non-efficiency-based)”competitivebehavior:itisunverifiable—atleastnotbythe making economic sense standard—because “economic sense” is a subjective anddynamicnotion. It isdifficult todrawaclear linebetweenmakingeconomicsenseandmakingsense in the“but for”exclusionaryway.135It ismoredifficultwhenwefactor instrategic interactions between firms. Under certain circumstances, conduct can beprofitableandatthesametimeexclusionaryinanon-efficiency-basedway.Thisisthecasewithexclusivedealingagreements,whereprofitablebutinefficientexclusionscanriseiftherearecollectiveactionproblemsamongthecustomersofadominantfirmorifthese customers cancolludeas intermediarieswith thedominant firm toexploit end-consumers.136Inthatsense,theeconomicsensestandardisunreliablefordistinguishingefficiency-based(lawful)andnon-efficiency-based(unlawful)conductandforpreservingtheefficiencies that it issetout topreserve. Incomparison, thewelfare-balancingtestaddressesthisissue,137butasdiscussedearlier,thistestfailstherequirementofbeingadministrable.Inotherwords,this“goodandbadcompetitivebehavior”presuppositionisintuitivebutfutile by itself when confronted by the fact that rival exclusions are inherent in the

133OrdoverandWillig,“AnEconomicDefinitionofPredation,”13–14.134Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”393.Seealso,ibid.,13(defining“predatorybehavior”).135As Salop pointed out, “it couldwell be that the firmwould be able to recover itsinvestment cost inamoreefficient technologyonly if itwereable togainamonopolymarketshareandraiseitsprice”(emphasisinoriginal).Salop,“ExclusionaryConduct,”346.136Elhauge, “DefiningBetterMonopolizationStandards,”284–92.Seealso, IlyaRSegalandMichaelDWhinston,“NakedExclusion:Comment,”AmericanEconomicReview90,no.1 (2000): 298–305 (showing that, even if buyer do not have coordination problems,exclusivedealingcanstillprofitablyexcluderivalsbyexploitingthecontractexternalitiesamongbuyersaslongasbuyerdiscriminationispossible).137 Salop, “Exclusionary Conduct,” 315–16 (emphasizing that this test is particularlyusefulforassessingconductthatisharmfulbyraisingrivals’costsratherthanbybeingpredatory).

23

competitive process,138and the no economic sense test championed by it provides noobjective and coherent criteria for applying the presupposed distinction.139Melamedsuggestedusing“intent”todrawthedistinction,140butanticompetitiveintentionsassucharenotlegallycondemnable.AsexplicatedintheArticle102caselaw,theyarerelevantonlyasaproxyforascertainingthelikelinessofthecompetitiveharmstranspiring.141Inthe same vein, the absence of fault does not guarantee exemption either. 142 This isdiscussedinthenextsubsection.

3.3.3. TheelementofintentIntentplaysasignificantroleintheconceptionofinherentimpropriety.Borkconsideredexclusionaryintentionstobeanecessarybutnotsufficientconditionforfindingunlawfulconduct.Ashestated,

Antitrustshouldattacknopracticeorarrangementonthegroundsthatitisexclusionaryor foreclosingunlessdeliberatepredationcanbeproved.Improper exclusion (exclusion not the result of superior efficiency) isalwaysdeliberatelyintended.143

The hard question is how “improper exclusion” intentions could be qualitativelydistinguishedfromefficiency-basedcompetitiveintentions,because,asBorksaidhimself,“Wheneveracompetitorcompetesheintendstotakebusinessawayfromrivals,whichinvolvesexcludingthem.”144Inotherwords,underthepremisethatallbusinessactivityexcludes,145the idea that there is a difference between “good” and “bad” intentions isincompatiblewith the adoption of efficiency as an external benchmark for separatinglawfulandunlawfulconduct.Thisisbecauseintentdoesnotcorrelatewithefficiency.Inthatsense,byrequiringthattheremustbeafindingofanimproperexclusionaryintentionbefore finding an unlawful practice, Bork narrowed the antitrust prohibition scopewithoutjustification.

138Gavil, “ExclusionaryDistribution Strategies byDominant Firms,” 21 (identifying anunansweredquestionintheAspenjudgmentbytheUSSupremeCourt:“whatistheproperapproach to assessing dominant firm conduct under Section 2when it producesbothanticompetitiveeffectsonrivalsandconsumers(i.e.,inefficiencies)andprocompetitiveeffectsfortheallegedexcludingfirm(i.e.,efficiencies)?”emphasisinoriginal).139Forexample, itwaspointedout that in theUnitedStates v.MicrosoftCorp, theD.C.Circuit adopted a no economic sense rationale to assess certain unilateral conduct.Insteadofassessingtheconductinquestion,itdefactopresumedthatefficiency-basedexclusionaryconductisnaturallybenignandnotcapableofcausinganycompetitiveharmwhereas non-efficiency-based exclusionary conduct is incapable of having efficiencyjustification.SeePopofsky,“DefiningExclusionaryConduct,”445–447.140Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”393–94.141Seenotes78–79andaccompanyingtext.142CaseT-65/89BPBIndustriesPlcandBritishGypsumLtdvCommissionoftheEuropeanCommunities[1993]ECRII389,para70.143Bork,TheAntitrustParadox,160.144ibid.,39.145ibid.,137.

24

Areedaalso resorted to theelementof intent for tracing the inherent impropriety.Heconsideredrelevanttheexamination“whetherthedefendanthadanintentiontoexcludebyimpropermeans”,146buthedidnotclarifywhatwashisunderstandingof“impropermeans”.Onthataccount,takingtheinherentimproprietynotionatfacevaluerisksbeingtrapped in circular reasoning: there is a preconceived notion of what is proper andimproperinthewaysthatfirmscompete,andwewouldjustknowwhenweseeit—eitherin actual manifestations or in the undertaking’s intentions, and this, in turn, wouldreinforcethepreconceivednotion.

3.3.4. ArepositioningofthenoeconomicsensetestunderArticle102Tosumuptheprecedingsubsections,theideaofinherentimproprietyisillusorybecauseexclusionsareinherentintherivalryprocess.Thedistinctionbetweenefficiency-basedandnon-efficiency-basedexclusionsisaquantitativeoneratherthanaqualitativeone.Accordingly, the rationale should be to find conduct unlawful when the exclusionaryeffects exceed a certain level that is benchmarked against certain external criteria.Attemptingtodefineunlawfulnessanyotherwaywouldbefutileatbestandmisleadingatworst.Tobeclear,thenoeconomicsensetestdoesshowstrengthforassessingcertaintypesofabuses.Forexample,itworkswellonrefusaltodeal,becausenormallyitwouldmakenosense for a firm to willingly forgo marginal profits unless the firm has an ulteriormotive.147But this featurecanalsobeaccountedforbydevisingpresumptions. In thatlight,thereisnopointinchoosingthemorecostlyoptionofadopting“noeconomicsense”asaunifyingtest.Instead,weshouldlimittheuseofthenoeconomicsensetest(includingthe profit sacrifice test) as a presumption of illegality applicable only to the types ofunilateral conduct that can be identified as “having no plausible prospect other thancompetitionrestriction”(orputdifferently,“nakedexclusions”148)byexternal,objectivecriteria.Oneoftheobjectivecriteriaistheinfringementofnon-competitionlawsasameanstotheabuseofdominanceatissue.TakeforexampletheconductofmisusingthepatentprocessinAstraZeneca,149orthedestructionofarival’sinfrastructureinLithuanianRailways.150Inthelessblatantformercase,thedeceitfulnessofthisconductwascontrarytotheruleslayingdownthepatentapplicationprocedures,andbecauseofAstraZeneca’sdominantposition,itraisedArticle102concerns.WhatmatteredwasthatAstraZenecaknowinglyviolatedtherelevantpatentrules—as indicatedbytheobviousdeceitfulness,andthat

146Areeda,“EssentialFacilities,”852(emphasisinoriginal).147Hovenkamp,“ExclusionandtheShermanAct,”157.148Forexamplesofreferencestothe“nakedexclusions”concept,seeMelamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”399;Salop,“ExclusionaryConduct,”317; Popofsky, “Defining Exclusionary Conduct,” 447 (characterizing “nakedlyanticompetitiveconduct”as“reasonablylikelytocauseanticompetitiveeffectsandlackedanyjustification”).149TheCommissiondecisiononAstraZeneca,recital144.150CommissionDecisionof2October2017relatingtoproceedingsunderArticle102oftheTreatyontheFunctioningoftheEuropeanUnion(CaseAT.39813–BalticRail),recital48.

25

violation provided a shortcut for ascertaining the ensued competitive harms. Otherobjective criteria include but are not limited to the intentions behind the contentiousconduct.151These criteria serve the purpose of confirming the presupposed inherentimpropriety that is at theheartof thenoeconomicsense/profit sacrifice test.Overall,considerationsof thesecriteriaaretobeunderstoodaspartof the“allcircumstances”approachtoabuseassessment152andacorrespondencetothecontext-baseddichotomyof“byobjectversusbyeffectrestrictionsofcompetition”underArticle101.153Inshort,thisarticlearguesthatthe“noeconomicsense”rationaleshouldonlyapply,asarebuttable presumption of illegality, to unilateral conduct that is found to have noplausible pro-competitiveness after a contextualized, objective assessment. If thisassessmentsuggestsotherwise,weneedastandardagainstwhichcompetitiveharmscanbebenchmarkedtofindabuse.

3.4. Thebenchmarkingofexclusionaryeffectsandtheasefficientcompetitortest

3.4.1. DisproportionalityasanalternativetoinherentimproprietyforfindingabuseElhauge (2003) proposed an abuse test alternative to the no economic sense test.Heconsidered that all rivalries exclude and suggested that, when determining theunlawfulnessofexclusionaryconduct,thekeyistotrackhowtheconductcontributedtothefurtheringofmonopolypower.Iftheconductinquestionfurtheredthemonopolybyenhancing the firm’sownefficiency, the conduct shouldbe cleared;but if the conductfurthered themonopolypowerbyundermining rival efficiency, the conductshouldbeunlawful,irrespectiveofwhethertheconductenhancedthefirm’sownefficiency.154ThisisshowninTable2.Underlying this test is the idea that “the undermining of rival efficiency” and “theenhancementoffirm’sownefficiency”aremutuallyexclusive,evenincases(ofverticalrestraintsforexample)whereambivalenteffectsareidentifiable.Accordingly,beingtheresultofjustifiableexclusionarypractices,afirm’sownefficiencyenhancementswouldnotcomeattheexpenseofrivalefficiency.Anyrivalexclusions—andtheimpliedex-postinefficiencies—resultedfromsuchpracticeswouldsimplybethepermissibleoutcomeofthe firm’s exploitation of its greater efficiency, rather than an undermining of rivalefficiency.155Table2:ThepresupposedanddeduceddistinctionsoftheElhaugetest

151Seenotes78–79andaccompanyingtext.152Seenote72andaccompanyingtext.153Seenotes81–82andaccompanying text.TheGenericsUKrulingputs “the relevantlegalandeconomiccontext”tothecenterstageforidentifyingrestrictionsbothbyobjectandbyeffect.Accordingly,adoptingacontextualizedapproachtoidentifyinga“byobject”restrictiondoesnotunderminethe“byobjectversusbyeffect”dichotomyunderArticle101.154Elhauge,“DefiningBetterMonopolizationStandards,”256.155ibid.,318–20(agreeingwiththeUSSupremeCourt’sperseruleoflegalityinAspenon“amonopolist’s efforts tomake itselfmore efficient than rivals and then exploit thatgreaterefficiencyinawaythatdrivesoutrivals”).

26

Presupposeddistinction Efficiency-based

exclusionary conduct:lawful

Non-efficiency-basedexclusionary conduct:unlawful

Deduceddistinction Conduct enhancing thefirm’sownefficiency

Conduct underminingrivals’efficiency

Thistestidentifiestwoscenarioswhereanexclusionarypracticewouldbefoundunlawful.Thefirstscenarioiswhereapracticeunderminedrivalefficiencywithoutenhancingitsown efficiency (such as pricing below cost). Indubitably, it should be outlawed. Thesecondscenario ismore delicate. It iswhen a practice (such as exclusive dealing andrefusaltodeal)underminedrivalefficiencyandatthesametimeenhancedtoacertainextent the firm’s own efficiency. There, the practice in question should still beoutlawed,156becauseitwasadisproportionatewayofenhancingorexploitingthefirm’sownefficiency.Forexample,arefusaltodeal—whichissupposedlyafirm’sexercisingofits property rights and a reflection of its efficiency—would be unlawful if it wasdiscriminatorilytargetingatrivals,becausethisdiscriminationelement—whichcausedthe undermining of rival efficiency—went beyond the exclusionary effect that wasnecessary for securing the property rights. 157 The same reasoning was applied toexclusivedealing: theexclusivityconditions“discriminateagainst thosewhodealwithrivals”, 158 and “in such cases achieving those efficiencies by internal expansion willgenerally be a less restrictive alternative to achieving them with exclusionaryconditions”.159This reasoning bears a striking resemblance to the “ability to exclude” conception ofdominance under Article 102, which entails a special responsibility on a dominantundertakingofnottodistortthecompetitiveprocess.160Inthatsense,essentiallywhatElhauge did was upholding the Chicago School’s efficiency objective (as the criterionaccordingtowhichlawfulandunlawfulconductisdistinguished)andatthesametimefactoringinthedominant/monopolymarketstructure.Consequently,Elhaugerefusedtotakeatfacevaluetheideathatdominantfirms’behaviorisgenerallywell-alignedwiththe robustself-correction forcesof themarkets, and insteadadoptedamorenuancedapproach. 161 Elhauge’s test can be compared with the AEC rationale: the formeraccentuates the need to preserve ex-ante efficiencies (which permits rival-exclusionsflowing from the exploitation of superior performance)whereas the latter focusesonascertaining ex-post whether the exclusion of certain rivals was indeed the result ofefficiency enhancement and exploitation. Despite the different viewpoints, they reachsimilar conclusions concerning the idea of competition on the merits: it is all aboutbenchmarkingtheexclusionaryeffects.

156ibid.,256–57.157ibid.,323.158ibid.,321.159ibid.,256,324.160ibid.,327.161ibid., 325–26, 336 (contrasting the robust forces of free market competition with“privateself-regulation”ofthemonopolists).

27

3.4.2. TheasefficientcompetitortestJudgeRichardPosneriscreditedforpioneeringtheAECtest.Heacknowledgedthat“apracticemay be at once exclusionary and efficient” from an ex-post viewpoint,162andsuggestedusingtheAECbenchmarktoweighthesetwokindsofeffects.163Supposedlybythe weighing, the AEC benchmark distinguishes non-efficiency-based exclusions fromefficiency-basedones.164Posnerunderstoodnon-efficiency-basedexclusionsastheonesthateventuallyresultin“anincreaseinthemarketpriceabovethecompetitivelevel”.165Notably, his advocacy for the AEC test was accompanied by a distrust of usingexclusionary intent as a proxy for identifying non-efficiency-based exclusions. 166Accordingly, this test dives directly into the measuring and weighing of efficiency-impairingand-enhancingeffects,andforthat,ithasnootherchoicebuttorelyheavilyonprice-costcomparisons.167However,itisimportanttonotethat,althoughtheprice-costcomparison plays a big role in the application of the AEC test, it is not a necessaryconditionofthelatter,aspricingisnottheonlymeansofnon-efficiency-basedexclusions.AmajordoubtoftheAECtestisitsunder-inclusiveness.First,itwassuggestedthattheAECtestdoesnottakeadynamicaccountofhowlessefficientcompetitorscouldposeasignificant competitive restraint,168 especially when economies of scale or scope arepresent.169Second, itwaspointedout that theAEC testmaybe functionally irrelevant

162Posner,AntitrustLaw,193.163Ashewrote,“ineverycaseinwhichsuchapracticeisalleged,theplaintiffmustprovefirstthatthedefendanthasmonopolypowerandsecondthatthechallengedpracticeislikely in thecircumstancestoexclude fromthedefendant’smarketanequallyormoreefficientcompetitor.Thedefendantcanrebutbyprovingthatalthoughitisamonopolistandthechallengedpracticeexclusionary,thepracticeis,onbalance,efficient.”Seeibid.,194–95.164ibid.,196(“Onlywhenmonopolypowerisusedtodiscourageequallyormoreefficientfirmsandthusperpetuateamonopolynotsupportedbysuperiorefficiencyshouldthelawstepin.”)165ibid.,200–01.166ibid., 214–15 (“Anydoctrine that reliesuponproofof intent is going tobeappliederratically at best.”). For an opposing opinion this, see Nazzini, The Foundations ofEuropeanUnionCompetitionLaw,64.167 O’Donoghue and Padilla, The Law and Economics of Article 102 TFEU, 232–33(explainingtheprice/costtestproposedbytheCommissionforfindingabusivepricingpracticesundertheAECrationale).168Salop,“ExclusionaryConduct,”328(discussingtheinadequacyoftheAECtestfromtheperspective of potential competition). See also, Gavil, “Exclusionary DistributionStrategiesbyDominantFirms,”59(disapprovingtheAECstandardasbeing“asnapshotofcompetitionatabriefmomentintime”).169Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”388(“arivalthatislessefficienttodaymightbecomeequallyormoreefficientifpermittedtimetodeveloplearning-by-doingeconomiesorifitssalesgrewandenabledittogainscaleeconomies”).

28

whenitcomestopracticeslikethemisuseofthepatentprocess.Suchpracticeswouldeffectivelyexcludeallcompetitors,nomatterhowefficienttheyare.170Anotherdoubtisthepredictabilityofthetest.Firmswouldfinditdifficulttoassessex-antetheirbusinesspractices(especiallynon-pricingpractices),becausethecriteriathistestusesforfindingunlawfulconductmaynotbeforeseeableorcontrollablebythefirmresponsiblefortheconduct.171HeretheAECtestisincontrastwiththenoeconomicsensetest,whichrelieson“theobjectiveeconomicconsiderationsforareasonableperson”asthepreliminarycriterionforfindingillegalconduct.172Insum,theAECtesthaslimitedtoolsforassessingnon-pricingconductanditraisestwoformalrequirementconcerns.Critically,however,noneoftheseissuesareinherentintheAECrationale.Withsomecalibrationsinapplyingthistest,theseissuescanbeaddressed.Thisisdiscussedinthenextsection.

4. Concretizingcompetitiononthemeritsbywayofminimumefficientscale

4.1. TheconceptofminimumefficientscaleThe concept ofMES is associatedwith economiesof scale.When themarginal cost ofproductionisbelowtheaveragecostperoutput,witheveryadditionaloutputunitthereisaless-than-proportionalcostincrease,andsoitismorecost-effectivetoproduceatalargerscale.Asa featureofmarketstructure,economiesofscaleexistwhentherearelargefixedcosts.173Arelatedconceptiseconomiesofscope,whichexistwhenthetotalcostsofproducingtwoormoreproductsbythesamefirmare less thantheaggregatecosts of each product being produced by a different firm. This is because there arecommoncosts forproducingmultipleproductsor therearecostcomplementarities inproduction.174Economiesof scope can reinforce economies of scale in amultiproductsetting.Theextentofscaleeconomiesdependsonthemarginalcostcurve.ThisisshowninFigure1.SupposethatmarginalcostappearsasaU-curveasitusuallydoes.So,ithasthelowestpoint(A):marginalcostfallstoabottomastheoutputlevelincreasesuptothequantitycoordinateofthatpoint(qa),andstartstorisethereafter.Thereon,asmarginalcostkeepsrisingwiththeoutputexpansion,atacertainpoint(B) theamountofmarginalcost isbound to catchup to theamountof average cost andmarginal cost increasebecomesproportional to the total cost increase.175When that happens, economies of scale are

170O’DonoghueandPadilla,TheLawandEconomicsofArticle102TFEU,233;Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,74.171Melamed,“ExclusiveDealingAgreementsandOtherExclusionaryConduct,”388.Seealso,Rousseva,RethinkingExclusionaryAbusesinEUCompetitionLaw,427(summarizingtheCommission’sutilizationoftheAECrationaleforfindingabusesofdominance).172Werden,“NetworkEffectsandConditionsofEntry,”417–18.173Pepall,Richards,andNorman,IndustrialOrganization,69.174ibid.,75.175Inthatsense,onewaytodetectthepresenceofeconomiesofscaleistoverifywhetherthetotalcostincreaseisproportionaltothetotaloutputincrease.Thiscanbedoneby

29

exhausted.Thisexhaustioncanbedelayedbywideningthegapbetweenaveragecost(ca2)and marginal cost (ca1) at that lowest point (A). In other words, the larger the fixedproductioncostsare,themoreextensivetheeconomiesofscaleare.Figure1:Economiesofscaleandthemarginalcostandaveragecostcurves

MEScanthusbedefinedas the lowest levelofoutputatwhicheconomiesofscaleareexhausted.176Astheextentofscaleeconomiescorrelatespositivelywiththeleveloffixedcosts,thesizeofMESdoestoo.Inmarketswheretheproductionrequiresalargeamountoffixed(sunk)costs,afirmwouldneedtoberelativelylargetoreachMES.MEScanbemeasuredbyestimatingfirmcostfunctions,thesamewaythateconomiesofscalecanbemeasured.177For antitrust, the variable size of MES necessitates the consideration of the relevantmarketsize.Ifbydefinitionthetotalamountofdemandislargeenoughtoaccommodatemany firms, a large MES will not necessarily entail a concentrated market structure.However,ifbydefinitiontherelevantmarketislimitedandtheMESislarge,themarketstructureislikelytobeconcentrated.Normallymarketdemandincreasesastheprice(orcost)goesdown,anditreachesthefullpotentialatthepointwheretheproductioncostisdrivendowntoaminimum.Also,itmaybethecasethatthetotaldemandofamarketisincreasingordecreasinginter-temporally.178ThesefactorshavealimitingimpactonhowmanyfirmscanachieveMES.Tobesure, therearemanydeterminantsofmarketdividingthepercentageincreaseintotaloutputbythepercentageincreaseintotalcost.Seeibid.,70–71.176ibid.,70.177ibid.,83–85.178EinerElhauge,“WhyAbove-CostPriceCutstoDriveoutEntrantsAreNotPredatory—andtheImplicationsforDefiningCostsandMarketPower,”YaleLawJournal112,no.4(2003):713.

30

structure, and the impact of MES on the market structure can be negated orovershadowedbyotherfactors.ThisisshowninTable3.Theextremeofaconcentratedstructureisanaturalmonopoly.Morespecifically,iftheamountofmarketdemandat theminimumlong-runcost level (whichmeans the costcannotbeloweredanyfurtherbystaticefficiency)islessthantwicethesizeoftheoutputofafirm’sMES,themarketwouldbeanaturalmonopoly.179Inthatcase,thereshouldnorealistic expectation of long-term competition, and so non-efficiency-based exclusionswouldbeunlikely.AlessextremescenarioiswherethepotentialtotaldemandismorethantwicethesizeoftheMESandsoatleasttwofirms(anincumbentandanentrant)canstayviable.Inthatcase, the incumbentandtheentrantareequallyefficient fromthestart,180albeitsomeincumbencyadvantagesenjoyedbytheformer.Thereshouldbenoproblemunlesstheincumbentisusingtheadvantagestoforecloserivals,becausethennon-efficiency-basedexclusion becomes a valid concern. 181 There is also the likelihood of horizontalcollusion,182butthiswouldbewithinthepurviewofadifferentantitrustsegment.Table3:ThedifferentMESsizesinrelationtotherelevantmarketsize

-Thesizeoftherelevantmarket +

SmallMES and largemarket:fragmented

LargeMES and largemarket:indeterminate

SmallMESandsmallmarket:indeterminate

LargeMESandsmallmarket:concentrated

-ThesizeofMES+

So far this subsection describesMES from a static viewpoint.Market size can change,however. Ina fast-growingmarket,arival isunlikely tobe inefficientlyexcludedfrom

179ibid.,790.180ibid.,791.181Seegenerally,ChiaraFumagalliandMassimoMotta,“ASimpleTheoryofPredation,”TheJournalofLawandEconomics56,no.3(2013):595–631.(explaininghow,underthepremisesofscaleeconomiesandsequentialbuyers,anincumbentcanuseitsincumbencyadvantagetoexcludeamoreefficientrival).182Elhauge,“DefiningBetterMonopolizationStandards,”321.

31

reachingMESbecauseoftheplentifulcontestabledemands.183Moreover,inmanycases,it ismoreaboutcompeting for themarketand lessaboutcompeting in themarket,asdynamic competition drives firms to create new demand curves and new consumersurplusgains.184Insuchcases,themarketmaybenarrowlydefinedandtheMESfoundlarge(onaccountof,forexample,significantR&Dinvestments),butinefficientexclusionsofrivalswouldstillbeunlikely.

4.2. TheuseofMESunderArticle102

4.2.1. DeutschePostInDeutschePost,theCommissionresortedtotheMESnotionfordeterminingwhetherthedisputed fidelity rebates were abusive. As found regarding the relevant market,“Successful entry into themail-orderparcelservicesmarket requiresa certain criticalmass of activity”.185 This criticalmasswas quantified as some 100million parcels ayear.186Theunderlyinglogicisthatanewentrantcancompetewiththeincumbenthead-ononlyafteritgetsastablefootholdinthemarket.TheCommissionconsideredthattheexclusive-dealingrequirementsattachedtothedisputedrebatesprevented,bylockinginthecustomers,thecompetitorsfromachievingthatMES.187Forconfirmation,itpointedtothefindingthattherewasnoentryandtheincumbenthadaconsistentmarketshareofover85%fortenyears.188Thusanabuseofdominancewasestablished.Althoughplausible for takingan initiativeofquantitativeanalysis,189theCommission’suseofMESinDeutschePostwassomewhatsuperficial.Itdidnotexplainhowthe“100millionparcelsayear”quantificationcameabout.NordiditfurtherassesshowthislargeMESmightbeindicativeofthemarketstructure.Ifasuccessfulentryintotherelevantmarketindeedrequired“considerablesunkcosts”forinfrastructure,190thelackofentryattempts would make at least some efficiency sense: the incumbent (Deutsche Post)enjoyedsignificanteconomiesofscopeaswellasscale,asithadsignificantcommonfixedcosts for (1)providing the commercialmail-orderparcelservicesand (2) fulfilling thepublic obligation of maintaining a reserve capacity for over-the-counter parcelservices.191Thescopeandscaleeconomieswereinawaycompensation for thepostaluniversalserviceobligationitbore.192Undersuchcircumstances,anentrywouldnotbe

183FumagalliandMotta,“ASimpleTheoryofPredation,”616–17.184SidakandTeece,“DynamicCompetitioninAntitrustLaw,”600–01.1852001/354/EC:CommissionDecisionof20March2001relatingtoaproceedingunderArticle82oftheECTreaty(CaseCOMP/35.141–DeutschePostAG),[May5,2001]OJL125/27,para37(hereinafter,theCommissiondecisiononDeutschePost).186ibid.,paras32,37.187ibid.,paras39,42.188ibid.,para32.189Nazzini,TheFoundationsofEuropeanUnionCompetitionLaw,241.190TheCommissiondecisiononDeutschePost,para32.191ibid.,paras8–10.192 Pietro Crocioni, “On the Relevant Cost Standard for Price-Cost Test in Abuses ofDominance,”JournalofCompetitionLaw&Economics14,no.2(2018):282.

32

easyindeed,193buttobefair,thepotentialentrantsneedednottobearanypublicserviceobligationeither.Therefore,thefactthattherewasnoentryandthemarketsharewasconsistentlyabove85%mightnotbeentirelyattributabletothefidelityrebates.Ofcourse,thisdoesnotnecessarilymeanthatthefidelityrebateswerenotabusive,194butamorenuancedanalysisfollowingtheMESlogicwouldhavebeenappreciated.

4.2.2. TheContestabilityDistinctionintheGuidancePaperIn theGuidancePaper, theCommissiondistinguishedcontestableandnon-contestablesharesofdemandwhenintroducingtheAECrationaletorebatesanalyses.195Thepointistoacknowledgethatthedominantpositionyieldsanincumbencyadvantagethatrendersonly a certain proportion of the market open for competition, so it is necessary toascertainhowthenon-contestablepartcanbeusedasleveragetolockincustomersandtoforecloserivalsonthecontestablepart.196Accordingly,aprice-costcomparisonintheAECspiritmustaddressthe following issue:whenarebateschemeisbasedonall thepurchases from the incumbent, a competitor for a customer’s certain contestabledemands would need not only to match the incumbent’s price discounts on thosecontestable demands but also to compensate the customer’s loss of discounts on theremainingpurchasesfromtheincumbent(themainpartofwhicharenon-contestable).197Thisincumbencyadvantagecanbeexplainedby,amongotherfactors(suchaslearning-by-doing),thepresenceofsignificantsunkandfixedcostsforentryandexpansion.198Theincumbentmadesuchinvestments,buttherivalsareeitherconstrainedfromcommittingsuch investments or still waiting for the invested infrastructures or technologies tobecomeoperational (which could takeyears),199so the customershavenoalternativesupplysourcesbuttheincumbentforlargeproportionsoftheirdemands.Importantly,

193FumagalliandMotta,“ASimpleTheoryofPredation,”612(explaininghow,insuchascenario,thereservedmonopolyinonemarketcaneffectivelyexcludeamoreefficientrivalinasecondmarket).194SegalandWhinston,“NakedExclusion,”298–305.195The EuropeanCommission,Guidance on the Commission’s Enforcement Priorities inApplyingArticle82ECTreatytoAbusiveExclusionaryConductbyDominantUndertakings,[24 February 2009]OJ C45/7, recitals 41–44 (hereinafter, “the CommissionGuidancePaper”).196ibid.,recital39.197ibid.,recitals41–42(“Theeffectivepricethatthecompetitorwillhavetomatchisnottheaveragepriceofthedominantundertaking,butthenormal(list)pricelesstherebatethecustomerlosesbyswitching,calculatedovertherelevantrangeofsalesand intherelevantperiodoftime.”).198FumagalliandMotta,“ASimpleTheoryofPredation,”597,610.199CommissionDecisionof13May2009relatingtoaproceedingunderArticle82oftheECTreatyandArticle54oftheEEAAgreement(COMP/C-3/37.990-Intel),recital886(hereinafter, “the Commission decision on Intel”) (observing that, in the x86 CUPproductionindustry,“Expandingoutputrequiresadditional(sunk)investmentintonewproperty,plantandequipmentaswellasseveralyears’ leadtime.”).Seealso,Elhauge,“WhyAbove-CostPriceCutstoDriveoutEntrantsAreNotPredatory,”778(“Theremightbe economies of scale or scope at low output that are only available over time asproductionisrampedup.”).

33

thesunkandfixedcostsalsoyieldextensiveeconomiesofscale,whichgivetheincumbentasizeablecostadvantagecomparedtorivalsthathavenotyetmadesuchinvestments.Therefore, it is equally necessary to delineate this cost advantage before assessingwhetheranas-efficientcompetitorisbeingexcluded.AstheCommissionexplainedintheInteldecision,theAECanalysisis“ahypotheticalexerciseinthesensethatitattemptstoanalyzewhetheracompetitorwhichisasefficientasIntel,butwhichwouldnothaveasbroad a sales base as Intel, would be foreclosed from entering”. 200 To that end,determiningtheincumbent’sMESiscrucial,foritcanserveasanAECbenchmarkthatfiltersout theexcessive impactof the incumbent’sunrivaled salesvolumeon the coststructure.Unfortunately,however,thistaskofisolatingscaleeconomiesisnotalwaysthoroughlyundertaken. This is largely attributable to the conception of contestable and non-contestable shares: as an oversimplified account of the incumbency advantage, 201 itautomaticallydiverts theanalytical attention to “the leveragingof thenon-contestableshares to lock in customers” and away from “the delineation of the incumbent’s costadvantagederivedfromthescaleeconomiesthatapotentiallyas-efficientrivalcanbutdoesnotyethave”.This isproblematic, as the incumbencyadvantageofholdingnon-contestabledemandsharesandthecostadvantagefromhavingscaleeconomiesarenotthesamething:theformerisultimatelytheresultofhighbarrierstoentryandexpansionandthusraisespotentialconcerns,whereasthelatteristheresultoffixedandsunkcostsand a neutral feature of themarket at hand. The conception of contestable and non-contestablesharesemphasizestheformeranddownplaystheimportanceofthelatter.On top of that, the case law instructs a potential effects threshold 202 and a ratherunforgivingstandpointtowardsexclusivity.203Thisrendersunnecessarythelookingintoof the extent and possible changes of the non-contestable proportion in the givencircumstances. Consequently, the possible efficiency accounts that flow from scaleeconomiesareseriouslyoverlooked.ThisisshownintheCommission’sAECanalysisinIntel.

4.2.3. TheCommission’sAECanalysisinIntelTheIntelcaseisthenewestadditiontoalonglineofprecedentsonrebates.Atthecenterof this decade-long legal controversy is the Commission’s AEC analysis on rebatesfeaturinganexclusivityelement.There,theCommissionconsideredthepresenceofnon-contestableshares,204mostlyalongthe lineofhowtheymightbe leveragedtomake itimpossibleforarivaltomatchthepricediscounts.205ThemissionofisolatingscaleeconomiesforanaccurateAECanalysis206wasskewed,inthesensethatadeterminationoftheMESwasnowheretobefound.TheCommissiondid

200TheCommissiondecisiononIntel,recital1004.201O’DonoghueandPadilla,TheLawandEconomicsofArticle102TFEU,232.202SeeSection2.2.1above.203TheECJjudgmentonHoffmann-LaRoche,para89.204TheCommissiondecisiononIntel,recital1009.205ibid.,recital1006.206ibid.,recital1004.

34

notethelargeMESpresentinthemarket,butonlyasanindicatorofhighentrybarriers.207Asexplainedintheprevioussubsection,thisisamischaracterizationofMES.AlargeMESassuchisnotnecessarilyanobstacleforanas-efficientcompetitortoenterorexpand.Unlike the incumbency advantage, MES is attainable by all rival firms—unless theincumbentisactivelyinhibitingthemfromdoingso.TheCommissionrejectedIntel’sefficiencyjustificationofscaleeconomies,onthegroundsthat the justification was imprecise and that the rebates were still unnecessarilyrestrictive of competition.208However, following theAEC rationale, a question can beaskedastowhatwouldbea“necessary”or“proportionate”restrictionthatenablesas-efficient competitors to enter and expand while unproblematically ridding off less-efficientones.ThisbegsadiscussionontheMES intherelevantmarket.ButsincetheCommission understood MES as nothing more than an indicator of entry barriers, itshould not come as a surprise that the Commission dismissed Intel’s scale economiesjustificationsoeffortlessly.Insum,thesamecanbesaidaboutIntelasaboutDeutschePost:therebateschemesatissuemayverywellbeabusive,butthelackofanMESexaminationinvitesdoubtsabouttheabusefinding.TheGuidancePaperdevisedthecontestabilityandnon-contestabilitydistinction to introduce the AEC rationale to rebates analyses, but the result iscompromisedbecausethisdistinctionignoresthecost-savingeffectofscaleeconomies.ThiscallsforanintegrationofMESintotheAECtest.

4.3. AbiggerroleforMESAs discussed in Section 3.4.2., the AEC test has been criticized for overlooking theexclusion of potentially as-efficient competitors, which could be welfare-reducing.Another issue is that the AEC test tends to be associated with the predatory pricinganalogyforitsstrongrelianceonprice-costcomparisons.Consequently,whenitcomestonon-pricingconduct,thereisatendencytomistakethetoolkitlimitationsoftheAECtestforitsinapplicability.ThisisexemplifiedintheIntelcontroversy.SomescholarsobjectedtotheAECtestbeingappliedtoexclusivityrebates.Theyarguedthattheexclusive-dealingrequirementssettheserebatesapartfrompredatorypricingpracticesandsowarrantedapresumptiveillegalitytreatmentwithouttheneedofanAECanalysis.209Tothatpoint,itshouldbereiteratedthattheAECrationaledoesnothavetobereducedto a price-cost comparison,much like that an exclusive dealing requirement does notnecessarilycallforanillegalitypresumption.Thestill-existingillegalitypresumptionon

207ibid.,recital866.208 ibid., recitals 1635–36 (stating that Intel failed to demonstrate “how theexclusivity/quasi exclusivity conditionsof thediscountswas capableof achieving” theallegedefficienciesandwhythe“efficienciesofscalecouldnotbegeneratedbyanotherdiscountingstructureorbyundiscountedlowerunitpricesforallcustomers”).209WouterP. J.Wils, “The Judgmentof theEUGeneralCourt in IntelandtheSo-CalledMoreEconomicApproachtoAbuseofDominance,”WorldCompetition37,no.4(2014):423;PaulNihoul,“TheRulingoftheGeneralCourtinIntel:TowardstheEndofanEffect-BasedApproachinEuropeanCompetitionLaw?,”JournalofEuropeanCompetitionLaw&Practice5,no.8(2014):521,526,530.

35

exclusivedealingunderArticle102isnotbecausetheexclusivityelementisinexcusable,butbecausethespecialresponsibilityconceptisexpansivelyinterpreted.210Theconceptof MES can be helpful for both complementing the AEC test and calibrating theinterpretationofthespecialresponsibilityconcept.

4.3.1. MESandtheAECrationaleTheconceptofMESiscomplementarytotheAECrationale,inthesensethatfirmsbeingrestrainedfromreachingMESisasufficientbutnotnecessaryconditionforas-efficientcompetitorsbeingexcluded.Rivalfirmscanbecomeas-efficientonlyaftertheyreachanMES.Meanwhile, as-efficient firmswill regress to anMES orworse if they are beingforeclosed—morespecificallyiftheircostsareraisedtotheextentthattheycannotaffordto produce as-efficiently anymore.211 Therefore, if firms are found not being able toachieveMES,itmaybethattheyarebeingforeclosedfromreachingormaintaininganas-efficientcostlevel.Inanyevent,itisimportanttolookatthemarketcircumstancesastohowdifficultitistoreach an MES and whether the incumbent is actively contributing to that difficulty.Admittedlyinthatregard,theMESconceptmaynotbeverystraightforward.Asmoothapplicationismorelikelytohappenincircumstancesthatfeaturelargesunkcosts,suchastheonesofDeutschePost.TheapplicationwouldbemorecomplexincircumstanceswhereahandfuloffirmswereabletoachieveMESwhileotherscouldnot.Inthatcase,wewouldneedtoascertainfurtherifthesituationwastheresultofarelativelysmall-sizemarketorcertain“raisingrivals’costs”practicesareatplay.

4.3.2. MESandthespecialresponsibilityconceptThe MES concept respects and accentuates the role of dominance in the Article 102paradigm.ItdirectstheattentiontowhetheradominantincumbentisactivelymakingitmoredifficultforrivalfirmstoachieveMES.Inthatsense,itcorrespondstothespecialresponsibilityconceptbyimplyingthat,whenafirmholdsmarket-widecontrolofpricesandcompetition,ithastobesubjecttoaproportionalityprinciplewhenenhancingandreapingitsefficiencies.212Thespecialresponsibilityconceptcanthusbeconcretizedasadominantfirm’sobligationnottoimpedearivalfirmfromreachingMES.Inthesamevein,abuse can be understood as behavior that undermines rival efficiencies to the extentbelowtheMESbaseline.213Thespecialresponsibilitybornebythedominantfirmcanbeputonaslidingscale,onaccount of the size of the MES in relation to the market size in the particular casecircumstances. Take for example a stylized scenario: if theMES takes up 30%of themaximummarketdemand,thenthefirmisallowedofexclusivedealingupto70%ofthe

210 Jones, “Distinguishing Legitimate Price Competition from Unlawful ExclusionaryBehaviour,”150,152–53.211Thomas G. Krattenmaker and Steven C. Salop, “Anticompetitive Exclusion: RaisingRivals’CoststoAchievePoweroverPrice,”YaleLawJournal96,no.2(1986):234.212Elhauge,“DefiningBetterMonopolizationStandards,”332.213ibid.,321.

36

marketdemand.ButiftheMESis40%,theallowednumberbecomes60%atmost.IftheMESis50%,exclusivedealingwillnotbeallowedaltogether.214

5. ConclusionsThe competitive process inevitably results in rival exclusions. This necessitates thedistinguishingoflawfulandunlawfulexclusionaryconductunderArticle102.Thelawhasadopted the dichotomy of competition on the merits and abuse to underscore thisdistinction, but the conception of competition on the merits is susceptible to themisguidanceofinherentimpropriety.Considering the trade-off between enforcement precision and enforcement costs, thisarticlesuggeststhat,insteadoftheinherentimproprietynotion,Article102needssimplya presumption of illegality for conduct that has no plausible explanation other thancompetitionrestriction.Thisiswherethenoeconomicsensetestcanbeuseful,providedthatitadherestothecontextualizedapproachlaiddowninthecaselaw.Ingeneral(andforconductthathasambivalenteffectsinparticular),thedistinctionbetweenabuseandcompetitiononthemeritsshouldbethemeasuringofcompetitiveharmsagainstacertainbenchmark.Amongthevariousproposedabusetests,theAECteststandsoutasthemostqualifiedforthistask.TheAECtestiscritiquedasbeingpotentiallyunder-inclusiveandanalyticallylimitedtoprice-costcomparisons.Althoughvalid,thesecritiquesdonotunderminethetheoreticalstrengthof theAECrationale fordrawingthesaiddistinction.TheconceptofMEScanhelp address these issues: it calibrates the AEC rationale on account of dynamiccompetition; it also equips this rationalewithanalytical tools for non-pricing conductsuchasexclusivedealing.215The “abuse versus competition on the merits” distinction pertains to the abuseassessment,butweshouldnotforgetthatthedominancepremiseispivotalintheArticle102 paradigm. This pivotality is crystalized as the special responsibility concept.Unfortunately, this concept lacks specification, and often it functions nomore than ashortcutfortheabuseassessment.Thisultimatelyunderminesthedominancepremise.The MES concept may help address this problem; we can understand the specialresponsibility concept as a dominant firm’s responsibility not to prevent a rival fromreachingitsMES.Accordingly,aslidingscaleofspecialresponsibilitiescanbeformulatedunderthe“allcircumstances”analyticalframeworkinindividualcases.Insum,theMESconceptdeservesmoreattention forpotentiallymakingArticle102enforcementmorerobust.

214ibid.,324.215JoshuaD.Wright, “Moving beyondNaive ForeclosureAnalysis,”GeorgeMason LawReview 19 (2012): 1166–67, 1181; Daniel A. Crane and Graciela Miralles, “Toward aUnifiedTheoryofExclusionaryVerticalRestraints,”SouthernCaliforniaLawReview84(2011):639.