MERGERS, STANDARD OF PROOF AND EXPERT - ACLE

30 Pages·2006·0.18 MB·English

JournalofCompetitionLawandEconomics2(1),71–100
doi:10.1093/joclec/nhi027
AdvanceAccesspublicationdateFebruary2,2006
MERGERS, STANDARD OF PROOF AND EXPERT
ECONOMIC EVIDENCE
Yves Botteman(cid:1)
ABSTRACT
D
There is an increasing use of complex econometric modelling in EC merger o
w
controlproceedings.Thequestioniswhethereconometricsaresubjecttoastan- nlo
dard of evidence similar to that applicable to facts and theories that the ad
e
CwhoemthmeirsstihoenretsrhadoiutilodnbaellysoumseesmaanrgdinaorfticduislacrteestioinnleifttstomtehrgeeCrodmemcisisiosinosnoinr d from
the treatment and handling of econometric evidence. In the former case, EC h
ttp
coofuecrtosnwomoueltdriecxeevritdaenncien.teInnstihveelraetvteierwcaosfe,thEeCCcoomurmtsiswsoiounld’saudsoepatnadraarthtiecrudlaetfieorn- ://jcle
entialapproach.WhiletheissuehasnotyetbeendealtwithbeforeECcourts,this .o
x
articlesubmitsthattheCommissionshoulduseeconometricswithcautionand, ford
hence,should meetarelativelyhigh evidentiary threshold beforeadmittingthe jo
u
results of econometric models into evidence. Several guiding principles of evi- rn
a
dence are suggested, which are not intended to negate the Commission’s ls.o
margin of discretion, but would, nevertheless, ensure that econometrics be rg
subjecttoafairlyhighstandardofproof. at U/
n
iv
e
rs
ite
I. INTRODUCTION it v
a
n
It is now standard practice for economic experts from the European A
m
Commission (DGCOMP), the notifying parties and third parties to partici- ste
pateactivelyinshapingthewaymergersshouldbeassessed.Asrecenthorizon- rda
m
tal noncoordinated effects merger cases illustrate,1 economics has taken an o
n
increasingly greater role in the assessment of mergers under the EC merger N
o
control system.2 Economists sponsor seminars, workshops and conferences ve
m
to raise awareness among antitrust lawyers and the business community be
r 2
2
(cid:1) SeniorAssociate,Linklaters,Brussels.Email:[email protected] , 20
1
liketothankBernardvandeWalledeGhelckeandDavidBaileyfortheirvaluablecomments.All 3
errorsandomissionsare,ofcourse,hisown.TheauthoractedforPeopleSoftintheECaspects
oftheOracle/PeopleSoftmerger.ThisarticlewascompletedinlateJune2005andisbasedonEC
caselawavailableatthattime.
1 SeeCaseCOMP/M.3083—GE/Instrumentarium,Commissiondecisionof2September2003;
CaseCOMP/M.3216—Oracle/PeopleSoft,Commissiondecisionof26October2004.
2 SeepublicationsponsoredbytheCommission,“TechnicalReport—EffectsofMergersInvolving
DifferentiatedProducts,”COMP/B-1/2003/07,7October2004,http://europa.eu.int/comm/
competition/mergers/others/effects_mergers_involving_differentiated_products.pdf(visitedon
5July2005).
#TheAuthor(2006).PublishedbyOxfordUniversityPress.Allrightsreserved.
ForPermissions,pleaseemail:[email protected]
72 JournalofCompetitionLawandEconomics
aboutthevariouseconomictoolsavailable,including regressionanalysesand
bidandmergersimulationmodels.3Thoseinstrumentshavenotyetbeenfully
tested in actual cases and their use is sometimes subject to cautions and
controversy.4 It is nevertheless appropriate to identify what role economic
models can play in the discharge by DGCOMPof its burden to prove that a
merger will significantly impede competition in a substantial part of the
CommonMarket.5
Inanattempttocontributetoandcomplementthedebateonthisissue,this
article starts with an overview of the various methodological and evidentiary
stepsthatDGCOMPhastogothroughincarryingoutmergerreviewsgenerally.
D
Thecollectionofrelevantevidenceisatthecoreofitstasks,butasthisintroduc- o
w
toryoverviewshows,thestandardofproofapplyingtomergercasesisill-defined nlo
a
despite efforts made by the CFI in the Airtours, Schneider/Legrand, and Tetra de
d
Laval/Sideltrilogy.Adiscussionthenfollowsonthestandard(s)ofreviewthat fro
m
Community courts have applied in the past, with a particular focus on the h
long-awaited ruling from the Court of Justice in the Commission’s appeal of ttp
theTetraLaval/Sideljudgment.6Practicalimplicationsaredrawnfromthosedis- ://jc
le
cussionsbeforeweenterintothedebateastowhatrolesandcontributionsecon- .o
x
omic models can have in the evidentiary threshold that must be met before ford
DGCOMP can satisfy itself that its opinion regarding the compatibility of a jou
rn
mergerissupportedbya“cogentandconsistent”bodyofevidence. a
ls
This contribution identifies a paradox. While the review exerted by the .o
rg
Community judicature might admittedly be less intensive when the a/
Commissioncarriesouteconomicassessmentsonthebasisofthefactualevi- t Un
iv
dence contained in the file, the proposition whereby the Commission should e
rs
clear a transaction whenever it is confronted by “grey area” cases7 may be ite
it v
a
n
3 IBCconferencesheldinLondonon11–12March2004andinBrusselson27–28January2005 A
m
entitled“TheUseofEconomicsinCompetitionLaw”. s
4 RBBBrief12,“Theemperor’snewclothes?Theroleofmergersimulationmodels,”January terd
a
2004;BenDubow,DavidElliott&EricMorrison,“Unilateraleffectsandmergersimulation m
models,”25(2)ECLR114(2004);GregoryWerden,LukeFroeb&DavidT.Scheffman,“A on
Daubert discipline for merger simulation,” Draft of 16 February 2004 (on file with the N
o
author);David Scheffman&Mary Coleman, “Quantitativeanalysesof potentialcompetitive ve
m
effectsfromamerger,”9June2003;MarcIvaldi&FrankVerboven,“Quantifyingtheeffects b
e
fromhorizontalmergersineuropeancompetitionpolicy”,Int’lJ.ofIndus.Org.(vol23,n.9, r 2
December2005,pp.669–691);andRoyJ.Epstein&DanielL.Rubinfeld,“Mergersimulation, 2, 2
asimplifiedapproachwithnewapplications,”69AntitrustL.J.883(2004). 0
1
5 Thisquestioniscloselyassociatedwiththeyetunresolvedissueofthestandardofproofappli- 3
cable to merger cases. For a detailed discussion of that issue, see, inter alia, David Bailey,
“Standard of proof in EC merger proceedings: a common law perspective,” 40 CMLR 845
(2003); Bo Vesterdorf, “Standard of proof in merger cases: reflections in the light of recent
caselawofthecommunitycourts,”BIICL—ThirdAnnualMergerConference,6December
2004; Mario Siragusa, Judicial Review of Competition Decisions Under EC Law, UK
CompetitionCommissionPapers,2004.
6 ECJ,CaseC-12/03P,Commissionv.TetraLaval,notyetreported.
7 Thestandardofproofisprobablyhigherthanamere50/50probabilitythattheeconomicharm
willmaterializeasaresultofthemerger.
StandardofProofandEconomicEvidence 73
inoperative in practice. More specifically, how should the Court review a
decision where DGCOMP has selected an econometric model that is not
entirelyconsistentwithmainstreameconomics?SupposeDGCOMPpreferred
to stand by a minority view that showed with a relatively high level of prob-
ability that prices would go up by a significant margin while other economic
models/assumptionswouldleadtheCommissiontoalessclear-cutresult.
Ourbasicpropositionisthateconometricscanusefullycontributetotheevi-
dentiary burden that the Commission has to discharge, but econometrics
should not be treated differently from the other factual elements of a merger
investigation, at least from the perspective of the evidentiary threshold that
D
hastobesatisfied.Likewise,themerefactthatCommunitycourtshaveconsist- o
w
ently held that the Commission enjoys a certain “margin of discretion” nlo
a
when handling economic evidence should not constitute a barrier against a de
d
thorough review of the Commission’s approach and reasoning in its use of fro
m
econometric evidence. While undeniably the “margin of discretion” concept h
is intended to highlight the limits of the Court’s judicial review, it should ttp
onlybeconstruedasbarringtheCourtfromcarryingoutitsownmergerassess- ://jc
le
ment. In this regard, the Commission is in charge of the anticompetitive .o
x
scenario. The Court’s role is merely to verify that the competitive scenario ford
is reasonable andthat the factual evidencethat isusedby theCommission is jou
solid, consistent and adequately supports the points made. Econometrics rna
ls
shouldmakenoexception.However,econometricevidenceentailssomething .o
rg
morecomplexthancrudefacts.Econometricmodellingisindeedintendedto a/
explainwitharelativelyhighdegreeofconfidence(andprobability)whatcon- t Un
iv
sequencesamergerwillbringtothemarketplace.Inviewofitscomplexityand e
rs
purpose, theCourtsmaynothaveall theexpertise themselvesto“re-do”the ite
analysis of the Commission in order to verify whether it is valid and solidly it v
a
n
grounded onthefactsofthecase andconformswith mainstreameconomics. A
m
Theymayneedtoseekassistancefromanexpertorinvolvethepartiesinelu- s
te
cidatinghowthemodelswereconstructed,whattheyintendedtodemonstrate rd
a
m
andwhytheyareconsistentwithmainstreameconomics.Undeniably,thecom- o
n
plexityofeconometricsmayhavetheconsequencethatCourtsmaynotbekeen N
o
todelvetoomuchintotheir intricaciesand,for thesakeofjudicialeconomy, ve
m
ratherprefer tocarryoutarelatively“light”reviewofeconometricevidence. be
r 2
2
, 2
0
1
II. THECOMMISSION’SSTANDARDOFPROOFINMERGERS 3
This section provides an overview of the Commission’s standard of proof in
merger reviews. It articulates the basic principles of evidence applicable to
mergers and concludes that the Commission’s evidentiary burden in merger
cases is setat afairly high threshold.8
8 NotethatthissectiondoesnotdiscusswhatstandardofprooftheCommissionhastomeetto
satisfyitselfthatamergerdoesnotraisesignificantconcernsandshouldbeallowedtoproceed.
74 JournalofCompetitionLawandEconomics
A. The Standard of ProofGenerally
The standard of proof is commonly understood as the necessary evidentiary
threshold that an authority or a court must meet before it can reach a
certain opinion or determination in a particular case.9 Depending on the
typeofmatterathand,civilorcriminal,thestandardwillbemoreorlessinten-
sive. At one end, the burden is set at a relatively low threshold and is often
defined under the “preponderance of the evidence” standard. This standard
is usually employed in civil matters where the court will put all the facts and
arguments that the parties have produced into the balance before reaching a
conclusion.Attheotherendofthespectrum,thestandardissetataverystrin- D
o
gentthreshold,namelythe“beyondreasonabledoubt”standardusedincrim- w
n
inalmatters.Thislatterthresholddemandsfromthecourtthatitsconclusions loa
d
begroundedonverysolidevidenceandthatanypossibilitythattheaccusedis ed
not guilty has been considered and rejected. Between the two extremes, the fro
m
standard of proof can moveacross the spectrum. h
ttp
MeIrngemr eRrgeegrulcaatisoens,10thheasConmotmsispseiocinfieids ctohnefreovnitdeedntwiairtyh tthhreesfhaoctldthtahtatthiet ://jcle
.o
must meet to prohibit a merger from going ahead. As the following sec- xfo
tions illustrate, the European Courts have given some guidance on the rd
jo
relevantstandardtobeappliedbutseemtohavedonesoinarelativelyincon- urn
a
sistent manner. Indeed, the CFI uses alternatively the phrases “in all likeli- ls
.o
hood,” “very plausible,” “very probable”. True, each phrase suggests that rg
a/
the standard is more than a mere preponderance of the evidence test. t U
Beyondsemantics,though,theissuewouldseemtorelatetothelevelofprob- niv
e
ability that the Commission has to reach before a merger can be held to be rs
anticompetitive.11 iteit v
a
n
A
m
s
te
rd
a
m
Onthispoint,seetheCommission’sdecisionintheSEB/Moulinexmergerandthesubsequent on
rulingoftheCFIinCaseT-114/02,Babylissv.Commission,[2003]ECRII-1279. N
o
9 See on this point David Bailey, “Standard of proof in EC merger proceedings”, 40 CMLR ve
m
2003, 845 at p. 848. The standard of proof is to be distinguished from the neighbouring b
e
conceptofburdenofproof.Whiletheformerrelatestotheevidentiarythresholdthathasto r 2
bemetbeforeadecisioncanberenderedinacase,thelatterdealsprimarilywithprocedural 2, 2
rules that determine who has to discharge the evidentiary threshold. In this article, the 0
1
burdenofproofisnotdiscussedinanydetailsinceArticle2oftheMergerRegulationmakes 3
clear that it is for the Commission to meet a certain evidentiary threshold before it is
allowedtoblockamerger.
10 CouncilRegulationNo.139/2004of20January2004onthecontrolofconcentrationsbetween
undertakings,OJ2004L24/1(hereinafterthe“MergerRegulation”).
11 Certain commentators have suggested that the probability that a merger is anticompetitive
shouldbesomewherearound70%.Itshouldbenotedthatthedeterminationofthestandard
of proofisto someextentcloselyconnectedwiththe“falsepositives” and“falsenegatives”
probleminantitrustdecision-making.See,e.g.,KenHeyer,“Aworldofuncertainty:economics
andtheglobalizationofantitrust”,72AntitrustL.J.375(2005).
StandardofProofandEconomicEvidence 75
B. TheCommission’s Four Stepsof Analysis
TheMergerRegulationdisposesthattheCommissionhasthedutytomakean
appraisal as to whether a concentration significantly impedes competition in
the relevant market. The Merger Regulation is, however, silent as to the evi-
dentiary threshold that the Commission has to reach before declaring a con-
centration to be incompatible with the Common Market. One may identify
several steps before such a determination can be reached. In particular, the
Commission has to gather all relevant facts, develop the basic theoryof anti-
competitive harm, specify the applicable test(s) and select the appropriate
econometric tools that would support its opinion. These four steps do not D
o
take place all at once or within a specified order. Rather, the facts will w
n
determine the theories and tests to be used and vice versa. In this sense, the loa
d
process is relatively dynamic and evolving. Some or all the above steps may ed
be considered with varying degrees of depth and detail in the course of the fro
m
first phase of the investigation, depending on the issuesat hand. h
ttp
dutWytitohcroengdaurdcttao“tthheofraocutgs,hitanisdwpiadienlsytarkeicnogg”n1i2zeindvtehsatitgtahtieoCn.oNmammiesslyiointmhaussta, ://jcle
.o
among others, collect all relevant qualitative and quantitative information xfo
from the market in a comprehensive manner. The Merger Regulation has rd
jo
giventheCommissionwidediscretionarypowersastothetypeofinformation urn
a
that it may collect from the notifying parties, its competitors, suppliers, and ls
.o
customers. The CFI has also given a relatively wide margin of discretion as rg
a/
to the scope and depth of information that the Commission may request t U
pursuant to Article 11 of the Merger Regulation.13 niv
e
Aspartofthisinvestigatoryprocessandonthebasisofthedataprovidedby rs
ite
the parties in their Form CO and in response to the various information it v
requests, the Commission will start developing the basic anticompetitive an
A
harmtheorybehindthetransaction.Itwilltypicallypayattentiontothepoten- m
s
tialanticompetitiveeffectsthattheconcentrationmaygenerateusingasabasic te
rd
framework the two theories laid out in the Horizontal Merger Guidelines,14 as am
well as the conglomerate and/or vertical effects theories. In particular, the on
N
coordinated and unilateral effects theories under the Horizontal Merger o
v
e
Guidelines will often be considered useful instruments to better identify and m
b
frame the issues. The Commission’s task will not only be to articulate the er 2
theory of anticompetitive harm, but to go a step further by laying out the 2, 2
0
1
3
12 See Opinion of Advocate-General Tizzano of 25 May 2004 in Case C-12/03, Commission
v.TetraLaval,para74.Thissentencereferstothegeneralprincipleofgoodadministration
underwhichtheCommissionmustoperateincarryingoutitsmergerinvestigationsandisa
meanstoachievetheexpectedoutcome,i.e.thatthecasebesupportedbycogentandconsistent
bodyofevidence.
13 Seeinparticular,CaseT-310/01,SchneiderElectricv.Commission,[2002]ECRII-4071,para
109.
14 Guidelines on the assessment of horizontal mergers under the Council Regulation on the
controlofconcentrationsbetweenundertakings,OJ2004C31/5.
76 JournalofCompetitionLawandEconomics
testandtheconditionsthatmustbesatisfiedbeforeits“gutfeeling”or intui-
tivethinking rises tothe level of informed assessment.
Oncethetestshavebeenproperlyspelledout,theCommissionwillhaveto
useanumberofeconomictoolstoverifywhethereachpropositionofthetestis
fulfilled. The tools range from very basic instruments, such as market-share
information, diversion ratios,15 win/loss analyses, and critical loss analyses,16
to more sophisticated econometric instruments such as regressions,17 and
bid and merger simulation models.18 The appropriateness of each of those
instruments will have to be considered in light of the underlying theory of
anticompetitiveharm.19
D
o
w
n
lo
a
C. The Standard of Evidence inMergerInvestigations d
e
d
Thecaselawisalsounclearabouttheapplicablestandardofproofinmerger fro
m
investigations.20Fromtheoutset,threeconsiderationsshouldbehighlighted. h
ttp
Fwihriscth, ththeestacnasdeardisosftapnrdooinfgm. aFyovrariynsdtaenpceen,diantgtohnethenedstaogfepohfapsreocIe,ssthaet ://jcle
.o
Commission seems to enjoy relatively broad discretion to decide whether x
fo
there are sufficient elements in the file to warrant the initiation of an in- rd
jo
depth investigation. To the extent that an Article 6(1)(c) decision does not urn
a
affect the legal position of the undertakings concerned, the decision is not ls
.o
rg
15 Ibid,para29.DiversionratioscalculatetheamountoflostsalesbyproductAtoproductB at U/
whenthepriceofproductAincreasesbyacertainpercentage.Diversionratioscanbeestimated n
usingown-priceandcross-priceelasticities.Ahighdiversionratiobetweentwoproductsindi- ive
catesthattheproductsareclosesubstitutes.Alesssophisticatedalternativetodiversionratios rsite
includescustomerswitchingstudiesorwin/lossanalyses. it v
16 Criticallossanalysiscalculatesthevolumelossesthatwouldbenecessarytomakeapost-merger a
n
priceriseunprofitableanddetermineswhethersuchalosswouldbelikelyinviewofthecharac- A
m
teristicsoftheindustry.Criticallossisconsideredasausefultooltodefinemarkets,butcan s
te
also,incertaincircumstances,beusedtoassesstheunilateraleffectsofamergerbyproviding rd
a
theminimumlostsalesnecessarytodefeatapriceincrease.Criticallossanalysishasbeenused m
by the Commission in recent merger cases such as, Case COMP/M 3060—UCB/Solutia, on
Commissiondecisionof31January2003,para42. N
o
17 Regression analysis assesses the relationship between one variable called the explained, or ve
m
dependant,variableandoneormoreothervariablescalledexplanatoryvariables.Regression b
e
analyses are used as a quantitative technique to examine how the variable in question (e.g. r 2
price)isaffectedbyanumberofothervariables(e.g.reductioninthenumberofcompetitors 2, 2
or theincreasedlevelofconcentration).Asweshallseebelow,regressionmodelshavebeen 0
1
usedrecentlyinGE/InstrumentariumandOracle/PeopleSoft. 3
18 MergersimulationmodelshaverecentlyemergedinEUmergercasesasaquantitativetech-
niqueusedtoestimatethepriceeffectsofamerger.See,e.g.,CaseCOMP/M.3191—Philip
Morris/Papastratos,Commissiondecisionof2October2003,para32,whereamergersimu-
lationwaspresentedbythemergingpartiestodemonstratethatthepriceimpactoftheconcen-
trationwouldbeminimal.
19 Notethatthisarticleconcentratesontheuseofthemostcomplexeconometrictoolsinmerger
investigations.
20 SeeAliNikpay&FredHouwen,Casecomment,“TourdeForceoralittlelocalturbulence?A
hereticalviewontheAirtoursjudgment”,24(5)ECLR2003,196,at202.
StandardofProofandEconomicEvidence 77
subjecttojudicialreviewbytheCommunityjudicature.Sincethestandardof
proof is a concept closely related to the availability and intensity of judicial
review,21 it is not surprising to see that Courts have likewise not defined any
evidentiaryrequirement before acase can go into the second phase.
The second consideration relates to the shifting in the burden of proof.
While it is not the purpose of this article to cover this aspect in any great
detail, it seems clear that the notifying parties playa key role in the adminis-
tration of the evidence.22
Third, the standard of proof is admittedly much higher when the
Commission reviews past and current market conditions than when it starts
D
to rely on those facts to predict the likely competitive effects of the merger o
w
in the short and mid-term.23 While factual evidence should be established nlo
a
with a relatively high level of accuracy, reliability, and consistency, the stan- de
d
dard of proof should not likewise require the Commission to prove the anti- fro
m
competitive effects with “absolute certainty”. That is, given the predictive h
nature of merger control, the reviewing agency should not be required to ttp
prove the anticompetitive effects of the merger using the standard of proof ://jc
le
applicable to criminal proceedings, that is, the “beyond reasonable doubt” .o
x
standard. Nevertheless, the CFI in Tetra Laval/Sidel held that, when the ford
effects of the merger do not immediately affect the structure of the market, jou
rn
the Commission must be convinced that merger will “in all likelihood” allow a
ls
the parties to achieve a dominant position in the near future.24 That is, the .o
rg
further in time such effects are predicted to materialize, the higher the a/
burden the Commission must deliver. On appeal of the CFI’s judgment in t Un
iv
the Tetra Laval/Sidel case, Advocate General Tizzano suggested, however, e
rs
that giventhe prospectivenature of mergeranalysis, it is ite
it v
a
n
sufficient if on the basis of solid elements gathered in the course of a thorough and A
painstakinginvestigation,andhavingregardtoitstechnicalknowledge,theCommission m
s
is persuaded that the notified concentration would very probably lead to the creation or te
thestrengtheningofadominantposition.25(emphasisadded) rda
m
o
n
If the Commission is not able to reach such an opinion, it must clear the N
o
merger.TheCourtofJusticeseemstohaveadoptedasomewhatlessstringent ve
m
standardof proof: b
e
r 2
2
the quality of the evidence produced by the Commission in order to establish that it is , 2
0
necessarytoadoptadecisiondeclaringtheconcentrationincompatiblewiththecommon 13
market is particularly important, since that evidence must support the Commission’s
21 SeeSectionIIIbelow.
22 Thepartieshaveadutytoprovidecomplete,accurateandnon-misleadinginformation.See
Article14oftheMergerRegulation.
23 AliNikpay&FredHouwen,seeabovenote21,at199.
24 CaseT-5/02,Tetrav.Commission,[2002]ECRII-4381,para148.
25 OpinionofAdvocate-GeneralTizzanoof25May2004inCaseC-12/03,Commissionv.Tetra
Laval,para74.
78 JournalofCompetitionLawandEconomics
conclusionthat,ifsuchadecisionwerenotadopted,theeconomicdevelopmentenvisaged
byitwouldbeplausible.26(emphasisadded)
The standard of proof applying to the Commission’s prospective analyses—
that is, the projected anticompetitive effects flowing from the merger—raises
a number of questions. While the ECJ’s “plausibility” test sounds sensible
and approaches that applied in US courts,27 in practice its application is
subjecttoanumberofhurdles.Inparticular,whatdoesa“plausible”standard
mean? Does that mean that econometrics would be subject to less scrutiny,
especially where such instruments are used to predict the likely effects of a
D
merger? o
w
LetustrytoarticulatetheaboveECJstandardinpracticalterms.Asafirst nlo
a
step, the Commission must get the facts right and describe the pre-existing d
e
d
market conditions with accuracy. As the ECJ points out, the quality of fro
the data is key to the reliability of the Commission’s prospective analyses. m
h
The task of the Commission is particularly demanding. For instance, in the ttp
context of particularly complex merger cases such as the recent GE/ ://jc
le
InstrumentariumandOracle/PeopleSoftcases,thishasresultedinaveryburden- .o
x
some data gathering process.28 fo
rd
Once facts are established, the Commission has to assess the facts and jo
u
determine whethercertain conditions and requirements are met. Allrelevant rna
ls
factors must be taken into account and carefully reviewed before any con- .o
rg
clusion is drawn as to their significance and impact on market conditions. In a/
thisrespect,suchconclusionsmustsatisfytherequirementsof“logic,coherence t U
n
andappropriateness”.Inparticular,theCommissionmaynotusethesameset ive
of facts to support twocontradictoryfindings.29 rsite
Next, the Commission’s opinion entails a complex technical assessment, it v
a
n
“based not on the application of precise scientific rules but on criteria and A
m
principles which are open to question, such as economic ones”.30 In this s
te
regard, one should pay particular attention to the selection, use, and cali- rd
a
m
bration of economic theoriesand instruments. o
n
It is quite clear that the selection and use of economic theories should be N
o
supported by the facts. For instance, while a broad-brush review of the case ve
m
may suggest that a particular merger may be explained using a unilateral b
e
effects scenario, the Commission has to ensure that the facts and the r 2
2
, 2
0
1
26 ECJ,C-12/03P,Commissionv.TetraLaval,notyetreported,para44. 3
27 Inpreliminaryinjunctioncases,UScourtsapplythe“preponderanceoftheevidence”standard.
28 OnemayquestionwhetherthestricttimetablessetforthintheMergerRegulationconflictwith
theevidentiaryrequirements.Itseemstousthat,inviewofthenecessitytobaseitsopinionon
solidfactualgroundstheCommissionwillprobablycontinuetomakeextensiveuseofthe“stop
theclock”provisionunderArticle11ECMR.Forinstance,intheOracle/PeopleSoftmerger,
theclockwasstoppedtwiceandthereviewperiodwasextendedbyabout6months.
29 CaseT-342/99,Airtoursv.Commission,[2002]ECRII-2585,para142.
30 OpinionofAdvocate-GeneralTizzanoof25May2004inCaseC-12/03,Commissionv.Tetra
Laval,para73.
StandardofProofandEconomicEvidence 79
conclusions drawn therefrom support the unilateral effects theory. In this
respect, there is much less guidance as to the evidentiary value to be given
to econometric instruments. There is equally little guidance on the level of
confidence or probability that the Commission must arrive at in relation
to the results generated by the operation of those econometric models.
The ECJ only indicates that the economic effects predicted by the econo-
metric theories and models should be “plausible”. While the selection of a
particular econometric tool should logically be made in accordance with a
relatively well established set of guidelines, the way a particular instrument
should be used—the parameters and variables to be included in the model
D
and the admissible assumptions that the Commission may draw—and the o
w
data upon which it should rely are subject to debate, even among nlo
a
economists.31 de
d
For instance,inthecontext of adifferentiatedproductmerger raisinguni- fro
m
lateraleffectconcerns,theCommissionmustascertaintheextenttowhichthe h
elimination of competition that existed between the merging parties prior to ttp
the merger will be of such magnitude as to allow the merged entity to find it ://jc
le
profitable unilaterally to raise its prices post merger. Resolving that question .o
x
calls for an evaluation of the relevanttest to be applied using the principle of ford
logic, coherence, and appropriateness as well as the selection, use, and cali- jou
rn
brationofeconometrictoolsthatmaybeparticularlysuitabletoestablishcon- a
ls
vincingly each branch of the test. In the hypothetical situation at hand, the .o
rg
Commission’s first task will be to demonstrate the closeness of competition a/
between the merging parties.32 Suitable econometric tools include win/loss t Un
iv
data, provided that the underlying information is sufficiently precise and e
rs
reliable to indicate against whom each of the merging parties competed. In ite
this respect, the way the win/loss sample has been prepared (what products it v
a
n
are covered and which geographies) and presented is an important aspect. A
m
Anothersetofinstrumentsthatmightbeconsideredincludesregressionana- s
te
lyses with a view to evaluating the extent to which the presence of one party rd
a
m
significantly impacts the pricing behaviour of the other party. Both instru- o
n
mentsrequirenotonlythatthefact-findingprocesshasbeencarriedoutpro- N
o
perly, but also that the econometric tools are calibrated to the particular ve
m
circumstancesofthecase.Thislatterpartissubjecttosomeformofarbitrage be
and it is unclear, under the ECJ’s proposed standard, what justification the r 22
Commission has to provide in support of the choice of one particular cali- , 20
1
bration or the selection of acertainvariableagainst another. 3
31 See,e.g.,GregoryJ.WerdenandLukeM.Froeb,“Unilateralcompetitiveeffectsofhorizontal
mergers”, inAdvancesintheEconomicsofCompetitionLaw(MITPress),draftof14January
2005(onfilewiththeauthor).
32 Theotherlimbsofthetestincludeevaluatingtheextenttowhichthenon-mergingpartiescould
beconsidereddistantsubstitutesforthemergingfirmsandtheirabilitytorepositiontheiroffer-
inginresponsetoapriceincreasebythemergedentity.
80 JournalofCompetitionLawandEconomics
Toexplorethisissue further,it is appropriate toreview howthe European
judicaturehasarticulatedthescopeofitsjudicialpowersinexercisingitsover-
sightof the Commission’sworkin past mergercases.
III. STANDARD(S)OFREVIEWINMERGERCASES
Fromaconstitutionallawpointofview,judicialreviewcanbedescribedasan
institutional concept defining the scope of intervention of the judiciary in
relation to acts and decisions adopted by the administrative bodies of the
European Union. There is the idea that Courts may not be invested with
D
thepowertoreviewanadministrativedecisiondenovobutshouldnevertheless ow
n
inquireastothelegalityofthedecisionagainstasetofprinciples,forexample, lo
a
theprincipleofgoodandsoundadministration.Thiswouldsuggestthatthere de
d
is a certain level of deference that Courts should maintain and keep in mind fro
m
when reviewingthe legalityof administrative decisions. h
ttp
://jc
A. The Scope of Judicial Review inMergerCases le.o
x
fo
InaccordancewithArticle230oftheECTreaty,Commissiondecisionsmay rd
jo
beannulledonthegrounds“oflackofcompetence,infringementofanessen- u
rn
tialproceduralrequirement,infringementofthisTreatyorofanyofitsruleof als
.o
law relating to its application, or misuse of powers”. The Courts’review will rg
consist of verifying whether: (i) the Commission carried out its duties under at U/
the ECMR by conducting a “thorough and painstaking” investigation; n
iv
(ii)therelevantproceduralruleshavebeencompliedwith;(iii)thestatement ers
of the reasons for the decision is adequate;33 (iv) the facts upon which the ite
Commission’s assessment is based are correct; and (v) whether there has it va
n
been any manifest error of appraisal or misuse of powers.34 Grounds based A
m
on legal and factual infringements are captured by the “manifest error of ste
rd
appraisal” standard. In RJB Mining, the CFI explained how this concept a
m
shouldbe understood: o
n
N
as regards the evaluation of the situation resulting from economic facts or circumstances ov
e
underlying the contested decision, it is settled case law that the Court, in conducting its m
b
review,mustconfineitselftoascertainingwhethertheCommissionmisuseditspowersor e
manifestlyfailedtoobservetheprovisionsoftheECSCTreatyoranyruleoflawrelating r 22
, 2
0
1
33 ThiswouldincludewhetherthereasoningoftheCommissionsatisfies“requirementsoflogic, 3
coherence and appropriateness”. See ECJ, Joined Cases C-68/94 and C-30/95 France and
Othersv.Commission,[1998]ECRI-1375(“KaliandSalz”),para229–241.
34 See,e.g.,ECJ,Case42/84,RemiaandOthersv.Commission,[1985]ECR2545,para34,and
Joined Cases142/84and156/84;BATandReynoldsv. Commission, [1987]ECR 4487,para
62; Case C-7/95, John Deere v. Commission, [1998] ECR I-3111, para 34: “[The Court]’s
reviewofcomplexeconomicappraisalsmadebytheCommissionisnecessarilylimitedtover-
ifyingwhethertherelevantrulesonprocedureandonthestatementofreasonshavebeencom-
plied with, whether the facts have been accurately stated and whether there has been any
manifesterrorofappraisaloramisuseofpowers”.