A methodology to measure the quality of tax avoidance case studies: Findings from the Netherlands

JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

ContentslistsavailableatScienceDirect

Journal

of

International

Accounting,

Auditing

and

Taxation

A

methodology

to

measure

the

quality

of

tax

avoidance

case

studies:

Findings

from

the

Netherlands

Anna

F.

Gunn

_,

_Dirk-Jan

_Koch

_,

_Francis

_Weyzig

a_{PartneratGunnTaxCommunicationandAffiliatedtoLeidenUniversityandErasmusUniversityRotterdam,theNetherlands} b_{RadboudUniversityNijmegen,SpecialProfessorInternationalTradeandDevelopment,CooperationandChiefScienceOfficerof} NetherlandsMinistryofForeignAffairs,theNetherlands

c_{CentralPlanBureauoftheNetherlands,theNetherlands}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Availableonline15May2020

Keywords: Taxavoidance TheNetherlands NGOs Researchquality

Dataavailabilitymethodology

a

b

s

t

r

a

c

t

Inrecentyears,therehavebeensubstantialeffortstocombatcorporatetaxavoidance. Theseeffortshavebeenpropelledinpartbymediatizedcasestudies,conductedby non-governmentalorganizations(NGOs)andothergroups,onthetaxavoidancepracticesof multinationalenterprises.Thesecasestudieshavebeencriticizedbecausetheyallegedly lackquality,butthiscriticismhasnotbeenassessedacademically.Thisresearchseeks toaddressthatgap.Itproposesanewmethodologytoanalyzethequalityofthesecase studiessystematically.Weconstructtenindicatorsrelatedtoallegedweaknessesanduse thesetoassess14casestudiesinvolvingDutchcorporateentities.Wefindthatthequality ofthesecasestudiesisaffectednegativelybyalackofadequatedata.Thus,ifcompanies andgovernmentsenhancetransparency,thiscouldincreasethequalityofcasestudies byNGOsandothergroups.Inadditiontothis,wefindthattheNGOsandothergroups themselvescansometimesincreasethequalityoftheircasestudiesbyinvestingintechnical expertiseandadoptingpracticesthatfosterobjectivity.Themethodologydevelopedfor thisstudycouldalsobeofuseforothertopicaldebates,suchasdisclosurerequirements andcorporatesocialresponsibilityreporting.Whereaswewereabletoaddresschallenges relatedtointernalvalidity,theexternalvalidityofthefindingscouldstillbeimprovedby extendingtheselectionofcasestudies.

©2020TheAuthor(s).PublishedbyElsevierInc.Thisisanopenaccessarticleunderthe CCBYlicense(http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Inrecentyears,theareaofinternationaltaxationhasbecomeincreasinglyprominentinpoliticalandpublicdiscourse,

bothonthenationalandinternationalstage.Muchattentionhasbeengiventothecorporatetaxpositionofmultinational

entities(MNEs)and thebroadand oftenrather abstractnotionsof “aggressivetaxplanning”and“tax avoidance”.The

debateoninternationaltaxisnotstaticandcandifferbycountry.Furthermore,ittendstocenteraroundspecificinstances

regardingaspectsofthetaxsystemorthebehaviorofspecifictaxpayers.IntheNetherlands,issuessuchastheuseofmailbox

companies,confidentialtaxrulings,andunlawfulstateaidareextremelytopical.Thepublicdebatehasbeenshapedinpart

∗ Correspondingauthorat:RadboudUniversityNijmegenSpecialProfessorInternationalTradeandDevelopmentCooperation,P.O.Box9104,Nijmegen 6500HE,theNetherlands.

E-mailaddress:d.koch@maw.ru.nl(D.-J.Koch). https://doi.org/10.1016/j.intaccaudtax.2020.100318

1061-9518/©2020TheAuthor(s).PublishedbyElsevierInc.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/ by/4.0/).

2 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

bytheinfluenceofnon-governmentalorganizations(NGOs),suchasOxfamNovib,ActionAid,andtheTaxJusticeNetwork

(TJN).

ManyoftheseNGOsenteredthetaxdebatefromtheperspectiveofdevelopingcountries.Taxavoidanceindeveloping

countrieswasconsideredtobeespeciallyproblematicinlightofhighlevelsofpovertyindevelopingcountriesandtheneed

forthesecountriestoinvestinsectorsasinfrastructure,education,andhealthcare.NGOshavecontributedtothisdebate

inseveralways;crucially,thisincludesresearchintoconcreteinstancesof(alleged)taxavoidancebyaspecificMNE.Such

researchtendstofocusonspecificaspectsoftheworldwidetaxstructureofanMNE,consideringinparticularinstances

whereincomeissubjectedtoaveryloweffectivetaxrate.Casestudiesareakeyingredientforthepublicdebate.They

illustratetheimpactoftaxrules,whicharefrequentlycomplexandabstract,andindoingsomakeitpossibleforpeoplewho

donothaveaspecialistbackgroundintaxlawtoparticipateinthedebate.NGOsprovidesimplifiedyetrelevantexamples,

sometimesinvolvingawell-knownbrand(e.g.Grolschbeer),whichprovideaclearstartingpointforfurtherdiscussion.This

dynamicallowsNGOstoexertadegreeofinfluenceoverthesubjectmatterofthedebate,andtheframingoftheissueof

taxavoidance.ThecasestudyapproachbyNGOswasdevelopedasawayofcircumventingproblemsintheavailabilityof

data(Finér&Ylönen,2017).

Asfarasweareaware,nopreviousacademicresearchhastriedtosystematicallyanswerthequestion:whatisthequality

ofthesecasestudies?Inlightofthesignificanceofcasestudiesforthedebateontaxavoidance,theabsenceofsuchresearch

isnoticeable.Thislastpointisunderlinedbythefactthatcasestudieshave,inthepast,beencriticizedfortheir(alleged)

lackofquality,e.g.withrespecttotheir(alleged)biasandlackofexpertiseonthepartoftheauthorswhocarriedoutthe

casestudy.However,asfarasthepresentauthorsareaware,casestudieshavenotyetbeenanalyzedfromanacademic

perspective.Thismay,inpart,beduetotheabsenceintheinternationaltaxandaccountingliteratureofamethodologythat

couldbeusedtocompletesuchananalysisinameaningfulmanner.

Theprincipalaimofthisarticleistoassessthequalityofcasestudies(alsoreferredtoas’reports’).Weseektocontribute

totheexistinginternationaltaxandaccountingliteraturebyprovidinganewmethodologyforassessingthequalityoftax

avoidancecasestudies.Toenablesufficientdepth,wehavelimitedthetestingofthemethodologytocasestudieswithtax

avoidancestructuresinvolvingDutchcorporateentities.Thisstudyalsoaimstoprovideapositivestimulusforthequality

offuturecasestudiesbyNGOs.Itistothisendthatweincludeanumberofconcreterecommendationsinsection5.

Thestructureofthispaperisasfollows.Section2outlinestheanalyticalstartingpoints.Startingwiththedefinitionof

theterm“taxavoidance”,itthengoesontodescribethecurrentinternationaltaxdebateandtheroleofcasestudiesinthat

debate,aswellastherelevanceofthequalityofthesecasestudiestoNGOs’effectiveness.Section2closeswithanoverviewof

thedebateonthequalityofthesecasestudies.Basedonthisoverview,theparticipative,inclusive,anditerativequantitative

casestudymethodologyisdevelopedinsection3.Usingthatmethodology,section4systematicallyanalyzesthe14case

studiesthatinvolvetheNetherlands.Thequalityofthecasestudiesisappraisedindetailfromtheperspectiveofthree

elements:thelackofadequatedata,thelackofexpertise,andthelackofobjectivity.Section5discussesthefindingsand

providesseveralrecommendationstoincreasethequalityofthecasestudies.Oneofthemainrecommendationsconcerns

theavailabilityofdata,inwhichcompaniesandgovernmentshavearoletoplay.Section5concludeswithremarksonthe

roleofthesecasestudiesinthefuture.Interestingly,sincethemethodologywasdevelopedandappliedtothecasestudies

intheNetherlands,someofthequalitystandardsthatwerewithinthesphereofcontroloftheNGOsappeartohavebeen

partiallyadoptedbyNGOs.

2. Analyticalstartingpoints:thesignificanceofcasestudiesandareasofcriticism

2.1. Thecontroversialconceptof“taxavoidance”

Thereiscurrentlynoconsensusamongtaxexpertsonthemeaningoftheterm“taxavoidance”andseveralrelatedterms,

suchas“aggressivetaxplanning”and“taxdodging”(Heitzman&Hanlon,2010;Kovermann&Velte,2019;Wang,Xu,Sun,

&Cullinan,2020,forthcoming).InitsGlossaryoftaxterms,theOrganisationforEconomicCooperationandDevelopment

(OECD)describedtaxavoidanceasfollows(OECD,2020):

atermthatisdifficulttodefinebutwhichisgenerallyusedtodescribethearrangementofataxpayer’saffairsthatis

intendedtoreducehistaxliabilityandthatalthoughthearrangementcouldbestrictlylegalitisusuallyincontradiction

withtheintentofthelawitpurportstofollow.

Weconsiderthistobeausefulstartingpointforthispaper,althoughitshouldbemadeclearthattheOECDdescription

isnotcomprehensive,universallyaccepted,orindeedunambiguous.For example,ifacountryintentionallyadoptstax

legislationthatenablesMNEstoreducetheirtaxliabilityabroad,sucharrangementsmaystillberegardedastaxavoidance

fromtheperspectiveofothercountries.Sometimesalternativetermswithasimilarmeaningareused,suchas“aggressive

taxplanning”.Thispaperdoesnotaimtodefinetheconceptof“taxavoidance”orsimilarterms.Consequently,noclaims

aremaderegardingwhetherthetaxstructuresexploredbyNGOsinthevariouscasestudiesconstitute“taxavoidance”in

theeyesofthepresentauthors.Forpresentpurposes(i.e.developingamethodologytoassessthequalityofcasestudies),it

sufficestoestablishwhetheraparticularcasestudydefineskeytermsandconcepts(suchas“taxavoidance”or“aggressive

taxplanning”)andwhetherthatstudyincludesalegalanalysisoftherelevanttaxrules.Theobjectiveistoassesstowhat

A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 3

oftheriskofjudgingex-postwithreferencetolegaldevelopments,includingterminologyanddefinitions,postdatingthe

publicationofacasestudy.

2.2. Theinternationaltaxdebate

Inarelativelyshortperiod,issuesthatwerepreviouslymainlyofinteresttospecialists,havebecome“topicsofpublic

interestandpopularmediareporting”,whichhaveresultedin“newpoliticalinitiatives[which]havealsoprogressedatan

unexpectedhighpace”(Forstater&Christensen,2017,p.3).ThisisevidencedbytheOECD’songoingworkonbaseerosion

andprofitshifting(BEPS),whichwaslaunchedin2013andwhichseekstocombatcertainformsofharmfulorotherwise

undesirableinternationaltaxplanningbyMNEs.

TheEuropeanUnion(EU)providesanotherinterestingexample.In2012,theEuropeanCommissionannounceda

clamp-downontaxevasionandtaxavoidance(EuropeanCommission(EC)(2012)).Todate,thishasresultedinaseriesofnew

measures,includingtheautomaticexchangeoftaxrulingsbetweenEUMemberStatesandtheAnti-TaxAvoidanceDirectives,

requiringthephasedintroductionofa selectionofanti-abusemeasuresfrom2019onwards.Since2013,theEuropean

CommissionhasalsobeencounteringtaxavoidanceviatheEUstateaidrules.

Therehasbeenconsiderablemediaattentionaroundtheissueoftaxavoidance.Aseriesofdataleaks(e.g.LuxLeaks,

PanamaPapers,andParadisePapers)basedonresearchbytheInternationalConsortiumofInvestigativeJournalists(ICIJ)

hasbeenvitalinthisrespect,revealingthescaleofoffshoretaxplanningandprovidingconcretecasesinvolvinghigh-profile

personsandcompanies(Oei&Ring,2018).

2.3. Significanceofcasestudiesintheinternationaltaxdebate

NGOsplayaroleintheinternationaltaxdebate.Examplesofthisincludeparticipationinpublicconsultations(Seabrooke

&Wigan,2016,p.367)aswellasothertypesoflobbyingandadvocacyactivities.NGOsalsohaveadegreeofmediapresence,

regularlyasaresultoftheresearch(casestudies)thattheypublish.TheriseinthenumberoftimesthattheTaxJustice

Networkiscitedinthepresshasbeensubstantial(Dallyn,2017).Althoughsuchoutwardlyobservablefactsdonotreveal

theexactimpactandinfluenceofNGOsonthetaxdebate,itisclearthattheirrolein,forexample,theBEPSProjecthas

beensignificant.ItisalsoworthnotingthattwooftheEuropeanCommission’srecentinvestigationsintounlawfulstate

aidandMNEswerethedirectresultofcasestudiesbytradeunionsandtheGreens(EuropeanCommission,2015a;The

Greens/EuropeanFreeAlliance,2016c).

2.4. Threeareasofcriticismofcasestudies

TheeffectivenessofthecasestudiesasameansoffurtheringthepositionofNGOs,willdependtoagreatextenton

whetherotheractorsintheinternationaltaxdebateregardthecasestudiesasbeingsufficientlyauthoritative(cf.Seabrooke

&Wigan,2016).Consequently,thequalityofthecasestudiesisextremelyimportant.Ifotheractorsinthedebate,suchas

policy-makersorthemedia,donotacceptthevalidityoftheclaims,thisreducestheireffectivenessfurtheringtheposition

ofNGOs.Whilethequalityofparticularcasestudieshasreceivedpraise(Gunn,2015),therearealsoconcernsaboutthe

qualityofcasestudiesingeneral.Inthenextsection,weprovideanoverviewofthesecriticisms.Whiletherearepotentially

otheryardstickswithwhichtoassessthequalityofcasestudies,wefocusprimarilyontheabovementionedcriticisms.This

approachaimstoavoiddevelopingcriteriawhichseemrelevantfromthevantagepointofthe‘ivorytower’,butthatarenot

deemedtobesignificantbytheactorsinvolved.

Basedonahigh-levelanalysisoftheDutchmediacoverageoftheinternationaltaxdebate,thepresentauthorsdistinguish

threemainareasofcriticismofcasestudies.Thepresentauthorsdonotclaimthatthisisacomprehensiveoverview,as

othertypesofcriticismarevoiced,albeitlessoftenandlesscoherently.Forthepresentpurpose(i.e.thedevelopmentof

anassessmentmethodology),weholdtheviewthattheidentificationofthreetypesofcriticismissufficientlydetailedto

allowforameaningfulanalysisofcasestudies.Thefirstpointofcriticismconcernstheavailabilityofin-depthinformation

anddataneededtocarryoutacasestudy.Morespecifically,thepointofcriticismisthatavailableinformationanddatais

insufficientasabasisfortheanalysismadeinthecasestudy(area1).Thesecondpointofcriticismconcernstheallegedlack

ofobjectivityoftheauthorsofeachcasestudy(area2),whereasthethirdpointpertainstotheirallegedlackofexpertise

(area3).

Thethreeareasofcriticismaredistinct,yetsometimesoverlap,andcanbemutuallyreinforcing.Forinstance,alackof

objectivityonthepartoftheauthorsofcasestudiescouldpotentiallycontributetocomplacencyonthepointoftechnical

rigor.Inturn,alackofrelevantdatacouldcomplicateapropertechnicalanalysis,astheauthorsmaynotbeawareofall

facetsofaparticularcase.Insomeinstances,theboundariesbetweenthedifferentpointsofcriticismmaybeblurred.For

example,iftherepresentativenessofacasestudyisnotexplained,shouldthisqualifyasalackoftechnicalexpertiseor

insteadasalackofobjectivity?Fortheanalysisinsection3ofthisstudy,weaimtomakeourchoicesinthisregardas

4 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

2.4.1. Lackofin-depthinformationanddataaboutthecase(area1)

Thefirstareaofcriticisminvolvesclaimsregardingtheresearchers’allegedunderstandingofthecaseasaresultofa

lackofin-depthinformation.Taxpayersdonottypicallypublishtheir(entire)taxstructure,andrelevantinformationin

thisregardisoftennotmadepublic.IntheNetherlands,agreementsforpriorcertainty(“taxrulings”)providedbytheTax

Authoritiestotaxpayers,areconfidential,although,sincemid-2019,summariesoftheseagreementsaremadepublic(Snel,

2018).

Rulesonconfidentialityandprivacycanpreventresearchersfromobtainingtheinformationneededtoforma

compre-hensiveunderstandingofaparticularcase(Finer&Ylönen,2017).Althoughresearchersmaybeabletoidentifyacertain

leveloftaxplanning,itisgenerallyextremelyhard(orevenimpossible)toestablishtheexacttaxpositionofanMNE.Large

dataleaksoccurredinrecentyearsandshedlightoncertainaspectsofthetaxplanningofarangeofcompanies(Oei&Ring,

2018),butdonotalterthefactthatin-depthinformationcanbelacking.

2.4.2. Lackofobjectivityonthesideoftheresearchers(area2)

Thesecondareaofcriticismconcernstheobjectivityofresearchers.Thepersuasivenessofthecasestudiesistiedtothe

(perceived)expertiseoftheresearchers,whichinturnislinkedtonotionssuchasobjectivityandexpertise(seesection2.4.3).

CasestudiesbyNGOsarewritteninaspecificpoliticalcontextandhaveaparticularpurpose.Ithasbeensuggestedthat

researcherscanlackobjectivityandmaypotentiallytrytofindevidencethatmatchestheirpreconceivedconclusion(such

astheideathattheNetherlandsisa“taxhaven”).Insteadofpromoting“evidence-basedpolicymaking”,theseresearchers

arethoughttobeengagedin“policy-basedevidencemaking”(Seay,2015).

Insomecases,thecriticismthatresearcherslackobjectivityismadebycompaniestargetedbyaparticularcasestudy

(e.g.byFleuretteGroup(2016)).Thesecompanieshaveaninterestintheoutcomesofthecasestudies.Indeed,theymay

evenhaveaninterestindiscreditingthefindingsofacasestudy.Companieshaveafundamentallydifferentperspectiveon

theirtaxposition,astheyhaveaccesstoalltherelevantfactsandcircumstances(oftenincontrasttotheresearcherswho

wrotethecasestudy).Itisconceivablethathonestmistakesandmisunderstandings,whichareduetoalackofin-depth

informationonthesideofresearchers,areviewedbycompaniesasalackofobjectivity.

2.4.3. Lackofexpertise(area3)

Thethirdareaofcriticismconcernsthe(perceived)methodologicalshortcomingsofcasestudies,inparticularthe

(per-ceived)lackofexpertisebyresearchersofthecasestudy.Taxlawisaspecializedfieldoflaw,characterizedbyhighly

complicatednationallegislation,taxtreatiesandotherinternationalagreements,caselaw,andjurisprudence,aswellas

unwrittenprinciplesoflawandnationalpractice.

Internationaltaxlawisaspecialistfieldthatrequirestrainingandexperiencetonavigate.Itisoftendifficulttoestablish

towhatextentaresearcherofaparticularreportgraspsthelegalnuancesofthecaseathand.Theeducationalbackground

ofresearchersdoesnotprovideadefiniteanswertothisquestion,ifonlybecausetheresearchersmayseekadvicefrom

technicalexpertstosupplementtheirknowledge.Theperceivedlackoftechnicalskillsonthepartoftheresearchersis

flaggedasanareaofcriticism(e.g.Paladin,2015).WhendiscussingaresearcharticlebySikka(2010)ontaxavoidance,

HasseldineandMorris(2013),p.10)stated:“Wealsosuspectthattheresearcherdoesnotfullyunderstandsomeofthetax

arrangementsthatarediscussed.”

3. Methodology

Thefirstsectionofthemethodologyfocusesontheselectionandgroupingofthecases.Thismethodologicalsection

subsequentlyexplainswhyandhowwedevelopedaparticipatoryquantitativecasestudymethodology.Thethirdsection

highlightsthedevelopmentofgenericqualitycriteria.Thissectionisfollowedbyanexplanationofhowwedevelopedand

appliedascoringmethodology.Thismethodologicalsectionclosesbyindicatinghowthefeedbackoftheresearchersand

NGOswasincludedtostrengthenthequalityoftheappraisalmethodologyandoutcomes.

3.1. Selectionofthecasestudies

Tobeincludedinthisreview,acasestudyneedstofulfillthefollowingfourcriteria.Firstly,itneedstobeastandalone

casestudyfocusingontaxavoidanceoraggressivetaxplanning.Secondly,thecompanystructureaddressedinthecase

studymustincludeaDutchentitycontributingtoareductionintaxes.Thirdly,thestudywaspublishedbetween2013and

2017.Fourthly,thestudymustincludeaquantitativeestimateoftheamountofrevenueatstake.

OurstudyfocusesontheNetherlands.ThereiswidespreadconcernabouttheproblematicrolethattheNetherlands

playsinaggressivetaxplanningandtaxavoidance.Furthermore,theEuropeanCommissionissuedanegativedecisionon

stateaidgivenbytheNetherlandstoasubsidiaryofStarbucks(EuropeanCommission,2015b),whichwasannulledbythe

GeneralCourtin2019.Morerecently,theEuropeanCommissionopenedaninvestigationintopossiblestateaidgrantedby

theNetherlandstoInterIKEA(EuropeanCommission,2017)andNike(EuropeanCommission,2019).Dutchlegalentities

alsofeaturedprominentlyinthetaxplanningstructuresmadepublicbyLuxLeaksin2014,thePanamaPapersin2016and

theParadisePapersin2017.StudiesbyNGOshavefurthermoreputtheNetherlandsinthetopthreecountriescontributing

A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 5

TheDutchgovernmenthasvigorouslydeniedclaimsthattheNetherlandsengagedinharmfultaxpracticeand the

facilitationofaggressivetaxstructures(Accountantweek,2015).Overthepasttenyears,theNetherlandshastakenvarious

measuresaimedatcombattingtaxavoidance.Theseincludeunilateralmeasuresaswellastheimplementationofmeasures

followingfrominternationalagreements.TheNetherlandshasformallystatedthatitseekstocombatitsreputationasa‘tax

haven’,asthebadpresswasnegativelyaffectingthenationalinvestmentclimate(Kahn,2018).

Forthepresentpaper,41casestudieswereidentifiedoncompaniesthatwereallegedlyusingtheNetherlandsbecause

ofitstaxclimate.ThesearchforthesecasestudiesstartedbyverifyingthepublicationsbymajorNGOswhichfocusedon

internationaltaxationissuesintheirlobbyorcampaigns(e.g.ActionAid).Next,anacademickeywordsearchwascompleted

usingtermssuchas‘casestudies’and‘taxavoidance’toidentifyadditionalacademiccasestudies.Amoregeneralsearch

wasalsocompletedofthewebsitesoftheDutchandEuropeanparliaments,todetermineiftherewereanyothercase

studiesmentionedintheparliamentaryproceedingsthatwereneitherpublishedonthewebsitesofNGOsnorinacademic

journals.Finally,wecheckedforadditionalcasesfeaturedintheDutchmediaaroundkeymediamoments(e.g.LuxLeaks).

Afteridentifying41casesinthisway,weappliedthefourselectioncriteriamentionedabove(seeTable1;ascoreof1means

thatacriterionismet).Thislaststepresultedintheselectionof14casestudies.

The14casestudiesareclusteredintothreesubgroupsbasedontheirtypeofauthors,wheretheywerepublished,and

their(implicit)objectives(seeTable2below).Threeofthecasestudieswerecommissionedbyapoliticalparty,namely

theGreens/EuropeanFreeAlliance(EFA)(thisisarelativelysmallpoliticalpartyintheEUandhastaxjusticeasoneofits

priorities).Thesecasestudieswerepresentedonthewebsiteofthepoliticalparty.Themainobjectivesofthesecasestudies

appeartohavebeentopromoteadebateintheEuropeanParliamentandtoobtainstricterrulesagainsttaxavoidanceat

theEUlevel.

ThesecondsubgroupconsistedofthreecasestudieseventuallypublishedasarticlesinCriticalPerspectivesonAccounting.

WeincludedthecasestudiesinthesecondsubgroupbecausetheyarebasedonreportswrittenforNGOs.Althoughthe

objectiveofthesearticleswastocontributeacademicallytotheresearchfieldontaxavoidance,theobjectiveofthe

under-lyingcasestudieswascomparablewiththeobjectiveoftheothercasestudiesbyNGOsincludedinthescopeofourstudy.

Theassessmentofthecasestudiesincludedinthesecondgroupwasbasedonthecombinedpublications,i.e.theacademic

articleswereregardedascomplementarytotheoriginalcasestudieswrittenbytheNGOs.

ThethirdandlargestsubgroupconsistedofthecasestudieswrittenforanNGOandthenpublishedonthewebsiteofthat

NGO,withoutanaccompanyingacademicarticle.Themainobjectiveofthesecasestudieswastocreatepublicawarenessof

theissueoftaxavoidance.Aswillbecomeapparent,adetailedanalysisofthequalityofthecasestudiesshowssubstantial

differencesbetweenthethreedifferentsubgroupsofcasestudies.Table2showsthekeyelementsofthe14selectedcase

studies.

3.2. Developmentofaparticipatoryquantitativecasestudymethodology

Oneofthemainchallengesofthepresentpaperwastodevelopamethodtoassesscasestudies.Mostofthecasestudies

analyzedinthepresentcontextclaimedtobeintendedaseitherpolicydocumentsforadvocacyoraformofinvestigative

journalism,andarenotintendedasacademicstudies.Thisraisesanumberofquestions.Isitevenpossibletodevelopa

uniformsetofqualitycriteria?Isqualitynotsociallyconstructedincontext-specificways?Afterall,theintendedfunctions

and audiencesofthecasestudiesvary.Thedifferencesinintended functionsoftheselectedcasestudiesarerelevant

whenconsideringthequalityofcasestudies.Forexample,anNGOmightselectitscasenotonacademicgrounds,suchas

representativeness,butbecauseofthescopetopropel“emancipatoryaction”.1

Weare,however,oftheopinionthatitisjustifiedandimportanttodevelopqualitycontrol.Thesecasestudieshad

asubstantialimpact(seesection2.3)onthepublicdebateandaimedtoinfluencepolicy.SeabrookeandWigan(2016)

demonstratedthatinternationaltaxdebateexperts,suchasNGOs,canbeeffectiveinthisregardiftheycancombineclaims

tomoralauthoritywithdemonstrationsofexpertise.Theaimofdevelopingthismethodologyforassessingthequalityof

casestudies,istoenableexpertstostrengthentheir(demonstrationsof)expertisefurtherandcontributemoreeffectively

toemancipatoryaction.

Themethodologydevelopedinthisstudycombinesaquantitativeandaqualitativeapproachtoassessthequalityof

casestudies.ItisquantitativeinthesensethatitcriticallyexaminesquantitativestatementsinreportsfromNGOs,butdoes

notinvolveinterviewsorparticipantobservationasaresearchmethod.Themethodologyisqualitativesinceittakescase

studiesasthestartingpointandthenaimsforanin-depthunderstandingthereof.

3.3. Developmentofgenericqualitycriteria

Itisnecessarytoassesseachcasestudyintheappropriatelight,whileatthesametimeprovidingananalyticalframework

thatallowsforacomparisonbetweenthesecasestudies.Inordertoachievethisdelicatebalance,thepresentauthors

designedasetofqualitycriteriawhichare,firstly,basedon(A)thestandardsthattheNGOsthemselveshavesetintheircodes

ofconduct(Partos,2012;StichingOnderzoekMultinationaleOndernemingen(SOMO,2006).Secondly,thesestandardsare

6 A.F. Gunn, D.-J. Koch and F. Weyzig / Journal of International Accounting, Auditing and Taxation 39 (2020) 100318 Table1

Selectionofcasestudies.

Researcher Year Title Researched

companies Selection criterion1:tax avoidance(1= yes/0=no) Selection criterion2: Netherlands (1=yes/0=no) Selection criterion3: publication date(1=yes/0 =no) Selection criterion4: quantitative estimation(1= yes/0=no) Selected?(1= yes/0=no) Non-GovernmentalOrganisations

ActionAid 2015 Anextractiveaffair.

Paladin 1 1 1 1 1 AssociatedBritish Foods 1 1 1 1 1 SABMiller 1 1 1 1 1 FNVandSOMO 2017 Verborgen belastingpraktijkenvan Nederlandsebedrijven [Hiddentaxpractices ofDutchfirms] BCD 1 1 1 1 1 G-Star 1 1 1 1 1 Aegon 0 1 1 1 0 AholdDelhaize 1 1 1 0 0 GlobalWitness 2014

Glencoreandthe Gatekeeper.Howthe world’slargest commoditiestrader madeafriendof Congo’spresident$67 millionricher Fleurette 0 0 1 1 0 Glencore 0 0 1 1 0 ERNC/ERG 0 0 1 1 0

GlobalWitness 2016 OutofAfrica

Fleurette 0 1 1 1 0

Glencore 0 0 1 1 0

ERNC/ERG 0 0 1 1 0

Oxfam 2017 MakingTaxVanish RB 1 1 1 1 1

SOMO 2015 Fool’sgold EldoradoGold 1 1 1 1 1

SOMO 2013 BelastingOntwijkenin

TijdenvanCrisis [Dodgingtaxesin timesofcrisis]

Energiasde Portugal

1 1 1 1 1

TNIandStop

Wapenhandel 2015

Taxevasionand weaponproduction: Mailboxarms companiesinthe Netherlands LockheedMartin 1 1 1 0 0 Boeing 1 1 1 0 0 BAESystems 1 1 1 0 0 GeneralDynamics 1 1 1 0 0 Northrop Grumman 1 1 1 0 0 AirbusGroup 1 1 1 0 0 United technologies 1 1 1 0 0 Finmeccanica 1 1 1 0 0

A.F. Gunn, D.-J. Koch and F. Weyzig / Journal of International Accounting, Auditing and Taxation 39 (2020) 100318 7 Table1(Continued)

Researcher Year Title Researched

companies Selection criterion1:tax avoidance(1= yes/0=no) Selection criterion2: Netherlands (1=yes/0=no) Selection criterion3: publication date(1=yes/0 =no) Selection criterion4: quantitative estimation(1= yes/0=no) Selected?(1= yes/0=no) Politicalorganizations TheGreens/European FreeAlliance

2016a Taxshopping:

ExploringZara’stax business

Zara 1 1 1 1 1

TheGreens/European FreeAlliance

2016b ToxicTaxDeals BASF 1 1 1 1 1

TheGreens/European FreeAlliance

2016c Ikea:Flat-packtax avoidanceTAAKS AVOYD

Ikea 1 1 1 1 1

Academics

Finér,L.,&Ylönen,M. 2017 Tax-driven_{chains:Amultiple}wealth_case studyoftaxavoidance intheFinnishmining sector

AgnicoEagleMines 1 1 1 1 1

FQM(Inmet)Pyhäsalmi 1 1 1 1 1

Ylönen,M.,&M.Laine 2015 Forlogisticalreasons only?Acasestudyof taxplanningand corporatesocial responsibility reporting StoraEnso 1 1 1 1 1 Other Profundo 2016 Taxavoidanceby miningcompaniesin developingcountries -Ananalysisofpotential Dutchpolicyinitiatives

AsiaCoalEnergy Ventures

1 1 1 0 0

BumiResources 1 1 1 0 0

CentralAsiaMetals 1 1 1 0 0

ChinaInvestment Corporation 1 1 1 0 0 EurasianResources 1 1 1 0 0 Glencore 1 1 1 0 0 GoldFields 1 1 1 0 0 NewmontMining 1 1 1 0 0 PeabodyEnergy 1 1 1 0 0 RioTinto 1 1 1 0 0 Trafigura 1 1 1 0 0

8 A.F. Gunn, D.-J. Koch and F. Weyzig / Journal of International Accounting, Auditing and Taxation 39 (2020) 100318 Table2

Overviewoftheselectedcasestudies.

Publisherorjournal Researcher Year Title Researchedcompanies Assessedestimate

Cluster1:politicalpartyreport

Greens/EFA M.TataretBatalla 2016a Taxshopping:Exploring

Zara’staxbusiness

Zara Intheperiod2011−2014,thefirmavoidedat leastEuro(EUR)585Mbyreportingcertain profitsintheNetherlands,Ireland,and SwitzerlandinsteadofSpain.

Greens/EFA M.Auerbach 2016b Toxictaxdeals BASF IntheperiodFY2010−2014,EUR923Mtax

avoided,includingEUR73Mduetoavoidance ofGermandividendtaxviaDutchholding,EUR 178MduetoDutchhybridloan,EUR72Mdue toDutchdeductionsforunrealizedcapital losses,EUR376Mduetointra-grouptrading profitsshiftedtolow-taxbranchesinPuerto RicoandSwitzerland,andEUR202Mdueto Belgiannotionalinterestdeduction.

Greens/EFA M.Auerbach 2016c Ikea;flat-packtax

avoidanceTAAKSAVOYD

Ikea ApproximatelyEUR1billionintaxavoidance inEUcountriesfortheperiod2009−2014 throughlicensefeespaidtoDutchentityfor trademarkownedinLiechtenstein(upto2012) andthentheNetherlands(from2012).The sectiononIKEAGroup’sBelgianloanstructures isnotassessedbecauseitdoesnotestimatean amountoftaxavoidedduetomissingdata.

Cluster2:NGOreport+academicpublication

CriticalPerspectiveson

Accounting L.FinérandM.Ylönen 2017

Tax-drivenwealth chains:Amultiplecase studyoftaxavoidance intheFinnishmining sector

AgnicoEagleMines EUR10.7MtaxavoidedinFinlanddueto excessiveinterestpaymentsandEUR2.7M duetoavoidanceofdividendwithholdingtax intheperiod2002−2014

FQM(Inmet)Pyhäsalmi EUR20.7MtaxavoidedinFinlanddueto excessiveinterestpaymentstoLuxembourg andEUR29.7duetoavoidanceofdividend withholdingtaxviatheNetherlandsinthe period2002−2014.

CriticalPerspectivesonAccounting M.YlönenandM.Laine 2015 Forlogisticalreasonsonly? Acasestudyoftax planningandcorporate socialresponsibility reporting

StoraEnso Intheperiod2005−2010,StoraEnsoavoided EUR47MtaxinFinlandbyroutingpurchases ofpulpfromBrazilthroughalow-taxedDutch entitywithataxruling(whilethepulpwas shippedfromBraziltoFinlandviaBelgium).

A.F. Gunn, D.-J. Koch and F. Weyzig / Journal of International Accounting, Auditing and Taxation 39 (2020) 100318 9 Table2(Continued)

Publisherorjournal Researcher Year Title Researchedcompanies Assessedestimate

Cluster3:NGOreportalone

ActionAid A.Dahlbeck 2015 Anextractiveaffair Paladin MalawilostUSDollar(USD)43Mrevenuein

theperiodFY2008−2014,including15.6Mdue tolowermineralroyalties,7.3Mdueto avoidanceofinterestwithholdingtaxviathe Netherlands,and20.2Mduetoavoidanceof withholdingtaxonmanagementfeesviathe Netherlands.

ActionAid M.Lewis 2013 Sweetnothings AssociatedBritishFoods USD17.7MwithholdingtaxavoidedinZambia

intheperiod2007–2012,includingUSD7.4M duetodividendroutedviaaDutchcooperative, USD7.4MduetoservicesfeespaidtoIreland, andUSD3MduetointerestpaidviaIreland.

ActionAid M.HearsonandR.Brooks 2010 Callingtime SABMiller TaxavoidancebythecompanyinAfricaand

IndiaofasmuchasGreatBritainPound(GBP) 20Mperyear,mainlyconsistingofGBP10Min AfricaduetoroyaltiespaidtotheNetherlands, andGBP9.5MinAfricaandIndiadueto managementfeespaidtoorviaSwitzerland. OxfamGB(author:O.Pearce) M.Curtis,L.Finér,andM.

Ylönen

2017 Makingtaxvanish RB AroundGBP200millionintaxavoidance

worldwidefortheperiod2014−2016, includinguptoGBP60millionindeveloping markets.

SOMO V.KiezebrinkandK.

McGauran 2017

Verborgen

belastingpraktijkenvan Nederlandsebedrijven [Hiddentaxpractices ofDutchfirms]

BCD Intheperiod2011−2013,BCDavoidedatleast EUR15.3McorporatetaxintheNetherlands throughintra-groupinterestpaymentsto Curacao.

G-Star Intheperiod2011−2014,G-staravoided approximatelyEUR3Mtaxperyearinthe Netherlandsthroughtaxdeductionsfor intangibleassets(withoutacorrespondingtax chargeelsewhere).

SOMO I.Römgens,I.Hartlief,R.

vanOs,andK.Mcgauran

2015 Fool’sgold EldoradoGold TheGreekgovernmentlostoverEUR0.7Min

withholdingtaxesduring2009−2013and/or EUR1.7Mincorporatetaxduring2012−2013 becauseofinterestpaymentsto/viathe Netherlands(thesetwofigurescannotbe addedup).

SOMO R.Fernandez,K.McGauran,

andJ.Frederik

2013 BelastingOntwijkenin TijdenvanCrisis[Dodging taxesintimesofcrisis]

EnergiasdePortugal(EDP) Intheperiod2008-mid2012,beforethe advancepricingagreementwaswithdrawn, EDPreportedmorethanEUR140Mprofitsin theNetherlandsandpaideffectively5.36% corporatetaxontheseprofits,resultinginatax benefit;whentheagreementwaswithdrawn, thetaxchargeoverthisperiodretroactively increasedbyEUR34.7M.

10 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

combinedwith(B)somemoregeneralscientificresearchnormsbasedamongotherthingsonPopper’sfalsifiabilitypremise (Popper,1959),Kuhn’snotionontheimportanceofexpertreviews(Kuhn,1996),andthenotionof‘constructvalidity’

(Trochim,2006).

3.4. Developmentofdetailedappraisalguidelines

Thethreeareasofcriticismmentionedabovearebrokendownintotenindicators.Thetenindicatorsaresubdividedinto

23questions(seeTable3).Detaileddescriptionsweredevelopedforeachpossiblescore(0,0.5,or1pointforeveryquestion)

tofacilitateanequitableassessmentmethodforallthereportsandarefoundintheAppendix.The14casestudieswere

analyzedusingthesequalitycriteria.

Oncethefirstversionoftheresearchmethodologyhadbeendeveloped,itwastestedandfurtherdevelopedinthree

ways.First,themethodologyandtheappraisalguidelinesweresharedwiththeresearchersandtheorganizationsthatwrote

thecasestudies.Basedontheirresponses,themethodologywasrefined.

Second,theevaluativequestionswereansweredbytwoofthepresentauthorsindependentlyforallthecase

stud-iestodeterminewhetherthere wereanydiscrepancies betweenthetwo authors.Assomediscrepancieswerefound

betweentheappraisals,thescoringguidelinesweremademoredetailed,anditwasdecidedthatalltheremaining

stud-ieswouldcontinuetobeassessedbythetwoauthorsindependentlyfromeachother.Thedifferenceswerereconciled

afterward.

Third,around-tablediscussionwasorganizedwithfiscalexpertswithdiversebackgrounds(e.g.academics,tax

profes-sionalsworkinginacorporateenvironment,civilservants,andexpertsaffiliatedtoNGOs)todiscussthemethodologyof

thispresentpaperandourinitialfindings,andtodevelopmeaningfulrecommendations,whicharepresentedinsection

4.Basedonthisround-tablediscussion,thepossibilityofanintermediatescoringof0.5withcorrespondingguidancewas

addedforcertainquestions(4b,7a,and9c)whichwereinitiallyscoredas1or0.Furthermore,aquestionwasaddedabout

thelegalanalysisoftheapplicabletaxrulesinthecasestudy(8c).

3.5. Appraisalofthequalityofcasestudiesandverificationofthefindings

Aftertheinitialappraisals,thefindingsweresharedwiththeresearchersandtheorganizations,whoprovidearesponse.

Sendingdraftfindingstothesubjectsofresearchandrequestingtheirfeedbackonthesefindingsisincludedasaquality

assurancestep,inlinewithMalschandSalterio(2016),p.13).Nearlyalltheresearchersresponded.

Insomeinstances,theresearchersnuancedsomeelementsoftheirappraisal.Oneresearcheralsorespondedby

shar-ingacomprehensivemethodologicalbackgroundreportandalltheunderlyingcompanyaccounts.Whensuchadditional

informationwasverifiableorcredible,andthequalityappraisalguidancedidnotrequirethattheinformationwaspublicly

accessibleorincludedinthecasestudyitself,thescoringwasreviewed.Inafewcases,thishadasubstantialimpactonthe

assessmentofthecasestudyandthisprovedanimportantfindinginitself:sometimesthequalityofacasestudyishigher

thanitseemsatfirstsight,sincesomecasestudiesomitdetailsaboutthemethodologyandresearchprocess.

4. Appraisalofcasestudyquality

Table4providestwotypesofinformation.Ontheleft-handside,itprovidesinformationregardingtheaveragescores

foreachofthetenqualitycriteria–brokendownpersubgroupofresearch.Thehighlightedelementsarethosethatscore

0.5orbelow(whenroundedtoonedigit),andwhich,therefore,indicatethatthereareconcernsaboutthequalityofthe

researchontheseelements.Forexample,theaveragescoreforthesevencasestudieswrittenbyNGOs(listedinthecentral

column)forthecriterion‘accesstodata’(firstrow)is:0.3.Forthiscriterion,whichconsistedofjustonequestion,noneof

theNGOcasestudiesscoreda1,fourscoreda0.5,andthreeofthema0:whichcomestoa0.3scoreforthiscriterion.Ifa

criterionconsistedofmultiplequestions,thescoresonthequestionswereaveraged(equalweight).Thislowoutcomeisin

ourviewproblematic.Thesecondtypeofinformationisontheright-handsideofthetable,namelytheaveragescoresfor

eachofthe23questions.Theaverageiscalculatedbytakingthemeansofthe14casestudiesbrokendownperquestion.

Belowarethemostsalientfindingsperqualitycriteria.

4.1. Lackofadequatedata

4.1.1. Accesstodata

Regardingtheaccesstodata,mostofthecasessurveyedsignaledlimitationsintermsofaccesstorelevantdata.An

importantsourceofinformationistheannualaccountsofindividualsubsidiariesofMNEs.Suchaccountsaremore

read-ilyavailableinsomejurisdictions(e.g.theNetherlandsandLuxembourg),thaninothers(e.g.SwitzerlandandCuracao).

Therefore,anNGOmayhaveaccesstoinformationforonepartofthestructure,butnotforanotherpart.

4.1.2. Qualityofdata

Asregardsthegatheringofinformation,inmostofthecasessurveyed,theauthorsofthecasestudiesindicateddifficulties

A.F. Gunn, D.-J. Koch and F. Weyzig / Journal of International Accounting, Auditing and Taxation 39 (2020) 100318 11 Table3

Keyquestionsforqualityassurance.

Areaofcriticism Appraisalcriteria Qualityappraisalquestions

LACKOFADEQUATEDATA 1.Quantityofdata 1aDidtheresearchershaveaccesstothedatathatwereneeded?

2.Dataquality 2aWerethereissueswiththequalityofdata?

LACKOFOBJECTIVITY

3.Transparencyonmethods,techniques,anddata 3aArethemethodsandtechniquesexplicit?

3bArethe(underlying)dataavailableinanaccessibleformat?

4.Subjectsofstudyhadrighttoreplyandresponsesintegrated

4aDidthecompanyhavetherighttoreply?

4bDidthecompanygetenoughtimetoreply?(2weeks) 4cWasthereplyintegrated?

4dWasthereplypubliclyavailable?

5.Peerreviewbyindependentexperts 5aWastheresearchreviewedbyindependentexperts? 6.Conclusionsandrecommendationsarebasedontheresearch

6aAretheconclusionslimitedtotheresearchitself? 6bArelimitstotheresearchmadeexplicit?

6cWeretherecommendationsalogicalfollow-upfromthefindings?

LACKOFEXPERTISE

7.Selectionofaparticularcasestudyexplained 7aWasitexplainedhowthecasestudywasselected? 7bWastherepresentativenessofthecasestudydiscussed? 8.Definitionofconceptsandlegalanalysis

8aWerethekeyconceptsdefined?

8bWerekeyconceptsusedinaccordancewithcurrentterminology? 8cDoesthecasestudyincludealegalanalysisofapplicabletaxrules?

9.Qualityofcalculationsandcounterfactualoftaxavoidanceestimate

9aWerethecalculationscorrectandsufficientlysubstantiated? 9bWasthecounterfactualmadeexplicit?

9cWasthecounterfactualreasonableandcomplete?

9dDidtheresearchershaveanexplicitandcorrectstrategyhowtodealwithmissingdata? 10.Qualitycontrolandrobustness 10aDidtheresearchersindicateadegreeofuncertainty?

12 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 Table4

Resultsofappraisalofcasestudies.

*1indicatesgood,0.5indicatesaverage,and0indicatespoor.

subsidiarycompanieswithintheMNE.Consequently,formoststudiesthescorewithrespecttothedataqualityandquantity is0.5orbelow,indicatingthatdatalimitationsaremajorelementsnegativelyinfluencingthequalityofthesestudies.Looking atcasestudiesthatwerenotselected,thedatalimitationsreportedareevenmoresubstantial,becausecaseswiththemost severedataconstraintstypicallydidnotallowforquantitativeestimates(seeTable1).Moreover,theauthorsofcasestudies

writtenbyNGOsindicatedthattheyhadsometimesexploredthepossibilityofresearchingothercompaniesthantheMNE

ultimatelyanalyzedinthecasestudies,butthatthishadnotresultedincasestudiesforpublicationastheydidnothave

A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 13

accesstodatacannotbeblamedontheresearchers,astheyaimedtofindcasestudieswithenoughdata.Instead,thedata

constraintslimitthepossibilitiesforhigh-qualitycasestudies.Basedonourresearch,theauthorsofcasestudiesseemto

havepursuedcasesdespitetheexistenceofdataissues,forthesimplereasonthatitisexceptionaltofindexamplesoftax

avoidancewithoutanyproblemsregardingdataaccessandquality.

4.2. Lackofobjectivity

Thesecondsetofindicatorsusedtoanalyzethequalityofthecasestudiesconcernstheissueofobjectivityinabroadsense,

includingopennesstoexternalscrutiny.Overall,thescoresarerelativelyhigh.Ofthethreeconstrainingfactorsidentifiedin

ourmethodology,apotentiallackofobjectivityhasthesmallestimpactontheresearchqualityoftheselectedcasestudies.

Still,someelementsleaveroomforimprovement,especiallyasregardstherightofreplyforcasestudiespublishedbya

politicalparty.

4.2.1. Transparencyonmethods,techniques,anddata

Thefirstindicatorofobjectivityconcernsthetransparencyoftheresearch,allowingreaderstoascertainwhetherthe

datawasusedandinterpretedinanunbiasedway.Itconsistsoftwoquestions:(1)Arethemethodsandtechniquesmade

explicit?(2)Arethe(underlying)dataavailableinanaccessibleformat?Thelatterquestionisabouttheprimarydataused

andwhethertheseareinthepublicdomainorobtainableonrequest.Allofthecasestudiescontainedatleastanindication

ofthemethods,techniques,anddataused,althoughtheywereincorporatedinvariousdifferentways.Examplesofthis

includeincorporationinthemaintextofthecasestudyreport,asatable,infootnotes,orasaspecialsection.Sometimesthe

fullcalculationscanbeobtainedonrequest.Thelevelofdetailvariedconsiderably.Mostreportsincludedafulldescription

ofmethodsandtechniques.Somereportsincludedanexplanationofonlysomespecificaspectsofthecalculation.Allofthe

casestudiessurveyeddocumentedtheirsources.Sometimesthesources(primarymaterials)wereavailabledirectly,which

greatlysimplifiedtheprocessofunderstandingthecasestudies.Sometimesthedatacanbeprovidedonrequest.Accessto

informationcanbelimitedincasesinwhichtheNGOprovidesalinktoawebsiterequiringpayment.

4.2.2. Righttoreply

Thesecondindicatorofalackofobjectivityistherighttoreplyforthesubjectsofcasestudies.The“righttoreply”is

understoodtomeanthepossibilityforacompanytohaveinsightsintoacasestudybeforepublicationandtohavethe

possibilityofsubmittingadditionalinformation.Therighttoreplyisassessedusingthefourfollowingquestions:(1)Did

thecompanyhavetherighttoreply?(2)Ifacompanyhadtherighttoreply,didithaveatleasttwoweekstoreply?(3)

Wasareplyintegratedintothecasestudy(ifthisispossible)?(4)Isthereplypubliclyavailable?Elevenofthe14case

studiesreferredtotherighttoreplyorindicatedthatthecompanieswereapproachedforareaction.Interestingly,thethree

instancesinwhichacasestudydidnotrefertotherighttoreplywereallthecasestudiespublishedbytheGreens/EFA.It

ispossiblethattheauthorsworkingforthispoliticalpartydidnotaskforareplybecauseitsreportswereonlyintendedto

drawattentiontoabroaderproblem,illustratedbythecasestudies,andtomovethatproblemupthepoliticalagenda.Itwas

noticeablethatnoneofthestudiesincludedacleardeadlinewithinwhichacompanyshouldrespond.Weidentifiedtwo

casesinwhichtheamountoftimeallowedforareactionseemedtooshort.Thepossibilityofaccessingcompanyresponses

tothirdpartiesvariedamongthestudies.Insevencases,thereportsreferredtoawebsitecontainingtherepliesofthe

companies.By2018,however,thelinksincludedforthreeofthosesevenstudieswerenotoperationalordidnotprovide

therelevantinformation.Theextenttowhichrepliesfromcompanieswereincludedinthecasestudiesdependedontheir

natureandlevelofdetail.Themostimportantobservationinthisregardisthatthemajorityofstudiesthatexpresslydealt

withtherepliesofcompaniesdidnotaddressalloftheissuesraisedinthereplies.

4.2.3. Reviewbyindependentexperts

Thethirdindicatoristheextenttowhichthecasestudieswerereviewedbyindependenttaxexperts.Inthecaseof

academicstudies,thisislikelytobethroughpeerreview.Inothercases,theindicatorisappliedwithreferencetoasuitable

thirdparty.Tenofthecasestudiesexplicitlyreferredtosomeformofexpertreview.Intwoothercases,informationon

anexpertreviewwasprovidedlaterbytheresearchersinresponsetothedraftfindingsofourstudy.Notsurprisingly,the

clusterofcasestudiesthatwerealsopublishedasanacademicpaperscoredhigherthantheotherclusters.

4.2.4. Conclusionsandrecommendationsarebasedontheresearch

Thefourthandfinalindicatorregardingtheissueofobjectivityconcernstheconclusionsandrecommendationsofthe

casestudies.Itconsistsofthreequestions:(1)Aretheconclusionsbasedontheresearchitself?(2)Arethelimitationstothe

researchmadeexplicit?(3)Aretherecommendationsalogicalfollow-upfromthefindings?Infourcases,theconclusions

wereindeedlimitedtotheresearch.However,intheremainingtencasesclaimsweremadethatwerenotsubstantiated

bytheactualresearch.Forexample,almostallofthecasestudiesmentionedthelimitationsofthetheircasestudyintotax

avoidance.Thecasestudies,orthebroaderreportsofwhichtheyareapart,containedabroadvarietyofrecommendations

14 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

manystudies,mostrecommendationsfollowedlogicallyfromthecase,butweregeneralizedtoothercomparablesituations.

However,somestudiesalsocontainedunrelatedrecommendations.

4.3. Lackofexpertise

4.3.1. Explanationofthechoiceforaparticularcasestudy

Thisindicatorconsistsoftwoquestions:(1)Wastheselectionofthecasestudyexplained?(2)Wastherepresentativeness

ofthecasestudyexplained?Thisindicatorshedslightontheexternalvalidityofthecasestudy.Themajorityofthecase

studiessurveyedcontainedsomeexplanationregardingtheselectionofthecase.Insomecases,thetriggerwasaspecific

publicstatementaboutacompany(e.g.astatementintheEuropeanParliament).Inotherinstances,theselectionwaslinked

toamoregeneralpublicoutcryovertaxavoidance.Someofthecaseswerechosenasafollow-uptoapreviousinvestigation

intoaparticularcompany.Theavailabilityofinformationwasalsocitedasareasonforselection.Ourreviewincludedfour

casestudiesthatformpartofbroaderpublications.Suchbroaderpublicationsprovidedagoodoverallexplanationofhow

thevariouscasestudieswereselected,thusleadingtomaximumscoresforthisquestion.Bycontrast,somestudiesdid

notprovideanyexplanationforwhyacertaincasewaschosen.Allacademicarticlesincludedadetaileddiscussionofthe

representativenessofthecases,whereasallotherpublicationscontainedonlyabriefdiscussionor,moreoften,nothing.

Onthespecificpointoftheexplanationofrepresentativeness,thedifferentnatureofacademicpublicationsisconduciveto

higherquality.

4.3.2. Definitionofconceptsandlegalanalysis

Thesecondindicatorconcernstheuseofconceptsandlegalanalysisinthecasestudies.Itconsistsofthreequestions:

(1)Arethekeyconceptsdefined?(2)Aretheyusedinaccordancewiththecurrentterminology?(3)Doesthestudyinclude

alegalanalysisoftheapplicabletaxrules?Theterm“currentterminology”referstointernationalusage,forexample,inthe

contextoftheOECD’sworkontaxation.Thisquestionwasincludedtoflaginstanceswhereusageincasestudiesappears

nottobeinlinewiththeusagewhichmightbeexpectedfromaninternationaltaxperspective.Forallthreequestions,the

reportspublishedbytheGreens/EFAoverallscoredmuchlowerthantheotherstudies.Almostallcasestudiespaidspecific

attentiontotheusageoftechnicalterms.Moststudiesdefinedorexplainedthekeyconceptinthemaintext,boxes,or

footnotes,sometimeswithillustrations.AfewNGOreportsalsoincludedaseparateglossary.Inthemajorityofcases,the

conceptsusedinthecasestudiessurveyedarefullyinlinewiththecommonusageofthesetermsinthefieldoftaxlaw.

Mostcasestudiesprovidedeitheracomprehensiveanalysisoftheapplicabletaxlegislationornoneatall.

4.3.3. Calculationsandcounterfactualoftaxavoidanceestimates

Thethirdindicatorconcernsthetechniquesappliedtothequantificationoftaxavoidanceusedbytheauthorsofthecase

studiesbyNGOs.Whereasthepreviousindicatorfocusedonlegalaspects,thisindicatorconcentratesoneconomicaspects.

Thequalityofthequantitativetechniqueswasassessedusingthefollowingfourquestions:(1)Arethecalculationscorrect

andsufficientlysubstantiated?(2)Isthecounterfactualmadeexplicit?(3)Isthecounterfactualreasonableandcomplete?

(4)Dotheresearchershaveanexplicitandcorrectstrategyfordealingwithmissingdata?Therewasvariationbetweenthe

casestudiesregardingthemethodsusedforquantifyingtax.Itwasnoticeablethatinalimitednumberofcases,itwasnot

possibletoreconstructthecalculationsmadebytheresearchers.Certaincasestudiescontainedminorerrorsincalculation

orfollowedaquestionablemethodologyforassessingtaxavoidance.Insomecases,assumptionsweremadethatcanbe

considereduntenable.Surprisingly,inonecase,theresearchersmadeacalculationerrorthatsuggestedalessbeneficialtax

outcomeforthecompany.Nearlyallthecasestudiescontainedaclearcounterfactualforeachtaxavoidanceelement.Only

intwocasesthecounterfactualwasnotpresentedclearlyorwasclearonlyforpartofthecalculations.Overall,forthecluster

ofreportspublishedbytheGreens/EFA,thequalityofthecounterfactualsismuchlower.Almostallcasestudiessurveyed

hadanexplicitandcorrectstrategyfordealingwithmissingdata.

4.3.4. Qualitycontrolandrobustness

Thefourthandfinalindicatoroftheexpertiseoftheresearchersisthequalitycontrolofthecasestudies.Thisindicator

wasconsideredwithreferencetothedegreeofuncertaintymentionedinthecasestudiesandthetriangulationofdatawith

othersources.Thedegreeofuncertaintycanbebasedonalternativeassumptionsorcalculationmethods,thusinvolving

robustnesschecksthatusedifferentspecificationstoestimatetheamountoftaxavoidance.Althoughthecasestudiesdidnot

useeconometricmodelsthatcouldberunwithalternativespecifications,somestudiescontainedcalculationsthatcouldbe

performedusingalternativeassumptions.However,sixcasestudiespresentedanamountwithoutanysuchqualifications.

Noneofthecasestudiessoughttoquantifyanexactmarginoferror.Theeffortstotriangulatedataforonecasewererather

impressive,involvinginterviewswithgovernmentofficialsinvariouscountriesandevenundercoverresearchtoinvestigate

theprovisionofintra-groupservices.Someresearchersdescribedtheirattemptstocollectadditionalinformationfrom

othersources,butstatedthattheywereunsuccessful.Onlyafewcasestudiesdidnotmentionanyattemptstoobtainother

A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 15 5. Discussionandconclusion

5.1. Discussiononandrecommendationsforthequalityofthecasestudies

Whereasallthefactorshadanimpactontheoverallquality,thedataavailabilityanddataqualityaretheonlyfactorsthat

scoredanaveragebelow0.5whenallthestudieswerecombined(0.45).Objectivityscoredthehighest,withanaverageof0.68

andnoneoftheindividualindicatorsscoringonaveragebelowthe0.5mark.Thelackoftechnicalexpertisedemonstrated

inthestudiesliessomewherebetweenthescoreswithrespecttodataandobjectivity(0.63).Thisdoesnotmeanthatno

improvementscanbemadetoincreasetheobjectivityofthecasestudies.Itisrecommendedthatallresearchersprovide

companieswiththeopportunitytorespondandgivethemenoughtimetodoso.Itisalsosuggestedthatresearchersmake

theresponsesfromcompaniesavailableandshowhowtheyhaveincorporatedthemintotheircasestudies.

Theindividualindicatorsthatscoredbelow0.5onaverageacrossallthetypesofstudiesarethedataavailability,the

discussionontheselectionofcasestudies,andtheindicationofadegreeofuncertainty.Morescoreswouldbebelow0.5

iftheywerebasedononlypublicinformation.Therefore,itisrecommendedthatresearchersmakemoreofthedataand

methodologyavailabletothepublic,suchasinseparatebackgroundpapers.Thethreescoresthatarebelow0.5arethe

focusofourrecommendations.

Regardingtheavailabilityandaccessibilityofdata,bothgovernmentsandcompanieshavearoletoplay.Thequality

ofthecasestudiescouldlikelybestrengthenedifmoreandbetterinformationwereavailableandmoreeasilyobtainable

byresearchers.Crucially,governmentscouldrequirethefilingofunconsolidatedannualaccountsofindividualentities,

withexplanatorynotes,andensurethatthesecanbeobtainedeasilyforfreeoratalowcost(Koch,2017).Thisappliesto

allcountries,includingdevelopingcountriesaswellascountriesusedincorporatetaxplanningstructures.Forcountries

wherethisisusuallythecase,suchastheNetherlandsandLuxembourg,thedataavailableforthecasestudiesisrelatively

good.Forcountrieswheresuchfilingsarenotavailable,suchasSwitzerlandandCuracao(whichsurfacedregularlyinthe

selectedcasestudies),thedataavailabilityishighlyproblematic.

Furthermore,companiesthemselvescouldpublishkeydata,suchasrelevantfinancialfiguresonacountry-by-country

basis.TheGlobalReportingInitiative(GRI),whichincludessustainabilityreportingstandardsthatarewidelyused,

pro-videssomeguidanceforthisintheguiseofanewtaxstandard.Moreover,recentexamplessuchasShell’spublicationof

comprehensivecountry-by-countrydatafor2018demonstratethatitisfeasible.Companiescouldalsoproviderelevant

explanationsabouttransferpricing,groupfinancing,intellectualproperty(IP),theuseoftaxtreaties,thelevelof

trans-parencywithforeigntaxauthorities,andtheuseoftaxholidays,taxrulings,andadvicebytaxadvisors(e.g.positionpapers,

memoranda,andlegalopinions).Inallcases,thepublicationofsuchinformationneedstobedoneinaformatthatis

acces-sibleandcomprehensibletoresearcherswithoutaspecificbackgroundininternationaltaxation.Wenotethatthenatureof

theinformationneededwilldependontheresearchobjectivesofaparticularstudy.Combinedwithcompanies’legitimate

concernsaboutconfidentialityandissuesofefficiency,agoodapproachtoincreasingtheavailabilityofdataisfor

compa-niestoprovideinformationonrequest,allowingtheresearcherstoaskquestions.Thiswouldalsoincreasetheefficiency

oftheresearchprocess.Allcasestudieswouldhavebenefitted,tovaryingdegrees,frommoreandmorereadilyavailable

information.

Regardingtherepresentativenessofthecasestudies,researcherscanimprovetheirapproach.Whileitcanbelogicalfor

researcherstoselectcasesbasedontheiremancipatorypotential,insteadoftheirrepresentativeness,theycouldmakemore

effortstodescribeanddocumentthedegreetowhichtheselectedcaseisexceptionalorcouldbeexemplaryofabroader

situation.Forexample,thiscouldbeachievedbysystematicallycomparingtheglobaleffectivetaxrate(ETR)withthatof

peercompanies.AcomparisonofthetotaltaxesbornebyanMNEcouldalsocontributetoabetterunderstandingofthe

taxpositionofaparticularMNErelativetoitspeers.Furthermore,wenotethatabreakdownofdifferenttypesofrelevant

taxescouldcontributetoamorenuancedanalysis.Forexample,inthecaseoftheextractiveindustry,itwouldbehelpfulto

distinguishcorporatetaxesontheannualprofitfromextractivetaxesleviedovertheexploitationoffinitenaturalresources.

Thiscouldenhancetheexternalvalidityofthecasestudy.

Lastly,regardingthedegreeofuncertaintyofcasestudyfinding,moreusecouldbemadeofalternativeestimatesto

highlightthesourcesofuncertaintyandindicatearangeoftaxrevenuesforgoneratherthanasinglefigure.Thiswouldalso

reducetheriskofmisinterpretationorexaggerationfromtreatinganestimateasahardfigure.

Inshort,forthequalityofthecasestudiestobeimproved,thecompaniesandthegovernmentsareinthedrivingseat.

Nevertheless,NGOscouldthemselvesbemoreaccountableandtransparentabout,forinstance,themethodsanddatathey

used.Insteadofalwaysbeingpubliclyavailable,sometimesthisinformationwasonlyavailableonrequest.

5.2. Limitationsofthecurrentresearch

Inourstudy,wewereabletoreachseveralconclusions.Forinstance,onaverage,thequalityofthecasestudiesbyNGOs

isbetterifcollaborationissoughtwithacademics.Furthermore,thequalityofcasestudiesbyNGOsishigherthanstudiesof

politicalorganizations.Ofcourse,certainlimitationsapplywithregardtotheinternalandexternalvalidityofourfindings

(Trochim,2006).Wesoughttominimizetheadverseimpactoftheselimitationsinavarietyofways.

Becauseoftheiterativemethodologyofourresearch,theinternalvalidityoftheresearchwasstrengthened.The‘feedback

16 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318

couldbeattributedtoalternativeexplanationsthatarenotexploredinthestudy(Girden,2001).Using10indicatorsand

23questions,wesoughttodevelopanencompassingdefinitionofquality.Theriskofpotentialbiaseswasfurtherreduced

becausethepresentauthorscomefromverydifferentprofessionalbackgrounds(includingaformeremployeeofanNGO,

aninternationaltaxspecialistwithabackgroundataBigFourfirm,andaseniorcivilservantwithexpertiseinthefieldof

internationaldevelopment).

Toascertainwhetherastudyhasexternalvalidity,thequestionwhichmustbeanswerediswhetherthefindingsapply

tootherresearchsubjectswhoseplace,times,andcircumstancesdifferfromthoseofthesubjectswhoparticipatedinthe

study(ResearchConnections,2016).Astudy’sexternalvalidityiscloselyrelatedtothegeneralizabilityofthefindings.In

thisregard,wefacelargerproblemsasoursamplesize(N)isrelativelylow(14,andfortwosubsetsofcasestudies,those

performedbypoliticalpartiesandbyacademics,Nisjust3),eventhough-tothebestofourknowledge-ourselectioncovers

alltaxavoidancecasestudieswithaquantitativeestimatepublishedduring2013–2017andincludingaDutchelement.In

2018and2019,newcasestudiesbyNGOswithtaxavoidanceestimateswerepublished,forexamplebyOxfamaboutUS

andAustralianfirms.NotalloftheseinvolveDutchentities.Thisraisesthequestionofwhetherandtowhatextentthe

currentfindingscanbeconfidentlygeneralizedtosuchlatercasestudies.Furthermore,thequestionarisesastothedegree

ofcertaintywithwhichourfindingsregardingthe(lackof)qualityofthestudiesbythepoliticalpartiescanbetransferred

tootherstudiesbypoliticalparties.Wewouldarguethatmodestyisrequiredonthepointofexternalvalidity.Forinstance,

allthreestudiesbypoliticalfoundationsthatweanalyzedwereproducedbythesameorganization,whichreducesthe

robustnessofthefindings.

Onekeywaytoenhancetheexternalvalidity,wouldbetoincreasethenumberofobservations.Forexample,thiscould

beachievedbyassessingmorerecentcasestudies,includingsomewithoutDutchentities.Lackofdatamightthenemerge

asanevenlargerqualityconstraintbecausetheavailabilityofdetailedunconsolidatedfinancialaccountswasrelatively

goodforDutchentities.Mostoftheassessmentframework(indicators1–8)couldalsobeappliedtocasestudieswithout

aquantitativeestimate.Inanycase,thequalityofguidelinesthatimplicitlyfollowfromtheassessmentframeworkalso

seemsvalidfornewcasestudies.

5.3. Potentialforscale-upofcurrentmethodologytoNGOresearchfocusingonotherdomains

NGOspublishcasestudiesonmanydifferentaspectsof(alleged)corporatemalpractice.Thiscanrangefromtheimpact

ondeforestationtoissuesinvolvinglaborconditions,andfromhumanrightsabusestoirregularitiesincorporatesocial

responsibilityreporting.Wenotethatthereislittlesystematicqualityappraisalforresearch,eventhoughourstudyshows

thatmuchcouldbelearnedfromsuchanappraisal.Cantheappraisalmethodologydevelopedbyusalsobeusedtoappraise

thequalityofcasestudiesoutsidethefieldoftaxavoidance?

Letusfirstrecapthekeystepsandfeaturesofthepresentmethodology.Thefirststepwasthedevelopmentofgeneric

qualitycriteria(section3.3),thenextstepwasthedevelopmentofdetailedappraisalguidelines(section3.4),andthelast

stepwastheappraisalofthequalityofcasestudiesandverificationofthefindings(section3.5).Throughoutallsteps,a

participative,inclusive,anditerativeapproachwastaken.Itwasparticipativeduetotheextensiveuseofqualitycriteria

proposedbytheparticipantsofthestudy(theresearchers).Inclusivenesswasguaranteedbythediversityofexpertsinvolved

inthequalitycontrolsystem:fiscalexpertsfromallkeydomainsparticipated(e.g.academics,taxprofessionalsworkingin

acorporateenvironment,civilservants,andexpertsaffiliatedtoNGOs).Theapproachcanberegardedasiterativeasthe

scoringmethodologywasadaptedaftertrialappraisals.

Wethinkthatthecurrentstepsandfeaturesofourmethodologycouldalsobehelpfultoappraisethequalityofcase

studyresearchonotherbusinesspractices.Theycouldalsobelinkedtothefieldofauditsand(socialandenvironmental)

accountingwherethereissubstantialacademicdebateondisclosure(e.g.Elkins&Entwistle,2018)andoncorporatesocial

responsibilityreporting(e.g.(Cai,Lee,Xu,&Zeng,2019)).Themethodologycouldalsopotentiallybemodifiedtoassess

companycasestudiesinotherdomains,suchashumanrightsandenvironmentalissues,whilepreservingitsparticipatory

andinclusivefeatures.Forotherdomains,itmaybenecessarytoverifythekeyfactorsthataffectthe(perceived)research

qualityofcasestudies.However,somefactorsidentifiedappeartohavebroaderapplicabilityandcould,therefore,forma

usefulstartingpoint.

Thedegreetowhich ourgenericqualitycriteriaanddetailedappraisalguidelinescanbetransposedand/orrequire

adjustmenttoapplytonon-fiscalcase studiesdifferspercriterion.Thefourindicatorswithrespecttoobjectivity(i.e.

transparencyonmethods,techniquesanddata;therightofreplyforthesubjectofthecasestudyandtheintegrationofthe

replyinthereport;peerreview;andfindingswhicharelimitedtotheactualresearch)appearimportantinalldomains.

Itseemsreasonabletoexpectthatthetwocriteriarelatingtodata(i.e.dataquantityanddataquality)arealsocrucialfor

othertypesofcasestudies.However,forissuesofamorequalitativenature,suchasharassmentofworkers,itmaybethe

casethatappraisalguidelineswouldneedtoputmoreemphasisonthereliabilityandinterpretationofthedata.Thefour

indicatorsfor‘expertise’(i.e.theexplanationoftheselectionofcasestudies;theinclusionofadefinitionofkeyconcepts

andlegalanalysis;thequalityofcalculationsandcounterfactuals;andqualitycontrol)aremorecloselylinkedtospecifics

ofthetaxavoidancedebate.Fortheappraisalofcasestudiesabouthumanrights,theuseofacounterfactualmaybeless

A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 17

whatextenthumanrightsissuesarestructuralanddirectlyresultfromcompanypoliciesandpractices.Thus,theappraisal

guidelinescanbetailoredtospecificexpectationswithrespecttoresearchqualityinotherdomains.

5.4. Thefutureofcase-studybasedtaxavoidanceresearch

Weaimedtocontributetotheinternationaltaxandaccountingliteraturebyprovidingatooltoanswerquestionsrelating

tothequalityoftaxavoidancecasestudies.Wefoundthattheuseofcasestudiesfrequentlyraisesquestionsregardingthe

(lackof)availabledata.Indeed,theuseofcasestudiesasawayofexploringtaxavoidanceseemstobelinkedprecisely

totheneedtocircumventdataproblems(Finér&Ylönen,2017).Doesthismeanthatcasestudiesontaxavoidancearea

“dead-endstreet”withlittlefuturerelevance?Wewouldargueagainstthis.Whileweareincertainrespectsquitecritical

ofthequalityofthecasestudies,wearehopefulthattheirqualitywillimprove.Amongstotherthings,thisisbecausemore

dataisbecomingavailableonline,forinstance,datarelatedtotheissueofbeneficialownershipofcompanies.Tostimulate

theuptakeofourfindings,weorganizedvariouspresentationsonourmethodologyandthefindingstopotentialusers.

Interestingly,oneorganizationofwhichthereportswereassessedhasstartedintegratingsomeoftherecommendations

resultingfromthisresearch(SOMO,2018).Forexample,theresearchedcompanyreceivedamonthtorespondtothereport,

anditsresponseletterwaspublishedonthewebsiteofSOMO.

Inthispresentpaper,wedevelopedamethodologyandsubsequentlyappliedthatmethodologytoasetoftaxavoidance

casestudiescarriedoutbyNGOsandothergroups.Whiletherewasalreadycriticismofsuchcasestudiesbytheaffected

companiesandsomejournalists,therewasnotyetacademicresearchintothequalityofthesestudies,inpartbecausea

suitablemethodologywaslacking.Bydevelopingcriteriaandguidelineswhichcanpotentiallybeusedtoappraisecase

studiesinotherareasofcorporatebehavior,thisstudyhascontributedtofillingthatgap.Thefindingsoncasestudies

oftaxavoidanceshowwaystoenhancethequalityofresearch,whilerecognizingthatgovernments,companies,andthe

researchersthemselvesallhavearoletoplay.

AppendixA. Supplementarydata

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/

j.intaccaudtax.2020.100318. DeclarationofCompetingInterest

Theauthorsdeclarethefollowingfinancialinterests/personalrelationshipswhichmaybeconsideredaspotential

com-petinginterests:FrancisWeyzighadprovidedtechnicalcommentsonsomeoftheselectedcasestudiespriortopublication,

uponrequestasanexternalexpert.However,theassessmentforthecurrentpaperfollowsastandardizedassessment

frameworkwithdetailedguidanceforscoring.Theanalysisisalsomoreelaborate,inpartbecauseitfocusesonstructures

involvingaDutchlink,whilesomecasestudiesincludeotherstructurestoo.Moreover,theassessmentcoversmanyaspects

beyondtheusualscopeofexternalexpertcomments,suchasavailabilityofcompanyresponses,datadisclosure,casestudy

selection,andrepresentativeness.

References

Accountantweek.(2015).Wiebes:Nederlandisgeenbelastingparadijs[Minister:Netherlandsisnotafisalparadise]. https://accountantweek.nl/artikel/wiebes-nederland-is-geen-belastingparadijs

Cai,W.,Lee,E.,Xu,A.L.,&Zeng,C.(2019).Doescorporatesocialresponsibilitydisclosurereducetheinformationdisadvantageofforeigninvestors? JournalofInternationalAccountingAuditingandTaxation,34,12–29.

Dallyn,S.(2017).Anexaminationofthepoliticalsalienceofcorporatetaxavoidance:AcasestudyoftheTaxJusticeNetwork.AccountingForum,41(4) http://dx.doi.org/10.1016/j.accfor.2016.12.002

EC.(2017).Pressrelease:Stateaid:Commissionopensin-depthinvestigationintotheNetherlands’taxtreatmentofInterIKEABrussels. http://europa.eu/rapid/press-releaseIP-17-5343en.htm

EC.(2019).Stateaid:Commissionopensin-depthinvestigationintotaxtreatmentofNikeintheNetherlands.EuropeanCommissionPressRelease. https://ec.europa.eu/commission/presscorner/detail/en/IP 19322

EC.(2015a).Pressrelease:Stateaid:CommissionopensformalinvestigationintoLuxembourg’staxtreatmentofMacDonald’s.Brussels:EC.

EC.(2015b).StatementbyCommissionerVestageronCommissionstateaiddecisionsregardingillegaltaxadvantagesgrantedbyLuxembourg(Fiat)andthe Netherlands(Starbucks).http://europa.eu/rapid/press-releaseSTATEMENT-15-5881en.htm

Elkins,H.,&Entwistle,G.(2018).Acommentaryonaccountingstandardsandthedisclosureproblem:Exploringawayforward.JournalofInternational AccountingAuditingandTaxation,33,79–89.

EuropeanCommission(EC).(2012).Clampingdownontaxevasionandavoidance.Commissionpresentsthewayforward.PressRelease. http://europa.eu/rapid/press-releaseIP-12-1325en.htm

Finér,L.,&Ylönen,M.(2017).Tax-drivenwealthchains:AmultiplecasestudyoftaxavoidanceintheFinnishminingsector.CriticalPerspectiveson Accounting,53–81.

FleuretteGroup.(2016).Fleurettegroupstatementre:KCCroyaltiesRetrievedfrom:.

http://fleurettegroup.com/media-center/press-release/2016-11-15-fleurette-statement-re-kcc-royalties.pdf

Forstater,M.,&Christensen,R.C.(2017).NewPlayers,NewGame.Theroleofthepublicandpoliticaldebateinthedevelopmentofactiononinternationaltax issues..EuropeanTaxPolicyForumResearchPaper.http://www.etpf.org/files/ETPFPaper-ForstaterandChristensenFinal.pdf:

Girden,E.R.(2001).Evaluatingresearcharticlesfromstarttofinish.London:SagePublications.

Gunn,A.F.(2015).WiezijntochdieNGO’s?[WhoarethoseNGOs?]May9,RetrievedfromArtikel104.nl.https://artikel104.nl/wie-zijn-toch-de-ngos/ Hasseldine,J.,&Morris,G.(2013).Corporatesocialresponsibilityandtaxavoidance:Acommentandreflection.AccountingForum,37,1–14.

18 A.F.Gunn,D.-J.KochandF.Weyzig/JournalofInternationalAccounting,AuditingandTaxation39(2020)100318 Heitzman,M.,&Hanlon,S.(2010).Areviewoftaxresearch.JournalofAccountingandEconomics,50(2-3),127–178.

Kahn,M.(2018).Dutchsetoutplantocountertax-havenreputation.FinancialTimes.https://www.ft.com/content/ef9b87ac-1af5-11e8-aaca-4574d7dabfb6 Koch,D.-J.(2017).Invented,Ignored,Invisble:Theunintendedeffectsofinternationalcooperation.Inauguraladdress:RadboudUniversityNijmegen. Kovermann,J.,&Velte,P.(2019).Theimpactofcorporategovernanceoncorporatetaxavoidance–Aliteraturereview.JournalofInternationalAccounting

AuditingandTaxation,36(100270),1–29.

Kuhn,T.S.(1996).Thestructureofscientificrevolutions.Chicago:ChicagoUniversityPress.

Malsch,B.,&Salterio,S.(2016).“Doinggoodfieldresearch”;assessingthequalityofauditfieldresearch.Auditing:AJournalofPractice&Theory,1–22. OECD.(n.d.).GlossaryofTaxTerms.RetrievedfromCentreofTaxPolicyandAdministration:http://www.oecd.org/ctp/glossaryoftaxterms.htm#T. Oei,S.-Y.,&Ring,D.(2018).Leak-drivenlaw.UCLALawReviewUniversityofCalifornia,LosAngelesSchoolofLaw,65(3),532.

Oxfam.(2016).Taxbattles-thedangerousglobalracetothebottomoncorporatetax.

https://www.oxfam.org/sites/www.oxfam.org/files/bp-race-to-bottom-corporate-tax-121216-en.pdf Paladin.(2015).Paladinenergy’sltdresponsetotheactionaidanextractiveaffairarticleRetrievedfrom:.

https://mininginmalawi.files.wordpress.com/2015/07/paladin - action aid response.pdf

Partos.(2012).Partoscodeofconduct.https://www.partos.nl/fileadmin/files/Documents/Gedragscodev2012.pdf Popper,K.(1959).Thelogicofscientificdiscovery.NewYork,NY:BasicBooks.

PublishWhatYouPayNorway.(2011).Pipingprofits.

http://www.publishwhatyoupay.no/sites/all/files/1005c-Version2PWYPNorwayPipingProfitsDOWNLOAD.pdf

ResearchConnections.(2016).Assessingresearchquality.https://www.researchconnections.org/content/childcare/understand/research-quality.html Seabrooke,L.,&Wigan,D.(2016).Poweringideasthroughexpertise:Professionalsinglobaltaxbattles.JournalofEuropeanPublicPolicy,23(3),357–374. Seay,L.(2015).ConflictmineralsinCongo:Consequencesofoversimplification.InA.Waal(Ed.),Advocacyinconflict-Criticalpespectivesontransnational

activism.London:ZedBooks.Chapter6).

Sikka,P.(2010).Smokeandmirrors:Corporatesocialresponsibilityandtaxavoidance.AccountingForum,34,153–168. Snel,M.(2018).Nieuwerulingpraktijk-briefaantweedekamer.DenHaag:MinisterievanFinanciën.

SOMO.(2006).SomoResearchcodeofconduct..SOMO.https://www.somo.nl/wp-content/uploads/2016/07/Codeofconduct.pdf SOMO.(2018).Miningtaxes.https://www.somo.nl/nl/mining-taxes/

TheGreens/EuropeanFreeAlliance.(2016).Ikea:Flat-packtaxavoidanceTAAKSAVOYDBrussels. https://www.greens-efa.eu/en/article/news/ikea-flat-pack-tax-avoidance/

Trochim,W.(2006).Researchmethodsknowledgebase.https://www.socialresearchmethods.net/kb/concval.php

Wang,F.,Xu,S.,Sun,J.,&Cullinan,C.(2020).Corporatetaxavoidance:Aliteraturereviewandresearchagenda.JournalofEconomicSurveys,forthcoming.

Furtherreading

ActionAid.(2015).Anextractiveaffair..ActionAid.https://actionaid.org/publications/2015/extractive-affair:

Dillard,J.,&Vinnari,E.(2017).Acasestudyofcritique:Criticalperspectivesoncriticalaccounting.CriticalPerspectivesonAccounting,88–109. FNV,&SOMO.(2017).VerborgenbelastingpraktijkenvanNederlandsebedrijven..SOMO.

https://www.somo.nl/nl/verborgen-belastingpraktijken-van-nederlandse-bedrijven-2/rapport-fnv-verborgen-belastingpraktijken-juni-2017/ GlobalWitness.(2014).GlencoreandTtheGatekeeper.Howtheworld’slargestcommoditiestradermadeafriendofCongo’spresident$67millionricher.

https://www.globalwitness.org/documents/17897/glencoreandthegatekeepermay2014.pdf:

GlobalWitness.(2016).Congosignsoverpotential$880mofroyaltiesinglencoreprojecttooffshorecompanybelongingtofriendofcongolesepresident Retrievedfrom:. https://www.globalwitness.org/en/press-releases/congo-signs-over-potential-880m-royalties-glencore-project-offshore-company-belonging-friend-congolese-president/

Hendrix,C.H.,&Wong,W.H.(2014).Knowingyouraudience:HowthestructureofinternationalrelationsandorganizationalchoicesaffectAmnesty International’sadvocacy.TheReviewofInternationalOrganizations,9,29–58.

OrganisationforEconomicCooperationandDevelopment(OECD).(2011).Guidelinesformulti-nationalenterprises.Paris:OECD. Oxfam.(2017).Makingtaxvanish.https://www.oxfam.org/en/research/making-tax-vanish

Profundo,&OffshoreKenniscentrum.(2016).Taxavoidancebyminingcompaniesindevelopingcountries. https://www.rijksoverheid.nl/documenten/rapporten/2016/12/23/tax-avoidance-and-mining-report

SOMO.(2013).Belastingontwijkingintijdenvancrisis..SOMO.https://www.somo.nl/nl/belasting-ontwijken-in-tijden-van-crisis/ SOMO.(2015).Fool’sgold..SOMO.https://www.somo.nl/fools-gold/

StopWapenhandel,&TNI.(2015).Taxevasionandweaponproduction:LetterboxarmscompaniesintheNetherlands..TNI. https://www.tni.org/en/publication/tax-evasion-and-weapon-production-letterbox-arms-companies-in-the-netherlands

TheGreens/EuropeanFreeAlliance.(2016a).Taxshopping:ExploringZara’staxbusinessBrussels.https://www.greens-efa.eu/en/article/tax-shopping/ TheGreens/EuropeanFreeAlliance.(2016b).ToxictaxdealsBrussels.https://www.greens-efa.eu/en/article/news/corporate-tax-avoidance-6465/ Ylönen,M.,&Laine,M.(2015).Forlogisticalreasonsonly?Acasestudyoftaxplanningandcorporatesocialresponsibilityreporting.CriticalPerspectives