WHOSE REGULATORS ?COMPETING PERSPECTIVESON

BUREAUCRATIC AUTONOMY –THE CASEOFTHE DUTCH

COMPETITION AUTHORITY

Papertobepresentedatthepanel“Bureaucraticautonomyandregulatory enforcement:Aremoreautonomousregulatorsalsobetterregulators?”ofthe ECPRGeneralConference,tobeheldatPisa,Italy,6-8September2007 Workinprogress KutsalYesilkagit UniversityofUtrecht UtrechtSchoolofGovernance Bijlhouwerstraat6 3511ZCUtrecht Netherlands +31302538649 [email protected]

1 1.Introduction Regulationafterdelegationcanevolveinprecipitatedways,astheformalinstitutional arrangements,whichcircumscribethestructure,jurisdictionandcompetenciesof independentregulatoryauthorities,“donotdeterminethebehaviorofIRAsandtheir relationshipswithotheractorsbecausepowersandcontrolscanbeused[byelected politiciansandregulates]inmanydiversewaysandinstitutionalframeworksare incomplete”(Thatcher2002:955).Afterdelegation,theformalinstitutionaldesignof agenciesturnsouttobeaweakpredictorofagencybehavior(Krause&xxx).The arrangementswithwhichpoliticiansintendtoexantehardwireorstack-the-deckof regulatoryagenciesprovetobeworkinginunintendedways(Balla1998).Findingsofa numberofrecentstudiesofregulationconcludethatagenciesdevelopdifferentpatterns ofbehaviorandthatotherkindsofinteractionbetweenagenciesandtheirdirect environmentdevelopthanthosethatwereexpectedonthebasisoftheirdesign. Examiningbusiness-regulatoryrelationshipsinGermanyandtheUK,Coenshowsthat aftertheyhavebeencreatedregulatorsinbothworkwithbusinesstodefine theirdomesticregulatoryrolesandcreatetheirowndistinctpoliticalspacefrom governmentpressures”suchthat“theregulatoryrelationshiphasevolvedbeyondthat envisagedintheinitialdelegationofpowerstotheregulator”(Coen2005:394-395,375). ExaminingtheinteractionsbetweenDutchcentralgovernmentalministriesandtheir independentagencies,Yesilkagitcametoasimilarconclusionashefoundthat“[R]ules, procedures,andcompetencesmayseemclearonpaperastheyinformallactorsaboutthe formalintentionsofthedesigners.However,theagencyleadershipcaninterprettherules inadifferentwaythanpoliticianshadinmindwhentheydesignedtherules”(Yesilkagit 2004:535). Oneofthefearsofrulingpoliticiansindemocraticsystemsisthatbecauseofa ‘poor’design,regulatorswilldriftawayfromthegoalssetinlegislation,eitherbecause regulatorsbecomecapturedbythesectororbecausetheydevelopandpursueinterestsof theirown.Delegationofauthoritymaybecomesynonymousforrunawaybureaucratsand thefearoflosingcontrol(Christensen&Laegreid2006).Thegoalofthispaperisto delvedeeperintothedeterminantsof‘real’bureaucraticautonomyofregulatory

2 agencies.Moststudiesonautonomyfocusonthedelegationofformalautonomyandthe formaldesignofagencies(Christensen2001;Epstein&O’Halloran1999;Gilardi2002). Giventhegap,thequestionthispaperiswhatthedeterminantsofinformal agencyautonomyare.Whenandunderwhichcircumstancescanweexpectformally independentagenciestobemoreorlessindependentthanformallyprescribed?Which factorscausethedegreeofinformalorrealautonomyofregulatoryagencies?Orplainly: Whoseinterestsdoregulatorsserveafterdelegation? Thesequestionshavebeenaddressedfromavarietyoftheoreticallensesand models(seeBaldwin&Cave1999),butthemajorityofthesemodelsfallsineitherone ofthefollowingtwolinesofperspectives.Oneperspectiveconsidersthedegreeof bureaucraticautonomyofanadministrativeagencyasafunctionoftheinfluenceexerted bypoliticalactorsintheenvironmentoftheagency.Theorieswithinthisperspective, whichIwillrefertoasthe externalcontroltheories ofbureaucraticautonomy,agencies withthehighestlevelofautonomyhavethelowestlevelofpoliticalconstraintsimposed uponthembytheirprincipals.Theotherperspectiveregardsthehighlevelofautonomy anagencyenjoysasbeingtheresultantofanagency’sorganizationalanditsleadership’s capacitiestoforgeasolidbaseofautonomyofitsown.Highlyautonomousagencies fromwithinthisperspective,i.e. administrativebehaviouraltheories ,arethosethathave managedtocreateawideconstituencybasis,havestrongsenseofidentityandmission, andaclearlydemarcatedaswellascoherentsetoftasks. Tothisend,Iwillexaminetheextenttowhichthebureaucraticautonomyofthe CompetitionAuthorityintheNetherlands( NederlandseMededingingsAutoriteit – NMA)isaffectedbythecontrolinstrumentsofitsprincipalsandtheinternalbureaucratic capacitiesoftheagencyitself.Theprimarypurposeofchoosingthiscaseisthatof replication:towhatextentdothefactorsthatarereportedinUSstudiesashavingan effect(ornot)onthebureaucraticautonomyofregulatoryagenciesperformwithina Europeanparliamentarysetting?Thevariationininstitutionalsettingallowsustoassess whethercontrolandreputationvariableswillperformthesamewithinafundamentally differentinstitutionalsetting(Weaver&Rockman1993;butseeHammond&Butler 2003).Ifthisisthecase,wewouldcomeclosertohavingmoregeneralizable mechanismsthataffecttheautonomyofbureaucraticagencieswithindemocratic

3 systems.Section2presentsthetwocompetingsetsoftheories,followedbysection3 wheretheresearchdesignofthisresearchispresentend.Section4presentstheresults andthepaperendswithaconclusion. 2.TheoriesofBureaucraticAutonomy:TheoriesofExternalControlversus AdministrativeBehavioralTheories ExternalControlTheoriesofBureaucraticAutonomy Theredoesnotexistasingleexternalcontroltheoryofbureaucraticautonomy.Wecan distinguishbetweentheoriesofexternalcontrolthathavetakenanarrowconceptionof whatconstitutes‘external’andtheoriesthathavecasttheirnetsomewhatwider. 1 Theorieswithanarrowconceptiononlytakeintoaccounttheinfluenceofformal politicalinstitutionsonbureaucraticautonomy,notablythePresidentandCongress; studieswithabroaderconceptionalsotakeintoaccounttheinfluenceofothernon- politicalactors,suchasthecourts,interestgroups,andcitizens.Thereadershouldbe awarethatthisclassificationoftheoriesiscontingentonthepoliticalsystemwithin whichtheseweredeveloped.Asthesetheorieshavebeenmainlydevelopedwithinthe US,theclaimsdevelopedbythesetheoriesarehighlyrestrictedtotheUSsystemof separation-of-powersandtheexistentsystemofchecks-and-balancestherein.Therefore,I willdescribethisstrandofautonomyresearchasitasbeendevelopedinUSpolitical science.AfterthisdiscussionIwillbrieflyassesstheirapplicabilitytoparliamentary systems. Externalcontroltheoriesofbureaucraticautonomywithanarrowfocusonhow andtowhatextentformalpoliticalinstitutions,i.e.CongressandthePresidency,have influenceonagencydiscretion.WithintheUS,thesetheoriescorrectedthelong-time dominantviewwithinUSscholarshipthatbureaucracywasoutofcontrolbecauseonlya “smallfractionofresources[were]goingintocongressionaloversight,thehaphazard natureoftheoversightactivitiesthatdidtakeplace,thelackofexpertisebymembersof Congressandtheirstaffs,andthedisregardsofbureaucratsfortheirmembers”(Miller 1ButseeKrause(1996)whohasmadethesamedistinctionadecadeago.

4 2005:209).Duringthe1980sthe‘congressionaldominance’theoriesofbureaucratic politicsemerged(McCubbins1985;Calvert,McCubbins&Weingast1989),showing thatCongressexertedbothexpostcontrolsthroughtheappropriationprocess,legislation andlegislativechanges,andoversighthearings(Weingast&Moran1983;Bendor&Moe 1985;Kiewiet&McCubbins1991;Epstein&O’Halloran1999)andexantecontrols throughagencydesign,administrativeprocedures,auditsandcompulsoryreporting (McCubbins1985;McCubbins,Noll&Weingast1987;Macey1992;Potoski&Woods 2001;Whitford2002).Presidentialinfluencetheoriesofpoliticalcontrolwereareaction tothisschool,pointingouttothecongressionaldominancescholarsthatPresidential powersstretchoverthebureaucracythankstotheauthoritytoappointagencydirectors, reorganization,andbudgetarypowers(Beck1982;Moe1982,1985;Wood1988;Bendor &Moe1985;Howell&Lewis2002;Lewis2003). Feelingthatbothcongressionaldominanceandpresidentialinfluencestudies focusedonjustonesideoftheprocessofpolitical,multipleprincipalstudiesof bureaucraticcontrolhavebroadenedthisviewsintwoways.First,thesestudieshave shownthattherearenojustifyinggrounds,boththeoreticallyaswellasempirically,for treatingtheinfluenceofthesepoliticalactorinisolationofeachother.Secondtheyhave shownempiricallythatbureaucraticagenciesperceiveotherthanthesecoregovernmental actorsasexertinginfluenceontheirdecision-making(Wood&Waterman1991; Ringquist1995;Hammond&Knott1996;Waterman,Rouse&Wright1998;Furlong 1998).FocusingonthePresidentandCongress,HammondandKnott(1996:163) concludethatbureaucraticautonomyisa“contingentmatter”dependingonthestrategies pursuedbythePresident,theHouseandtheSenateinagivencircumstance.Insome circumstances(e.g.events)someagenciesmayenjoymoreautonomythaninother circumstances,butthesystemicfindingisthatpoliticalcontrolisamatterof“joint custody”ofpresident,congressandthecourts.Furtherevidencefortheconcomitant powersofthetwocorepoliticalinstitutionsisbroughttobearbyWoodandWaterman (1991)whofoundthatpoliticalappointments,aninstrumentsharedbybothinstitutions, hassubstantiallymoreimpactonbureaucraticresponsivenessthanchangingbudgets, legislationandadministrativereorganizations(seealsoRingquist1995). Furlongfoundthatpoliticalcontrolisnotan“issue…ofcongressionaldominance

5 orpresidentialcontrol,butratherofsharedauthorityandtheabilitytoinfluencepolicyin onedirectionoranother”.However,andherehegoesastepfurtherthanhispredecessors, healsofindsstrongsupportfortheinfluenceofinterestgroups.Heconcludes:“interest groups’abilitytoaccessandshapepolicybothdirectlyandindirectlyisevidentandmust notbeoverlooked”(Furlong1998:61).Interestgroups,asMoeassertedonce,arekey playersinthepoliticsofdelegationandadministrativedesign,asthey,perhapsmorethan electedpoliticians,careaboutthecontrolregimeofadministrativeagenciesasthese agenciesareessentialtothepolicyprocessandthedistributionofpublicservicestotheir members(Moe1995). 2Finally,thebroadestfocusistobefoundintheworkof Waterman,WrightandRouse(1998)assertthatwasbettertospeakof“venuesof influence”thanofmultiple(individual)principalsexertinginfluenceonagencies, because“influence[onanagency]isexertednotjustbymultipleprincipalsbutby differentlyperceivedsetsofprincipals[towhichtheagency]respond[s]indifferent ways”(Watermanetal1998:19). AdministrativeTheoriesofBureaucraticAutonomy Acommoncritiqueoftheexternalcontroltheorystudiesisthattheyoverlookonecentral actorinadministrativepolitics:theagencyitself.Inlinewiththesestudies,Ringquist (1995)foundexecutiveappointmentsandbudgetchangestoplayadecisiveroleon bureaucraticdecision-makingbytheEPA.ButhealsofoundthatEPAautonomywas partiallydeterminedbybureaucraticdiscretion(andhowEPAofficialsusedthatspace), agencyresourcesandthesalienceandcomplexityofthepolicyarea.Krauseassertsthat thesestudies“havelargelyignoredtheindependentroleadministrativeagenciesare capableofplayingintheadministrationofpolicy”(Krause1996:1088).Essentially, then,controltheoriesofbureaucraticautonomyhaveseeminglydownplayedtheideathat Onceanagencyiscreated,thepoliticalworldbecomesadifferentplace.Agency bureaucratsarenowpoliticalactorsintheirownright:theyhavecareerand 2Anextremeformofinterestgroupcontrolofagenciesisofcoursethatofcapture(Stigler1971).Less stringentviewsaretobefoundininterestgrouptheoriesinwhichinterestgroupsparticipateinvarying formswithpublicpolicymaking(vanWaarden1992).

6 institutionalintereststhatmaynotbeentirelycongruentwiththeirformal missions,andtheyhavepowerfulresources–expertiseanddelegatedauthority— thatmightbeemployedtowardthese“selfish”ends.Theyarenowplayerswhose interestsandresourcesalterthepoliticalgame(Moe1995:143). Administrativebehavioraltheoriesofbureaucraticautonomyputtheagency,its interests,itscapacities,itsidentityanditsmissioncentraltotheirexplanatorymodel.The mainthesisofthislineofresearchisthattheautonomyofadministrativeorganizationsis afunctionsoftheorganization’scapabilitiestoforgeitsownautonomyfromitsdirect environment.Whereaswecouldsaythatintheformerperspectivetheorganizationis consideredasapassivetakerofthediscretionthatitsenvironmentgrantstoitself,inthis perspectivetheagencybecomesanactiveseekerorforgerofautonomy.Saliencefor reputationtheorieshasrecentlyrisenduetheworkofCarpenter(2001)andKrauseand Douglas(2005).Whereastheformerhasshownthatsomeagencies(e.g.thePostal) becamehighlyautonomousduringthedevelopmentaleraoftheAmericanbureaucracy becauseoftheskilfulentrepreneurshipofagencyleaderstobuildreputation-based networks,thelatterhaveempiricallyshownthatinstitutionaldesign,oneofthekeyex anteinstrumentsofpoliticalcontrolaccordingtocontroltheorists,haveproventobea lessstrongerpredictorofbureaucraticautonomythanprofessionalreputationofagency staffwithinfinancialandeconomicsupportagencies. Still,theserecentworkscanbesaidtobefromolderdescent.Interestin bureaucraticautonomyinpoliticalscienceandpublicadministrationshasbeen prominentlyfiguringwithinstudiesofpublicadministrationforoveracoupleofdecades (Selznick1957;Downs1967;Kaufman1960;Clark&Wilson1961;Wilson1989). Bureaucraticautonomy,orbetterput:thereachingofit,wasconsideredbytheseauthors a sinequanon fornothinglessthanthe‘survival’ofanagency.Agenciessurvive, accordingtoClarkandWilson(1961),whentheysucceedinofferingtherighttypeof incentives(material,solidary,purposive)totheirmembersandbeneficiariesoftheir programtosupportthepowerbasesoftheorganization.Forthisorganizationsneedto competewithotherorganizationsforthescarceresourceswithwhichorganizationscan producethoseincentives.Duringthiscompetition“[E]achorganizationseekstoassert

7 andmaintainitsautonomyordistinctivecompetenceinorderthatitmaylay unchallengedclaimtoastockofpotentialincentives”(ClarkandWilson1961:156).The preservingofone’sautonomyhasbecomethecriticaltask,accordingtoClarkand Wilson,ofanorganization,as“[A]utonomygivesanorganizationareasonablystable claimtoresourcesandthusplacesitinamorefavorablepositionfromwhichtocompete forthoseresources”(ClarkandWilson1961:158). Autonomymustbeforged,itdoesnotcomebyitself.Theliteraturedistinguishes betweentwomainsetsofconditionsforachievingautonomy(cf.Wilson1989:188-192). Onesetofconditionisthatagenciesdeliverservicesthatare“worthwhiletosomegroup withinfluenceoversufficientresourcestokeepitalive”(Downs1967:7).Itmustthus developaclientelethatbenefitsandperhapsisdependentupontheservicesdeliveredby theagency.Thisdependence,however,shouldalso“impressthosepoliticianswho controlthebudgetthatitsfunctiongeneratepoliticalsupport”formwhichpoliticianscan benefitintermsoftheirre-electionorimprovingoftheirpartypoliticalbase(ibid.). Carpentercallsthisthepoliticallegitimacybasisofanagency: Bureaucraticautonomyrequires politicallegitimacy ,orstrongorganizational reputationsembeddedinanindependentpowerbase.Autonomyfirstrequires demonstratedcapacity,the belief ,bypoliticalauthoritiesandcitizensthat agenciescanprovidebenefits,plans,andsolutionstonationalproblemsfound nowhereelseintheregime.Thesebeliefsmustalsobegrounded inmultiple networks ,throughwhichagencyentrepreneurscanbuild programcoalition aroundthepoliciestheyfavor(Carpenter2001:14,emphasisinoriginal). Asecondsetofconditionsforautonomyisthatagencieshavethecapacitiesto deliverthebenefitstotheirsupporters:itmusthavethestaffwiththerightskillsandthe dedicationtofulfilltheorganization’sultimategoalofsecuringastableandstrong autonomouspositionwithinthepublicspace.Kaufman’sstudyoftheforestserviceforms anexamplehere.ThehighdegreeofunityKaufmanobservedwithintheForestServiceis partially–nexttohierarchicalspecialization,communicationprocedures,amongothers– theresultantoftherecruitmentandsocializationoftherangers:theyareinfusedwiththe

8 valuesandnormsofheadquartersbuttheybehaveasifthesevaluesweretheirown.3In linewiththis,carefullydesignedrecruitmentinstrumentsiswhatSelznick(1957)sawas aprerequisiteforautonomy.Theleadershipofanorganization,ifitwantedtopassthe “initialsurvivalthreshold”(Downs1967:9)andtomaneuverthroughcriticalperiods ahead,ithadtocarefullyselect,first,the“socialbase”,i.e.thosesegmentsof“the [organization’s]environmenttowhichoperationswillbeoriented”(Selznick1957:104). Asecondcriticaltaskisthecreationofahomogenousstaff.Thisway,heagency’s leadershipcanbuildaninstitutionalcorewho“whenmaturedinthisrole,performthe essentialtaskofindoctrinatingnewcomersalongdesiredlines”andsecurethatfuture policiesoftheagencywillconform,bothinnormandspirit,tothedoctrinesandmission ofthefoundersoftheorganization(Selznick1957:105). Carpenterbringsthesetwoconditionsforbureaucraticautonomynicelytogether. Hedefinesautonomyasthecapacityofbureaucraticactorsto“changetheagendasand preferencesofpoliticiansandtheorganizedpublic”(Carpenter2001:15).Inthishe ascribesthebureauchiefsofadministrativeagencies–themezzolevel–acrucialrole. Agenciesthathavemanagedtoacquirebureaucraticautonomy,Carpenterargues,have acquiredthatthroughthecapacityofbureauanddivisionchiefs(1)topolitically differentiatetheiragencies’preferencesandmissionsformthoseofpoliticiansand organizedintereststhroughcarefulrecruitingandsocializingtheirpersonnel;(2)to developorganizationalcapacitiesforprogramming,planning,deliveringpublicservices, andaboveall,toinnovatetheirprograms;(3)andtoforgemultiplenetworksinthe environmentoftheiragenciesthatsupporttheiragencies’missionsandgoalsandinthe endtogainastrong legitimacy basefortheiragencies. 3.ResearchDesign CaseSelection:CompetitionPolicyRegulationinTheNetherlands 3TheleadershipoftheForestService“actuallyinfuseintotheforestofficersthedesiredpatternsofaction inthemanagementoftheir,sothattheRangershandlemostsituationspreciselyastheirsuperiors stoodlookingovertheirshoulders,supervisingeverydetail…FromtheRangers’pointofview,theyarenot obeyingordersorrespondingtocueswhentheyactionontheirdistricts…Itisnotcompulsionor inducementorpersuasionthatmovesthem;itistheirownwill.”(Kaufman1960:222).

9 Themainaimofthisresearchistotestthecompetingtheoriesofbureaucraticautonomy withinadifferentinstitutionalsettingthanwherethesetheorieswereinitiallydeveloped andtested.Tothisend,Iwillexaminetheextenttowhichthebureaucraticautonomyof theCompetitionAuthorityintheNetherlands( NederlandseMededingingsAutoriteit – NMA)isaffectedbythecontrolinstrumentsofitsprincipalsandtheinternalbureaucratic capacitiesoftheagencyitself.Theprimarypurposeofchoosingthiscaseisthatof replication:towhatextentdothefactorsthatarereportedinUSstudiesashavingan effect(ornot)onthebureaucraticautonomyofregulatoryagenciesperformwithina Europeanparliamentarysetting?Thevariationininstitutionalsettingallowsustoassess whethercontrolandreputationvariableswillperformthesamewithinafundamentally differentinstitutionalsetting(Weaver&Rockman1993;butseeHammond&Butler 2003).Ifthisisthecase,wewouldcomeclosertohavingmoregeneralizable mechanismsthataffecttheautonomyofbureaucraticagencieswithindemocratic systems. TheNetherlands’politicalsystemisaparliamentarydemocracyofthe consociationaltype(asopposedtotheWestminstertype,seeMoe&Caldwell1994), amongitsbasicpoliticalinstitutionalfeaturestheconsequentproductionofmultiparty governmentsfoundeduponastable,oftenoversized,parliamentarymajoritycoalitions thatusuallyformasaresultofthe’sproportionalelectoralsystem(Lijphart1999; Andeweg&Irwin2005).ThissystemalsodifferssharplyfromtheUSsystemduetothe ministerialportfoliosystem(Laver&Shepsle1990,1996).IncontrasttotheUS presidentialsystemwhereaclearseparationofpowersbetweentheexecutiveand legislativebranchesjuxtaposesPresidentandCongressonarelativelyequalfooting, withinparliamentarysystemsthelegislatureandexecutivearemoreorless‘linedup’ withthelatterbeingfullyresponsibleanddependentuponthelegislaturethroughavote ofconfidence. 4Thereisasingularchainofdelegationfromvoterstoparliament,from parliamenttocabinet,fromcabinetleadershiptoindividualminister,andfromthe individualministertothebureaucracy(Andeweg2000;Strøm2000).Attheexecutive 4ThisisthecaseintheNetherlands.Inotherparliamentarydemocracies,e.g.whereminoritygovernments arethenormsuchasinScandinaviancountries,therelativebalancebetweentheexecutiveandthe legislatureismoreinfavorofthelatterthaninthecaseoftheNetherlands.Parliamentarysystemsalso differalongarangeofothermoreformaldimensions,suchasthenecessityofaformalvoteofconfident. ForthesevariationspleaserefertoLijphart(1999),MüllerandStrøm(2003)

10 endofthischain,individualministersarethe“relevantprivilegedactors”.Theirportfolio assignstothemanautonomousroleformakingproposalsandevenpolicyforwithintheir portfolio(cf.Huber&Shipan2002).Thissystemmakesaministeranagentof parliamentbutatthesametimeaprincipalforthebureaucracythatisassignedtoher portfolio.Theautonomyofbureaucraticagenciesispartiallydependentontheformal powersoftheminister(Christensen2001).Withinparliamentarysystems,ministersare ultimatelyresponsibleforthedecisions,actionsandbehaviourofbureaucratsresiding undertheirauthority. 5 AnotherreasonforwhyIhavechosencompetitionpoliciesisthatofthespecific circumstancethatthe1998reformofDutchcompetitionpoliciesandthesubsequent establishmentoftheDutchCompetitionAuthorityasthemainenforcerofthisnewpolicy meantafundamentalbreakwiththepastpracticesandcultureoftheDutcheconomy.For decadescartelswerenotonlyadmittedbutwithinthecorporatistandconsociational politicaleconomiccultureoftheNetherlands,cartelswerealsoconsideredthe appropriateformofeconomicgovernanceintheNetherlands.Fullmarketcompetition wasconsideredasathreattothestabilityandsurvivalofvariouseconomicsectors. Hence,a1993reportoftheOECDontheNetherlandsreportedthatthe(secret) governmentregistersofcartelcontained245agreementstodividemarkets,about270to fixprices,50agreementsonexclusivedealings,andmorethan200agreementstocontrol competitionindistribution(inOECD1998:4).Thepre-1998periodintheNetherlandshas thereforebeenutterlydescribedasa‘cartelparadise’(deJong1990).Beinganewkidon theblock,theNMAwasdelegatedtheseeminglyinsurmountabletaskofliberalizingthe Dutcheconomicsectorsandbringitintolinewiththeprinciplesofliberalmarkets.In otherwords,acasestudyoftheautonomyoftheDutchcompetitionauthoritythatwas createdonlyin1998inordertoexecuteanti-trustlegislationinacartelparadise,ishighly

5IntheNetherlands,oneoftheinstrumentstodeterminethedegreeofdiscretionislaiddowninthe Guidelinesforlegislation( Aanwijzingenvoorregelgeving ).Indirectlyitprescribestoindividualministers whatkindofpowerstheycandelegatetoindependentbodiesaswellastheareasonwhichthisdiscretion maybeused.Article124f(1)statesthatonlyonorganizationalandtechnicalmattersandinspecialcases providedthattheministersbeholdshispowerto(dis)approvewiththeagency’sdecision.Furtherit prohibitsthroughArticle124f(2)thedelegationofadvisorytaskswithregardtotheformulationofgeneral bindingrulesinmatterspertainingtocentralgovernmentalpolicies.Aministermaydelimitthepolicy discretionofagenciesbyissuinggeneral(article124l,5b);issueguidelinestotheagencyofhowtoexecute theirtasks(article125l,6).

11 suitedforthepurposeofexaminingtheeffectsofexternalpoliticalandinterestgroup pressuresontheNMAaswellastoassertthevalueofadministrativebehavioural explanations. Afinalreasonforchoosingcompetitionpoliciesisthatitbringsthecrucial dimensionofEuropeanpublicpolicymakingintotheequation.Sincetheestablishmentof theEuropeanCommunitycompetitionpoliciesformacoreareaofEuropeanintegration. Article3(g)oftheECTreatystatesthatwithintheEUthereshallbea“systemensuring thatcompetitionintheinternalmarketisnotdistorted’.Articles81and82(anti-trust), Articles87-89(stateaids)andMergerRegulation139/2004(replaced4064/89)delegate totheCommissionextensivepowerstoimplementandenforcecompetitionpolicies withinthememberstates.CouncilRegulation1/2003‘modernized’ECcompetition policiesbyextendingthepowersoftheEuropeanCommission(i.e.DGIV,the competitionDirectorateGeneral)tothenationalcompetitionauthoritiesandnational courts.Inotherwords,theDutchcompetitionauthoritymightwellbeJanus-faced:ithas beendelegatedauthorityfromboththenationalaswellastheEuropeanprincipals.What thissituationimpliesfortheautonomyoftheNMAisanempiricalquestion.Itmakesthis casemorethanamerereplicationofUStheoriesofbureaucraticautonomyasitalso delvesintodevelopinganewtheoryofbureaucraticautonomywithintheEuropean multi-levelgovernancesetting(Hooghe&Marks2001).However,thisdoesnotenable anygeneralizationtobureaucraticautonomyinmulti-levelsettingsascompetition policiesarequiteuniquevis-à-visotherareasofEuropeanpublicpolicyintermsofthe extensivepowersEC’sDGIVenjoys. Methodology Casestudydesign,RodgersandJensen(2001)assert,shouldideallybestandardizedsuch thatthecausalmechanismsandtheoutcomesofeachindividualcasestudycanbe cumulatedandcomparedwithlargersetofcasestudyfindings.Thedesignofresearchin thispaperreflectsthislineofthought.BelowIwilldescribetheprimaryresearchentity aswellasthemaininterventionsandcriticaleventsthatarebelievedtohavehada

12 significantimpactontheprimaryentity. 6Morespecifically,thisresearchisaqualitative longitudinalcasestudythatanalyzestheimpactofaseriesofinterventionsandcritical eventsontheNMA’agendaanditsagenda-settingprocesswithregardstoanti-trust policies.Inasense,thiscasestudycombinesthebasicideaofalongitudinalcasestudy withthatofpre-posttypeofcasesstudies(Rodgers&Jensen2001:238).Thetimeframe startswiththecreationoftheNMAon1January1998andendson31June2007.

DEPENDENTVARIABLE .ThebureaucraticautonomyoftheNMAismeasuredin termsofautonomytheagencyhastosetitsownregulatoryagenda. 7Thedeterminationof itsownagenda,iswidelybeenconsideredasonethekeyindicatorsofbureaucratic autonomy.TheagendaoftheNMA,then,isthelistofprioritizeditemstowhich,when authorizedbytheNMAleadership,theorganizationwillallocateitsscarceresourcesin ordertotheitemsthathavebeenputonthelist.TheNMAisamulti-industry regulator(cf.Macey1992)thathastoapplyanti-trustlegislationacrossamultitudeof differentsectors. 8TheNMAhasalmostbydefinitionlimitedorganizationalresourcesbut inordertofulfilitscoretaskitmustchoosefromamongseveralindustriestomonitor; thisprocesscomestothevergeoforganizationaldecision-makingasisdescribedwithin classicstudiesofpublicadministrationandorganizationalsociology(Cohen,& Olsen1972). Thedependentvariableofisagenda-setting.Ihavefocusedontwoindicators.The firstisthenamesofindustries(e.g.housing,finance,healthcare)thathaveappearedon theagendaoftheNMA.Themainaimwithregardtothisvariableistoexaminetowhat extentthislistofprioritizeditemsistheresultantofpoliticalcontroloradministrative behavioural.Inotherwords,hastheNMAputtheseitemsitselfontheagendaorhasit putspecificitemsonitsagendabecauseofexternalpressures?Thesecondresearchentity istheagenda-settingprocedureitself.Thesettingoftheagendaisaroutineprocess withintheNMA.However,routinesmaychangeandtheproceduresmayreflectbiasesas

6RodgersandJensenusetheterms“outcomesforprimaryentity”and“interventions”fordependentand independentvariables,respectively.Notethatthereisnofundamentaldifferencebetweenthetwo approaches. 7Ichoseheretolookatagenda-settingautonomy.ThishascloseresemblancetowhatVerhoestandothers refertoaspolicyautonomy(Verhoestetal2004) 8SomesectorsfalloutsideofthejurisdictionoftheNMAandareregulatedbysingle-industryregulators, suchasthetelecommunicationsregulatorOPTA,HealthCareAuthority,electricityregulatorDte(whichis infactasemi-detacheddepartmentwithintheNMA,buthasadifferentjurisdictionthantheNMA).

13 towhichkindofitemshavealargerchanceofbeinglistedthanothers.Inotherwords, examiningthisvariablemayrevealtouswhetherthereexistfactorsthat‘favour’certain industriestobeprioritizedoverothers,andwhetherpoliticalcontrolfactorshavelarger orsmallerweightthanadministrativebehaviouralinthisprocess.Bothvariableschange overtimeandweneedtoaccountfortemporality(seebelow).

INDEPENDENTVARIABLES .Inaccordancewiththetwosetsofcompetingtheories, Idistinguishbetweentwosetsofmaininterventions.Thefirstsetofinterventionsare fromexternalprincipals(politicalcontrolexplanations).Thebriefliteraturereviewabove hasidentifiedanumberofdeterminantsthathavebeenclaimedtohaveaneffecton bureaucraticautonomy.Theseare,likeseveralcriticaleventsandconstantfactors,given inTable1. Fromthelistofpotentialpolitical controlvariables Iwillassesstheestimated effectsoftheappointmentoftheagencyheadandagencydesignontheagendaand agenda-settingproceduresoftheNMA.Thelistofpotentialcausalfactorsthatwe discussedaboveislarger,butgiventhestructureofdelegationinparliamentarysystems, i.e.executivedominanceandministerialautonomy,thesearethemainvariablesofwhich onecanempiricallyexpecttoexertaneffectonbureaucraticautonomy.Duringthe periodunderstudy,theMinisterofEconomicAffairshadtheformalauthoritytoappoint thedirectoroftheagency.Theministerusedhisauthoritytwice,in1998andthenin 2005.Asregardsagencydesign,theNMAstartedasadepartmentalunitoftheMinistry ofEconomicAffairs;in2005itwasreorganizedasanindependentadministrativebody. WemayexpectthattheautonomyoftheNMAtosetitsownagendahasrisenafteritis becomeformallyanindependentadministrativebody. Thesecondsetofindependentvariablesarethosethatestimatetheeffectsof administrativebehaviouralfactorsonthebureaucraticautonomyoftheNMA.Twosets ofvariablescanbedistinguished:politicallegitimacyandorganizationalcapacity.There aretwoproxiesforpoliticallegitimacy:theagency’sreputation,heremeasuredasthird actors’perceptionsoftheidentityandmissionoftheNMA,andthedegreetowhichthe agencyhasanindependentpowerbase,heremeasuredasthetypeandstrengthofthe networkswithinwhichtheNMAisanchored.Thesecondbureaucraticdeterminantis organizationalcapacity.Theliteratureoftenhasservicedeliveryagencies(e.g.schools,

14 police,socialwelfare)aspointofreferenceandmeasuresorganizationalcapacityin termsoftheextentinwhichtheagencyiscapableofdeliveringtheservicestothe constituencythatisneedoftheseservices.TheNMA,however,isaregulatoryagency thatadjudicates,licenses,orinvestigatesinthefieldofanti-trust,mergersandstateaids cases,andassuchitdoesnotdeliverservicesinthesenseasisjustdescribed.Rather,the organizationalcapacityoftheNMAthatisdeemedcrucialforitsautonomycanbe measuredbylookingatthetypeofindustriesorindividualcasesithasprioritized.Ahigh leveloforganizationalcapacitywillexpectedlyleadtheNMAtohandlebiggerandmore importantcaseswithinmoreimportantindustries.Wecanthenhypothesizethatthemore theNMAdecidesincaseswithhigherlevelsofcomplexity,incasesthatare(politically) sensitive,orcasesinwhichthe“bigfish”,i.e.largerfirmssuchasmulti-nationals,the moreautonomytheNMAenjoystosetitsownagendaanddetermineitsagenda-setting process.

CONTROLS .Nexttothesedeterminantswhichwederivedformthetwomain bodiesoftheoriesofbureaucraticautonomywemustincludeascontrolsoralternative explanationsatleastthreecriticalevents,i.e.eventsthathavehadcausaleffectsonthe agencyautonomyoftheNMAbutthatcouldnotbeforeseenbythetheory,intothe equation.Thefirstisthedisclosureofthelarge-scalebutuntilthenuncoveredcartels withintheconstructionindustry.Thetelevisionnewsprogram,Z embla broughtthenews andinterviewedawhistleblowerfromoneofthefraudulentconstructionfirms.This whistleblowerrevealedwithinthisprogramthathedidhavecontactwiththeNMAin early1999andthattheNMAhadthusbeenawareofthesepracticessince1999,buthad beenunwillingtotakeaction.ThisnewsarousedagulfofcritiqueontheNMAfromthe sideofparliamentandmedia,questioningtheagenda-settingcapabilitiesoftheNMA. ThesecondeventwastheenactingofECRegulation1/2003on1May2004.This regulationsignificantlynotonlyalteredtherole,powersandpositionoftheNMAvis-à- vistheCommission,butalsovis-à-vistheMinisterofEconomicAffairs.Thiseventlead twoDutchpubliclawyerstoassertthatthe“discretionaryroomfortheministerbecomes smallerthemoreoftentheCommissionandthenationalauthoritiesmanagetoreach policyagreementswithinthesettingoftheEuropeancompetitionnetwork”(Verheij& Verhey2005:219).Iwilltesttheirhypothesis.

15 Thirdly,on2November2006thelongawaitedFrameworkActofIndependent AdministrativeBodieswasenacted.Thelawhasbeendebatedsincethecritiquethat emergedduringthe1990sonthefragileconstitutionalembeddingofindependent agenciesintheNetherlands.TheFrameworkActwantedtoendthefloweringofdifferent regimesofindependentagenciesandstandardizeboththeproceduresaswellasthefinal designofindependentagencies.RegulatoryauthoritiessuchastheNMAarebynature andtasksdifferentfromthemajorityoftheolderandexistingindependentbodies– regulationversusservicedelivery–butthedebateconcerningtheindependencyof regulatoryauthoritieshasneverthelessbeensubjectedtothegeneraldebatesurrounding independentadministrativebodies(Verheij&Verhey2005). Finally,weneedtoaccountforthe“growthcycle”oftheNMA.Giventhe specificmomentswheremostoftheinterventionsandthecriticaleventsintervenedinthe courseofdevelopmentoftheNMA’sagendaautonomy,thisdesignsystematically accountsforthetemporaleffectsoftheobservations.However,weneedalsotoaccount forthegeneralsequenceofagencydevelopmentagainstwhichbackgroundtheevents andtheinterventionstakeplace.FromDownsandotherswelearnthatagenciesstruggle forautonomywhentheyarejustcreatedandthatitsneeds,missions,andtheorientations ofitsleadershipandstaffwilldifferonceanagencyhaspassedthe“initialsurvival threshold”(Downs1967:9).Inotherwords,wehavetoasserttowhatextentour observationsonagendaautonomythatarereactionstointerventionsoreventsispartly causedbythespecificphaseorstageinwhichtheagencyfindsitself.

16 Table1Interventions,criticaleventsandconstantfactorsbetween1998andJune2007 Description Measurement Dependentvariables Agenda Thelistofsectororcasesthathavebeenprioritizedbytheagency Namesofsectors andtowhichitwillallocateitsresourcesto Agendasettingprocedures Theprocedureswithwhichtheagencyselectssectorsorcasesthat Agendasettingandprioritizationprocedures willbeputontheagenda Independentvariables Externalcontrolvariables : appointmentofagencyhead Theappointmentofanagencyheadisfoundtobeoneofthemost Between1998and1July2005,NMAwasheadedbyaDirector- importantinstrumentstocontroltheautonomyoftheagency General;after1July2005byBoardofDirectors agencydesign Theoretically,itisclaimedthatregulatoryagenciesneedtobe Until1July2005,NMAwasaunitoftheMinistryofEconomic independentfrompoliticalinterventionsinordertoenhanceand Affairs;thereafteritbecameanIndependentAdministrativeBody maintainthecrediblecommitmentpoliticians. withoutlegalpersonality ministerialresponsibility Ministersareultimatelyresponsibleforthebehaviourandactions Theuseofoverridepowersortheblockingofanagencydecision ofbureaucraticagencieswithintheirportfolio.Thereforethey enjoyextensivepowerstooverrideorblockagencydecisions pressuregroups Pressuregroupsrepresentingestablishedfirms,newentrants, Thenumberandintensityoflobbies,aswellastheirentrance consumersandotherinterestswithinregulatedsectorshave pointintotheprocessofagenda-setting interestsinregulationsthatbenefittheirposition Administrativevariables: politicallegitimacy 1.Embeddedreputationin2.independentpowerbase 1.TheperceptionoftheNMAbyexternalactorsand2.networks withinwhichtheNMAisembedded organizationalcapacity TherelativesuccesswithwhichNMAimplementsanti-trust TheextentinwhichNMAhasbeensuccessfulinmakingcritical provisions decisionsdifficultandsensitivecases Alternativeexplanationsandcritical events constructioncartelaffair TheprogramZemblabroughtatelevisedreportofseriousfraud TheNMAwascriticizedforhavingfailedtodetectthisearlier.I (9November2001) andcartelpracticeswithintheDutchconstructionsector measurethereactionsoftheNMAtothisnewsandthe conclusionsofsubsequentparliamentaryhearing ECRegulation1/2003 ThisregulationdecentralizedECcompetitionpoliciestothe ThenumberoftimestheNMApartookinECregulatory (1May2004) memberstates’competitionregulators processesbeforeandaftertheregulation time Thetemporaldimension Theorientationsofanagencyleadershipwilldifferoncean agencyhaspassedthe“initialsurvivaltreshold”

17 4.Results InthissectionIwillassesstheestimatedeffectsofeachgroupindependentvariables.For eachvariableIwilldiscusstheextenttowhichthisvariablehashadadirect,indirect(or intervening)ornoeffectatallontheagendaandagenda-settingproceduresoftheNMA. Ofthesevariables,then,Iwillonlydescribeatsomelengththosevariableswithadirect orindirecteffect,andonlymentionthevariablesthathadnoeffect.Firsttheestimated effectsofthemodel’sfourcontrolvariableswillbeassessed,followedbythe administrativeandcontrolvariables,respectively. External/PoliticalControlExplanations Withinthisclusterofvariables,onlytheappointmentofagencyheadshadindirecteffect ontheagendaoftheNMA.Ifoundnodirectorindirect,positivenornegativeeffectfor agencydesign,ministerialresponsibility,andpressuregrouppolitics.Iwilldiscussfirst howappointmentshadanindirecteffect.ThenIwilldiscusstheremainingnon-effective controlvariables.Iwillfinallydevoteattentiontothe(non-observed)effectsofagency design,sincethisisaprominentvariableindelegationandpoliticalcontrolliterature. Theappointmentoftheheadofanagencycanoffertheappointingauthoritya substantialleverageovertheagendaoftheagency(Moe1985).Intheperiodunderstudy, theNMAhashadtwodifferentpersonsasdirector-generalandpresidentoftheboard, Mr.A.W.KistandMr.P.Kalbfleisch.Kalbfleischwasappointedasdirector-generalin 2003butwasreappointedaspresidentaftertheNMAgotbeholdofitsindependent status.Ingeneral:Whileideologyseeminglyhasnotplayedaroleintheirappointment, andthereisnodiscernibleinfluencebetweenideologicalorientationofbothpersonsand theNMAagenda,thetheredoesseemtobearelationshipbetweenthedominantpolitical ideologicalopinionsoftheirtimewiththechoiceofindustrialsectorsonwhichtheNMA focusedduringtheirreign. KistwasaformerpartneratalargelawfirminTheHague,specialisedin companylaw.HewasappointedinNovember1997bytheMr.H.Weijers,thefervent marketliberalMinisterofEconomicAffairsofthejuniorcoalitionpartner,thesocial liberalpartyofDemocrats66(D’66).OntheoccasionofhisappointmentKist,whohas beenallegedlyaffiliatedwiththeLabourParty(NRC29August2005),hemadeclear thattheNMAunderhisrulewoulddefendtherightsofconsumers:“Wearethe guardiansoftheinterestsofconsumersandkeepersofeconomicdemocracy(sic)”,Kist toldtheaudienceattheceremonyoftheinstallationoftheNMAon25November1997 (NRC,26November1997).Heexplicitlystatedthatconsumerswere“theconstituencyof theNMA”(FD,26November1997).However,justbeingsetup,theagendaoftheNMA hadbeenquiteoverloadedwithrequestsfromcombinationsoffirmstoobtain dispensationfortheirconcentration. 9Althoughsomecasesinnameof‘consumer interests’weretakenon,duringtheearlyyears,theNMAwaspreoccupiedwithaburden ofthesetransitionalarrangementthatemanatedfromtheintroductionoftheCompetition Actin1998.Theremainingcapacitywasallocatedtotheimplementationofanti-trust regulations,butasonecivilservantoftheMinistryofEconomicAffairstoldtothestaff oftheParliamentaryEnquiryCommitteein2002,theNMachosethecasesitwould investigateona“’make-a-case’principle.Itwasdeemedmoreimportant”,thecivil servanttoldthecommittee,“toshow[totheoutsideworld,KY]thattheregenuinely existedacompetitionpolicythantoactenforcethelaw”(KamerstukkenII,2002-2003, 28244,nr.12:142). HissuccessorKalbfleischwasformerlythevice-presidentoftheHaguecourtof justice.HewasappointedbythesocialliberalministerMr.P.Brinkhorst;hewasfrom thesamepartyasWijers,butknownassomewhatmoremoderatethanWijers.WhileKist hasbeenaffiliatedwiththeLabourParty,Kalblfleischhadnopoliticalaffiliation(ANP, 15August2003).Whenhewasappointeddirector-generalinOctober2003,theNMA wasstillpreoccupiedwiththeinvestigationsintheconstructionindustry.Aswewillsee below,thelargescaledetectionofcartelswithintheconstructionsectorinNovember 2001andforwhoseneglecttheNMAhasbeenseverelycriticizedbythepublicand

9Thenumberoftheserequestsexceededtheestimatedexpectednumberofsuchrequests:1040requests weresubmitted,while350wereexpected.Inaddition,theNMaexpectedtoprocessapproximately150 complaintswithregardtoabusesofdominantpositionsfromindividualsorfirmsfromdifferentsectorsor tipsfrominformersconcerningcartels:itwere266.Finally,50tipswereexpectedbut154weresubmitted. Althoughtheagencyhadcorrectlyprepareditselfwithregardtothetypeofworkthatwould,they miscalculatedtheamountofit(NMa1999 jaarverslag1998 ).

19 Parliamentandthelatter’sspecialinvestigativecommitteethatwassetupaftertheaffair cameintodaylight.Nexttotheconstructionsector,theNMAcontinuedinvestigating areasthathadbeenputontheagendabyKist,thefreeprofessions(e.g.lawyers)andthe financialsector(e.g.banksandinsurancecompanies).Atthesametime,thenumberof sectorsbeingliberalizedwasincreasingduringtheperiodKalbfleischassumedoffice. Thesectorsunderliberalizationwerepublictransport,energysector,andthehealthcare sector,whichenlargedthenumberofareastowhichtheNMAhadtoattend. TheagendaoftheNMAhasbeentoalargeextentbeentheconsequenceofthe choicesofthedirectors,transitionalarrangements,andtheincreaseoftheareasofthe Dutcheconomythatwereliberalized.Ministerialinfluenceontheagendathrough appointmenthasbeenindirectatbest,asnodirectpoliticalorideologicallinkagecanbe madebetweentheministers’politicalbackgroundandthetwodirectors.Whywereother variablesnoteffective?Agencydesignhashadnoeffectontheagendabecausethere werenosignificantchangesobservedintheagendaandagenda-settingproceduresthat couldbeattributedtothedesignandchangeofdesigntheNMAduringtheperiodunder study.Thesameistruefortheeffectsofministerialresponsibility. 10 Before independence,theministerhadtheauthoritytoissueinstructionsinrelationtoindividual casestothedirector-generaloftheNMA;afterindependence,thisauthoritywascurbed andreplacedbytheauthoritytoissueonlygeneralinstructionsastohowtheboardofthe NMAshouldtakeotherintereststhaneconomiconesintoaccount(vandeGronden&de Vries2006:61).Butthroughouttheperiod,nousehasbeenmadebytheministerofthese authorities. Iwillillustratetheabsenceofeffectsfromagencydesigninsomewhatmore depth,bylookingattheinteractionsbetweentheministerialdepartmentatthemezzo- level.Thecentraldepartmenthasdensecontactswiththe‘parastatal’agenciesthatfall undertheportfolioofaminister.Soforministerialinfluenceweneedacloserlookatthe interactionsbetweenthedepartmentandtheNMA. 11

10 InthecaseoftheNMa,theministermayprescribesuchguidelinestotheboard(Article5dMw),but theseguidelinesmustfocusonhowtheNMashoulddealwithnon-economicinterestswhendecidingon theexemptionslaiddowninArticle6(3)(thesameasArticle81(3)EC). 11 Theissueofindependencyhasbeenaconstantissueofdebate.Duetothepoliticalimplicationsof havinganindependentagencyregulatingcompetitionpoliciesintheNetherlands,whichatthattimewas

20 ContactsbetweentheNMaandtheMinistryofEconomicAffairshavebeen institutionalized.ThebasicruleisthatcontactsbetweentheministryandNMawillnot involvediscussionsorexchangesofinformationontheindividualcaselevel.Thereare severalcontactsatdifferentlevels,bothformalandinformal:thedirectorofcompetition attheministryandtheboardordirectorofthestaffunit‘strategyandcommunication’; bothsideshaveestablishedaccountmanagersforrelationshipswithNMaandEconomic Affairs,respectively,andatthelevelofindividualcivilservantspolicyunitsofthe ministryandthecompetitiondivisionoftheNMa.Atoneofthehighlevelmeetingsthe issueseparatingelectricityinfrastructurefromtheelectricityproduction.Atsuch meetings,aswasdescribedtousbyahigherNMaofficial,theydiscuss Allkindsofissues.Theykeepusinformedaboutdevelopmentsonhighstrategic issuesthataregoingonattheirside,andwekeeptheminformedaboutcurrent developmentsatouragency.Theythusforexampleinformusaboutthelatest developmentsconcerningthe‘separationfile’[separationofinfrastructurefrom productionofelectricity,KY].Wethensay,wellbeinformedaboutthefact, althoughitisnoneofyourbusiness,thatareportaboutsuchandsuchisaboutto appear[thatmightofrelevancetoyou]…It’sjustplainlyaboutkeepingeach otherinformed(respondentAfromNMa). Duringthesecontacts,thereisastrictdistinctionbetweenpolicyissuesandmonitoring issues :“We discuss policyissueswiththem,butweonly inform thedepartmentabout

stillacartelparadise,theDutchlegislaturedidnotdaretogoasfarastodesigntheNMaasanindependent agency(VerheijandVerhey2005:148).Atthesametimeitwascleartoallinvolvedpartiesthatthe regulationofcompetitioncouldonlybesuccessfulifpoliticianswouldabstainfrominterveninginthis area.ThefirstCompetitionActthereforecontainedaclausethatstatedthatanyinterventionbytheminister inaspecificcaseshouldbebasedonawritingthatwouldbeattachedtotheindividualcase’sfileinnorder toensureasmuchaspossiblethetransparencyoftheprocess(Article4dCompetitionAct(old), Staatsblad 1997,242).Thisclauseshouldalsoensurethatanygeneralpolicyinstructionsfromtheministertothe director-generaloftheNMashouldbepublishedasanofficialpublicationintheformofpolicyrules (beleidsregels )asspecifiedbytheGeneralActofAdministrativeLaw(KamerstukkenII,1995-1996, 24707,nr.3:60).Inotherwords,althoughtheNMawasdesignedasadepartmentalbodyunderthedirect hierarchicalsupervisionoftheminister,theindependenceoftheNMainthehandlingofcaseswasasmuch aspossibleshieldedfromministerialinterventions.Inlaterevaluations,itturnedoutthattheminister, despitetheseverelobbyingfrompressuregroups,hasneverintervenedintheagendaoftheNMawhenthis agencywaspartoftheministry.

21 issuesconcerningthemaintenanceofthelaw”(respondentformNMa).Informingmeans, forexample,thattheNMawillannouncetheraisingoftariffssothatthedepartmentcan preparetheministerforhowtoreactonthisnewswhensheisaskedbythepressaboutit. TheMinistryofEconomicAffairscaninfluencetheagendaoftheNMa,albeitin aindirectmanner.Whereastheministry,aswehaveseen,abstainsitselffromintervening intheagency’sdecisions,itdoeshavecompletecontroloverthefundingoftheagency. Thishastodowiththefactthatonlytheboardofdirectors,whichreplacedthedirector- general,hasbeengrantedindependency.Theagency’sbodyhasnolegalpersonalityand ispartofthedepartmentalhierarchy.Althoughnoneofthecivilservantscantakeorders ordirectionsformthedepartment,thedepartmentcaninfluencetoalargeextentthe budgetmechanism.TheNMaisdependentonthedepartmentinbudgetarysenseintwo ways:theyearlyworkplanandthereorganizations.Intheworkplan,theNMasetsout howitisgoingtoallocateitscapacityoverthevarioustasksandprioritizes,followingits yearlyagenda,thesectorsthatwillenjoyspecificscrutinyfromtheNMa.Thisworkplan isthensendtotheministryforafinancialapproval,itbecomesanissueofthehalfyearly departmentalbusinesscycle:theNMaisaccountabletothedepartmentaboutthetargets setandperformance.Thesamegoesforreorganizations(NMayearreport2004:36).The organizationoftheNMacanbeorganizedwithintheframethatisagreedwiththe ministry.Allchangestopersonnelandtaskallocationwithintheformalorganization chartistheresponsibilityoftheNMa;however,iftheNMadecidestosetupanewunit orsubunitwithinoneoftheexistingunits,thenthatbecomesareorganization,whichhas tobeapprovedofbythedepartment.However,thereisagreyareaoforganizational changesthatcanorcannotbedefinedasreorganization.Aseeminglysimpletransferofa numberofstaffmembersfromoneunittoanother“couldbeinterpretedbythe departmentasareorganization.Thenwehavetocoordinatethatchange”(respondentC fromNMa). Finally,pressuregroupsdidnotexertdecisiveinfluenceontheNMAagenda,but theywereactiveinpleadingtheircasetotheministeratthetimewhentheagencywas partoftheministry.Asamatteroffact,theimmensepressuresformthesectorswhich thethenMinisterofEconomicAffairs,Mrs.A.Jorritsma(1998-2002),experienced,was oneofhermainmotivestograntanindependentstatustotheNMA.Sheexplained:

22 “Nowthey[pressuregroups]cometomewiththeircomplaintseveryday.Butafter havingwatchedtheNMaoperateduringthepastsixmonths,Ihavethestrongfeeling thatIdonotneedtheformalpowerstointerveneinitsbusiness.Ireallywanttogetrid offofit”(JorritsmainFD,12September1998). 12 Despitethesepressures,industry lobbiesdidnotexertanydiscernibleeffectontheagendaoftheNMA.Theexceptionis consumerinterests,butthiswastheproclaimedconstituencyoftheNMA,asKisthas saidseveraltimesintheinitialphaseoftheNMA,tomakeclearthatthiswastheprime legitimacygrounduponwhichtheNMAwoulddoitsdifficultjob.Atthesametime,that thishasbeenalsothemottoofanagencyleadercreatingareputationforbeingtoughon cartelsisanissuewewillexploreinsomewhatmoredepthbelow. Administrativeexplanations Bureaucraticagenciesearnpoliticallegitimacyiftheycanbuildfavourablereputations, thatisbeliefsandperceptionsabouttheagency,amongactorsintheirnearenvironment andthemoredistantpublic,andwhenthesereputationsbecomevestedwithinnetworks ofgroupsthathaveaninterestintheagency’sprograms.Thecapacitytodeliverits productsisintricatelyrelatedtothereputationofanagency.Autonomyispartlytherefore forgedbythequalityoftheservicesitdeliverstoitsbeneficiaries.Howeverinthecaseof amarketregulator,therearenoservicestobedeliveredasinthevcaseofschools,or welfareagencies.Acompetitionauthorityadjudicatesandregulates,therebypotentially incurringharmtothegroupsitregulates.Thebeneficiariesareconsumers,butthereisno cleargroupofconsumers,exceptthanwhenrepresentedbysomeconsumerassociation, andtheyarenodirectlyaffectedgroupinrelationtotheservicesorproductswhichthe regulatordelivers.Therefore,thelegitimacyofaregulatorshouldthereforebemeasured intermsoftheloadandeconomicimportanceofthecasesthataredealtwithbythe authority. InthecaseoftheNMA,thetaskofbuildingareputationandsupportingcoalitions 12 DuringthedebateonthebudgetofEconomicAffairs,therewascriticismsraisedagainsttheintentionof theministertoautonomizetheNMasoquickly,withoutwaitingforthescheduledevaluationin2001ofthe CompetitionAct(FD14August1998).Thelargestinterestorganizationofemployers,VNO-NCW,was highlycriticalaswell,deeminganindependentagencythatdefinescompetitionandmarketstoo farreaching(FD15august1998).

23 befellupontheshouldersofthedirector-generals. 13 WhentheNMAwascreated,itwas positionedhierarchicallysubordinatetotheMinisterofEconomicAffairs,theDutch economywasa‘cartelparadise’,andpoliticalpressuresforliberalizingtheDutch economywerehigh.So,althoughtheformalpowersoftheNMAwereslim,it neverthelesshadtostartinthemidstofaperiodwhentheNetherlandshadtoshowtothe OECDaswellastheEuropeanCommission(seebelow),thatitwastomakeserious headwaywithfurtherstrengtheningofitscompetitionpolicies.Astrongreputationand supportivecoalitionswere sinaquanon forthesuccessandsurvivaloftheNMA. Onesteptowardsestablishingitsreputationasacartel-basherwassetduringthe openingceremonyoftheNMAinNovember1997(seeabove).Onthisoccasion,the RoyalDecree(i.e.cabinetorder)bywhichhewasappointedwashandedtohimbythe crownprinceH.M.Willem-Alexanderinfrontofanaudiencethatincludedthe,minister, thenEuropeanCommissionerofCompetition,Mr.K.vanMiert,representativesoflarge lawfirms,andrepresentativesofvirtuallyeverysectorthatwastoberegulatedbythe NMA.Thiswasanhighlyunusualsettingfortheinstallationofanordinaryhighercivil servant.Inadditiontosymbolismofthemeeting,Kistmadeclearinhismaidenspeech that,aswesawabove,consideredconsumersasitsagency’smostimportant constituency.Heconsideredthebreakingdownofcartelsthepathtowardseconomic democracy.Ofcourse,thespeechcontainedhighlysymbolicandnormativereferences, butthemessagewascleartoboththeaudiencethatwaspresentaswellasthebroader public. Inthetimeleftfromprocessingthedispensationrequests,theNMashowedits teethinanumberof‘pro-consumer’cases.IntheendofFebruary,theagencyannounced tocloselymonitortheintendedmergerbetweenNetherlandstwolargestconference centers(RAIinAmsterdamandJaarbeursinUtrecht)andinonesweepitannouncedto investigatetheallegedlydominantpositionoftheNVM,thelargestassociationofreal estateagentsintheNetherlands(FD,26and28February1998,respectively).Afew monthslater,theNMastruckagainwithitsadvicetothegovernmenttoprohibitthe

13 Adirectorate-generalisanordinarypositionwithinthehigherranksoftheDutchcivilservice.In general,aministryhastwoorthreedirectorates-generalandthedirectors-generalsformthetop administrativemanagementofaministrytogetherwithasecretary-general.

24 alliancebetweentheJointAdministrationOffice(GAK),oneofthefiveprivatesocial welfareexecutiveagencies,andthebankingandinsurancecompanyAchmea;itwasthe firsttimethatcompetitionagencydidadvisethegovernmenttoprohibitconcentration (FD,19June1998).Withinthesameperiod,NMaactedproactivelywhenithadthe MinisterofEconomicAffairsaskCommissionerVanMierttorevertthehandlingofthe intendedmergerbetweenVendexandKBB,twoofthelargestretailcompaniesholding largewarehouses,totheNMainsteadofhandlingthiscasebytheCommissionitself;the effectsofthemerger,theNMaargued,wereprimarilynational(FD,23April1998).And asifinpassing,theNMablockedthemergeroftwochemist’sretailers,Internatio-Muller andBrocacef(FD,18July1998).Butalsosectorsthathadnotthoughtaboutthemselves ashavinganythingtodowithmarketprinciples,wereaptlydeclaredtobebusinessesand weredeclaredtofallunderthejurisdictionoftheNMA.Thishappenedtotradeunions andhospitalsastheNMAannouncedthattheiractivitiesmadethempartofthe jurisdictionoftheCompetitionAct(FD,11June1998).Togiveafinalexample,in March2003,theNMaimposedafineonanumberofshrimpfishingfirms.Thisfineof4 millioneuroscausedasharpdeclineintheshrimpprices(from4to1,61euro),andas couldbeexpectedheavyprotestaccompaniedbybarricadingharborsandferriesbetween themainlandandthenorthernislandsbyproducers’interestgroupsinthefisheries.Itled toaseriesofparliamentaryquestionswhichweresupportedbythemajorityoftheMPs (cf.HandelingenII,2002-2003,Aanhangsel:1887).Thequestionswereaddressedtothe StandingCommitteeofAgriculture,NaturalResourcesandFoodQualitytoitsminister, butsoonthequestionspilledovertotheEconomicAffairsdepartment.TheNMa’s decisionhadbecomeapoliticalissuefortheministry.Unimpressedbytheseprotestsand parliamentaryupheavals,theNMainsteadofrevisingitsfuturelinetowardsthissector underthepressurefromthesector, 14 tolddirectlytotheMinistryofAgriculturethat“you shouldradicallyreformthatsector”(respondentAfromNMa). ThereputationoftheNMA,despiteoftheactivism,receivedaheavyblowfrom theconstructioncartelaffair.Theaffairprovedacriticalopportunityfortheopponentsof anindependentstatusoftheNMatodownplaythereputationoftheNMAasacartel-

14 “TheNMakeepscriticallymonitoringtheshrimpfisheriessector”,3December2006,on Visserijnieuws.nl ,at http://www.visserijnieuws.nl/ ,consultedon25July2007.

25 basher.Likemanyothercritics,theparliamentaryleaderoftheLaborParty,Mr.Ad MelkertmockinglysaidthattheNMaappearedtobetterathandling“barbersand physiotherapists”thanthebigcartels(FD16November2001).Thedirector-general defendedhimselfandhisagencybysayingthatthesecasesaredifficulttoproveandthat theNMalackscrucialpowers,e.g.toenterprivatehousesforsearchingevidence,to conductthiskindofinvestigationproperly.Butthiswastonoavail:thereputationofthe NMAwashemorrhaged,onlytoberestoredtosomeextentbyapositivereviewthe NMAreceivedfromtheGeneralAccountingOffice(ARKxxxx). AlternativeExplanationsandCriticalEvents

CONSTRUCTION CARTELS AFFAIR .On9November2001,thetelevisionnewsmagazine Zemblabroughtthenewsofahugecartelsintheconstructionsector.Thejournalists disclosedtheexistenceof‘shadowaccounts’inwhichallofthelargestDutch infrastructurebuilderswerementioned.Togethertheyshouldhaveriggedtheirbiddings ongovernmentalpublicworksprojectsonamassivescale. 15 Whatismore,the documentaryfurtherdisclosedthattheNMa,togetherwiththePublicProsecutor’s Office,hadbeenapproachedbyawhistleblower,hadalreadybeenapproachedand informedbyformerdirectorofoneoftheconstructionfirmsmorethantwoyearsago,but bothpublicorganizationshadnotundertakenanyactionagainstthecartel.Parliament reactedaghastatthesefacts.ItraiseddoubtsabouttheprioritizationsoftheNMaandthe agency’sinternalproceduresofprocessingexternalsignals,i.e.thatithadfocuseditself on“smallfish”andneglectedthelargercasestheduringthepastyears(FD15November 2001). Thereasonsforwhytheconstructionsectorhadnotbeengivetheutmostpriority hithertowererevealedduringtheinvestigationsofparliamentaryenquirycommittee (KamerstukkenII,2002-2003,28244,nr.12).First,ofthetotalof824informalsignals theNMareachedbetween1998and2001,notmorethan10camefromtheconstruction sector;ofthe454formalcomplaintsonly8werefromthesamesector.However,

15 Estimationsofexpertswasthatthebuildershadincreasedtheirpricewithabout8%inaverage.

26 measuredbythecriteria 16 theNMaemploysthedeterminethepriorityofthesesignals, noneofthemwereofsuchqualitythatitrequiredfurtherinvestigationbytheNMa.A secondreasonwasthatitwasn’tthepriorityoftheNMaatthattimetoactuallycheckthe informationortomonitorthenewCompetitionAct.AsonecivilservantattheNMatold thestaffoftheparliamentarycommitteein2002,iftheNMahaddecidedtomaintainthe act“theorganizationhadtomakeacompletedifferentprioritization”ofitstask (KamerstukkenIII,2202-2003,28244,nr.12:151),i.e.insteadofgivingprioritytothe assessmentofdispensationrequests,itshouldhaveallocatedthebulkofits organizationalcapacitiestomaintenanceofthelaw.Finally,thecapacitytheNMadid retainnexttothedispensationrequesttomaintainthelaw,itusedthisforcaseswitha highconsumers’interests,suchasthepresumedverticalpriceagreementsconcerninggas betweenservicestationsandtheiroilcompanies.Besideshavingaconsumers’profile, priceagreementsinthissectorwereverydifficulttoproveandwasareal“energy usurper”,asonerespondenttoldthecommittee. Followingtheaffair,theNMamadecertainalterationstoitsagendaandagenda- settingprocedures.First,itimmediatelyprioritizedtheconstructionasitsetupaspecial projectteamtoinvestigatetheconstructionsector(FD23January2002)andopeneda specialphonenumberfor(anonymous)informationoncartels.Kistalsoannounceda newproceduretosettheagenda.DuringthepresentationoftheNMa2001yearreport, Director-GeneralKistannouncedthattheNMawouldfromnowoninvolvestakeholders, i.e.Consumers’Association,inthesettingofitsagency’sagendainorderto“involve societyandtoincreasethetransparencyoftheNMa”(FD28June2002).TheNMa wouldremainresponsiblefordecidingwhichsectorsandcaseswouldbeinvestigated.

EC REGULATION 1/2003. Since1May2004,competitionpolicieswithintheEChave beenfundamentallyaltered.SincetheenactmentofRegulation1/2003,national competitionauthoritieshavebecomeresponsibleforthedirectimplementationofArticles 81and82EC.IntheDutchcase,thishashadatleastformallyconsequencesforthe authorityoftheministertointerveneinnationalanti-trustissues.Specifically,thisnew 16 Theformalcriteriaare1)economicinterests,2)consumerinterests,3)theseverityoftheoffence,and4) theexpectedeffectivenessoftheNMatoaddressthecase.See http://www.nmanet.nl/nederlands/ home/Actueel/Themadossiers/Procedures/index.asp ,consultedon25July2007.

27 regulationmeansthattheauthorityoftheministertoprescribegeneralguidelinesas regardsanti-trustmeasures,whicharegrantedtotheministeronthebasisofArticle5d oftheDutchCompetitionAct,havepracticallybecomevoid(VerheijandVerhey2005: 218).ThesamegoesforArticle5dofthesameactwhichstatesthattheministercangive theNMainstructionswhenevertheNMaattends(advisory)meetingsofcompetition relatedcommitteesatBrussels–thisaccordingtoonepubliclawyer,Sauter(2002, IN vandegronden&widdershoven ),isagainsttheauthorityoftheCommissioninEC competitionpolicies.Regulation1/2003notonlydelegatestheauthoritytonational competitionauthoritiestoimplement81and82EC,butithasalsoestablishedthe EuropeanCompetitionNetworkwiththemaingoalofharmonizingECcompetition policiesthroughregularmeetingsanddiscussionswithallofthenationalcompetition authorities.Theinfluenceoftheministerisshrinking,then,themorecompetitionpolicies becomeharmonizedattheEClevel.Theempiricalquestionthatinterestshereis:howfar arewe?ToanswerthisquestionwewillexaminetheextenttowhichtheNMahasbeen tiedtoandembeddedwithinEuropeannetworksofcompetitionpolicymaking.Firstwe willlookattheperiodpreceding1/2003,whichisactuallyinforcesince1May2004. 17 ThenwewillexaminetotheperiodfollowinMay2004. FromtheNMa’sfirstyearreport(1998)welearnthatithadalreadybeenactive intheadvisorycommitteesonanti-trustandmergers,aswellasthattheagencyhad partakeningeneral(expert)committees.IthadforthesepurposesestablishedaEU workinggroupwithintheagencythatcoordinatedtheagency’sEUactivities.Inaddition thedirector-generalpartookeachyearsteadilytotheDGconferencewhichwereattended bytheDG’sofalltheparticipatingnationalcompetitionauthorities.From2001 onwards,theattendancetotheseformalmeetings–wheretheNMatogetherwithEAand perhapsothersdepartmentsrepresentstheDutchpositionwithregardtocompetiytion policies–theNMastartspartakingwithinseveralinformalworkinggroups,ofwhichthe workinggrouponleniencypoliciesisperhapsoneofthemostimportantones.In2000 alreadytheNMapartakesinthefirstproposalstoharmonizeleniencypoliciesacrossthe EC.In2001,thisresultsintheestablishmentofaworkinggroup.Inaddition,another workinggroupiscreated,oneonthe‘multi-jurisdictionalmergers’,withtheaimtocome 17 ThedataisfromNMayearreportsunlessreportedotherwise.

28 toguidelinesonhowtohandlecross-bordermergercases.Attheendof2002,a EuropeanCompetitionNetwork(ECN)isestablished.Duringtherunuptothe modernizationoftheECcompetitionpolicy,TheECNissetupbytheCommissionwith thegoaloffurtherharmonizingthepoliciesandimplementationofcompetitionpolicies acrossthememberstates.In2003,theCommissionpublishesanumberof AnnouncementsconveyingguidelinesforhowtoapplytheECprovisions. 18 Theenactmentof1/2003bringschangestotheworkingproceduresoftheNMa. Thenewregulationnowprescribestoapplynotonlynationalcompetitionlawbutalso ECcompetitionlawtocasesthataffecttradebetweentwoormorememberstate countries.TheNMaisfurtherobligedtoinformtheCommissionaboutnew investigationsanddraftsofdecisionsinindividualcases,sothattheCommissioncan checkthatdecisionsmadebynationalauthoritiesareinlinewitheachotherandwiththe ECrules.Also,theregulationenablesunderstrictconditionsthesharingofconfidential informationbetweennationalcompetitionauthoritiesandtheCommission.Finally,the regulationdelegatestoNCAstheauthoritytoadvisenationalcourtsonhowtointerpret andapplyECCompetitionlegislation(as amicaecuriea ).Theseareformalpowers, whetherusedornot,thatenormouslyenhancethebureaucraticautonomyoftheNMavis- à-vistheDutchgovernment. DiscussionandConclusion InthispaperIcomparedandtestedthemainthesesoftwocompetingperspectiveson bureaucraticautonomyonasinglecasestudyoftheagenda-settingprocesswithinthe DutchCompetitionAuthority.Allthedisclaimersofacasestudyapply.Following RodgersandJensen,Ipresentedthecaseinawaythatcanbestandardizedandbe comparedwiththefindingsofothercasestudies.Morethanthis,however,Iexploredthe tenabilityofthemainlyUSbasedtheoriesofbureaucraticautonomywithinacasestudy ofcompetitionpoliciesandregulationwithinaparliamentarysystem.Itisnowtimeto assesstherelativeimpactofeachthegroupsofvariables. Staringwiththeindependentvariablesfromtheexternalcontroltheories.The 18 Seehttp://ec.europa.eu/comm/competition/antitrust/legislation/legislation.html

29 appointmentofanagencyiswithgreatconsensusamongUSscholarspointedoutasthe maininstrumentofpoliticalcontrolthatpoliticianshavetotheiravailtoinfluencethe agendaanddecisionsofanagencyundertheircontrol.Inthiscasestudy,Ifoundonlyan indirecteffectofappointmentauthority.Withindirectismeantthatthedirectorofan agency,heretheDirector-generaloftheNMA,doeshaveanimpactontheagendaofhis orheragency,butthatpartisanshipdoesnotplayaroleintheappointmentprocess.As regardsagencydesign,itisexpectedthatagenciesthataredesignedtobecome independentfromtheprincipalwillalsohavemoreactualautonomythanagenciesthat arenotformallyindependentfromtheprincipal.Herewesawthatevenduringtheperiod theNMAwasnotindependentfromtheMinisterofEconomicAffairs,itenjoyeda measureofautonomyparalleltothattheNMAenjoyedwhenithadbecomeindependent in2005.Themainreasonforthisistwofold.Firstsubsequentministers’heldhighesteem ofbuildingandmaintainingtheircrediblecommitmenttowardstheindustry;secondly independencysavedthemaloadofmonitoringtheagency,whichtheyotherwisehad beenforcedtobythepressuresofindustrylobbiesandparliament.HerewithIalready touchedtheroleofinterestgroups:theseexertedlargepressuresontheminister,buthad verylittleimpactontheNMA’sagenda.OnlyoneinterestclaimedtheNMAtodefend: thatofconsumers.Inconclusion,wefindverysmallsupportforexternalcontrol explanationswithincasestudy. Bycontrast,wefindmoresupportforadministrativebehaviourtheses.Political legitimacy,accordingtoadministrativebehaviouraltheories,understoodastheexistence ofsolidbeliefsheldbycitizens,interestgroupsandpoliticiansalikeaboutthecapacities oftheorganisationandtheexistenceofsolidandstrongadvocacycoalitionsaroundthe organizationisthesteppingstoneforbuildinganautonomybase.Thebeliefsmustbe raisedandthecoalitionsbebuilt,whichisthetaskoftheagency’smezzo-levelorinthis casethatofthedirector-general.ThisstudyhasshownthatfromthebeginningtheNMA leadershiphastriedtoforgeanimageofcartel-basher,guardiansofconsumerinterests andkeepersofeconomicdemocracy.Thissuccessfulimageframingwasreflectedinthe prioritizationofsectorswithhighconsumersalience:healthcare,housingmarket, fisheries,andretail.However,theNMA’sreputationasapro-consumeradvocatewas shatteredaftertheconstructioncartelsaffairswasless.Byitsverynature,itwasalsoless

30 successfulinforgingnetworks:amulti-industryregulatorhasmoreenemiesthanfriends. Inthiscase,wecouldarguethatforregulatorscatchingthe‘bigfishes’andhandling thesecaseswithoutmakingmistakesisamoreseriouspreconditionthanembeddingitself broadnetworks. Perhapsthemostsubstantialimpactonthepoliticsofagenda-settingwithinthe NMAwasthesingleeventofthedisclosureofcartelscandalwithintheconstruction industry.Ofcourse,thisevent’seffectworkeditselfoutthroughpoliticallegitimacy:the imagethattheNMAhasmissedthegiantcartelsinthissectordidraisepeople’sdoubts aboutthereputationandtheorganizationalcapacitiesoftheNMA.Sothisevent underscoresthesalienceofadministrativetheoriesofbureaucraticautonomy.Thisseems tobedifferentfortheeffectsthatECRegulation1/2003had.ThedecentralizationofEC competitionpolicyexertseffectsintwodirections.Thefirstdirectionisthatthefurther Europeanizationofcompetitionpoliciesgivesevenlesserroomforministerialand parliamentaryinterventionintheagendasoftheNMAanditscounterpartsinother memberstates.WithEC1/2003eventheministerialpowerofissuingageneral instructiontotheNMAinrelationtohowitshouldaccountfornon-economicinterests withinanti-trustpolicieshasbecomeobsoleteinthelightofCommunitylaw.The seconddirectionisthattheNMAwillbecomeembeddedinastrongcoalitionafterall, butnowthroughtheEuropeanCompetitionNetwork.Thestrengtheningofthisnetwork andthegrowingcooperationbetweentheEuropeanCommissionandthenational competitionauthoritieswillleadtoagreatershieldingoftheNMA’sagendafrom nationalpressuresandconcerns. Thefindingsofthiscasestudysuggestfurtherresearch.Towhatextentarethe findingsforexternalcontrolvariablesattributabletoinstitutionalsettingofparliamentary system,i.e.theNetherlands?Checksandbalancesarenotalientoparliamentarysystems, butmultiplepoliticalchecksasthatintheUSbetweenCongressandthePresidentis absentintheNetherlands.Doesthisimplythatadministrativebehaviouraldeterminants ofbureaucraticautonomywithinparliamentarysystems‘genetically’built-in?Further enquiryisneededoftheEuropeanizationofcompetitionpolicyfortheautonomyof nationalregulators.Inthislightitwouldbeinterestingtodelvefurtherintotheseveral emergingmulti-levelnetworksofagencies.

31 References Andeweg,RudyB.(2000),‘MinistersasDoubleAgents?TheDelegationProcess BetweenCabinetandMinisters’, EuropeanJournalofPoliticalResearch ,37,pp.377- 395 Andeweg,RudyB.andGalenA.Irwin(2005). GovernanceandPoliticsofthe Netherlands. Basingstoke:PalgraveMacMillan. Balla,StevenJ.(1998),‘AdministrativeProceduresandthePoliticalControlof theBureaucracy’, AmericanPoliticalScienceReview ,92:663-673. Beck,Nathaniel(1982),‘PresidentialInfluenceontheFederalReserveinthe 1970s’, AJPS ,26:415-445 Bendor,JonathanandTerryM.Moe(1985),‘AnAdaptiveModelofBureaucratic Politics’, AmericanPoliticalScienceReview ,79:755-774 Calvert,Randall,MatthewMcCubbinsandBarryWeingast(1989),‘ATheoryof PoliticalControlofAgencyDiscretion’, AmericanJournalofPoliticalScience ,33:588- 610. Carpenter,DanielP.(2001). TheForgingofBureaucraticAutonomy . Reputations, Networks,andPolicyInnovationinExecutiveAgencies,1862-1928 .Princeton:Princeton UniversityPress. Christensen,JørgenG.(2001)‘Bureaucraticautonomyasapoliticalasset’,inB. G.PetersandJ.Pierre(eds), Politicians,BureaucratsandAdministrativeReform , London:Routledge,pp.119-131. Christensen,TomandPerLægreid(eds.)(2006). AutonomyandRegulation. CopingwithAgenciesintheModernState .Cheltenham:EdwardElgar. Clark,PeterB.andJamesQ.Wilson(1961),‘IncentiveSystems:ATheoryof Organizations’, AdministrativeScienceQuarterly ,6(2):129-166. Cohen,MichaelD.,JamesG.March,andJohanP.Olsen(1972),‘AGarbageCan ModelofOrganizationalChoice’, AdministrativeScienceQuarterly ,17(1):1-25. Downs,Anthony(1967). InsideBureaucracy .Boston:Little,BrownandCo. Epstein,DavidandSharynO’Halloran(1999). DelegatingPowers.ATransaction CostsApproachtoPolicyMakingunderSeparatePowers .Cambridge:Cambridge UniversityPress.

32 Furlong,ScottR.(1998),‘PoliticalInfluenceontheBureaucracy.The BureaucracySpeaks’, JournalofPublicAdministration,ResearchandTheory ,8(1):39- 65. Hammond,ThomasH.andKnott,JackH.(1996),‘WhoControlsthe Bureaucracy?:PresidentialPower,CongressionalDominance,LegalConstraints,and BureaucraticAutonomyinaModelofMulti-InstitutionalPolicy-Making’, Journalof Law,Economics,andOrganization ,12(1):121-168. Hammond,ThomasH.andChristopherK.Butler(2003),‘SomeComplex AnswerstotheSimpleQuestion‘DoInstitutionsMatter?’’, JournalofTheoretical Politics ,15(2):145-200. Hooghe,Liesbet&GaryMarks(2001). Multi-LevelGovernanceandEuropean Integration .NewYork:Rowan&Littlefield. Howell,WilliamG.andDavidE.Lewis(2002).‘AgenciesbyPresidential Design’,JournalofPolitics ,64(4),pp.1095-1114. Huber,JohnD.andCharlesR.Shipan(2002). DeliberateDiscretion?:The InstitutionalFoundationsofBureaucraticAutonomy ,CambridgeUniversityPress. Jensen,JasonL.andRobertRodgers(2001),‘CumulatingtheIntellectualGoldof CaseStudyResearch’, PublicAdministrationReview ,61(2):235-246 Jong,H.W.de(1990),‘Nederland:hetkartelparadijsvanEuropa?’, EconomischStatistischeBerichten ,72:192-7. Kaufman,Herbert(1960).TheForestRanger.AStudyinAdministrative Behavior.Baltimore:JohnHopkinsPress. Kiewiet,D.RoderickandMathhewD.McCubbins(1991).TheLogicof Delegation.CongressionalPartiesandtheAppropriationsProcess.Chicago:University ofChicagoPress. Krause,GeorgeA.(1996),‘TheInstitutionalDynamicsofPolicyAdministration: BureaucraticInfluenceoverSecuritiesRegulations’, AmericanJournalofPolitical Science ,1083-1121. Krause,GeorgeA.andJamesW.Douglas(2005),‘InstitutionalDesignversus ReputationAffectsonBureaucraticPerformance:EvidencefromU.S.Government

33 MacroeconomicandFiscalProjections,’ JournalofPublicAdministration,Researchand Theory ,15(2):281-306 Levine,MichaelE.andJenniferL.Forrence(1990),‘RegulatoryCapture,Public Interest,andthePublicAgenda:TowardaSynthesis’, JournalofLaw,Economics,& Organization ,SpecialIssue:PapersfromtheOrganizationofPoliticalInstitutions Conference,April1990,6:167-198 Lewis,DavidE.(2003).PresidentsandthePoliticsofAgencyDesign.Political InstitutionsintheUnitedStatesGovernmentBureaucracy,1946-1997.Stanford:Stanford UniversityPress. Lijphart,Arend(1999).PatternsofDemocracy:GovernmentFormsand PerformanceinThirty-SixDemocracies.NewHavenandLondon:YaleUniversityPress. Macey,JonathanR.(1992),‘OrganizationalDesignandPoliticalControlof AdministrativeAgencies’, JournalofLaw,EconomicandOrganization ,8:93-110 McCubbins,MatthewD.(1985),‘TheLegislativeDesignofRegulatory Structure’, AmericanJournalofPoliticalScience ,29:721-748. McCubbins,MatthewD.,RogerG.Noll&BarryR.Weingast(1987), ‘AdministrativeProceduresasInstrumentsofPoliticalControl,’ JournalofLaw, EconomicsandOrganization ,3:243-277 Miller,GaryJ.(2005),‘ThePoliticalEvolutionofPrincipal-AgentModels’, AnnualReviewofPoliticalScience ,8:203-225. Moe,TerryM.(1982),‘RegulatoryPerformanceandPresidentialAdministration,’ APSR ,28:739-777 Moe,TerryM.(1985),‘ControlandFeedbackinEconomicRegulation:TheCase ofNLRB’, APSR ,79:1094-1116. Moe,TerryM.andMichaelCaldwell(1994).TheInstitutionalFoundationsof DemocraticGovernment:AComparisonofPresidentialandParliamentarySystems. JournalofInstitutionalandTheoreticalEconomics ,150(1),pp.171-195. Müller,WolfgangandKaareStrøm(eds.)(2003). CoalitionGovernmentsin WesternEurope .Oxford:OxfordUniversityPress.

34 OECD(1998), BackgroundReportomCompetitionRegulationandRegulatory Reform .Paris:OECD.ConsultedintheInternet: www.oecd.org/dataoecd/3/42/2497317.pdf ,20July2007 OECD(1999). RegulatoryReformintheNetherlands .OECDReviewsof RegulatoryReform,ParisOECD. Ringquist,EvanJ.(1995),‘PoliticalControlandPolicyImpactinEPA’sOfficeof WaterQuality’, AmericanJournalofPoliticalScience ,39:336-363. Selznick,Philip(1957). LeadershipinAdministration.ASociological Interpretation .NewYork:Row,PetersonandCompany. Stigler,GeorgeJ.(1971),‘TheTheoryofEconomicRegulation’, BellJournalof EconomicsandManagementScience ,2:3-21. Strøm,Kaare(2000),‘Delegationandaccountabilityinparliamentary democracies’, EuropeanJournalofPoliticalResearch ,37(3),pp.261-290. Verheij,LucasF.M.andNicoVerhey(2005),‘Demachtvanmarktmeesters. Markttoezichtinconstitutioneelperspectief’.In:AngelicaA.vanRossum,LucasF.M. VerheijandNicoVerhey, Toezicht/PreadviezenvanA.A.vanRossum,L.F.M.Verheijen N.Verhey .Handelingen Nederlandse Juristen-Vereniging(135,2005-1).Deventer: Kluwer Verhoest,Koen,Peters,GuyB,Bouckaert,Geert,andVermeulen,Bram(2004), ‘TheStudyofOrganisationalAutonomy:AConceptualOverview’, Public AdministrationandDevelopment 24(2):101-118. Waarden,Fransvan(1992),‘DimensionsandTypesofPolicyNetworks’, EuropeanJournalofPoliticalResearch ,21(1/2):29-52. Waterman,RichardW.,AmandaRouse,andRobertWright(1998),‘TheVenues ofInfluence:ANewTheoryofPoliticalControloftheBureaucracy’, JournalofPublic AdministrationResearchandTheory ,8(1):13-38. Weingast,BarryandMarkMoran(1983),‘BureaucraticDiscretionor CongressionalControl?RegulatoryPolicymakingintheFederalTradeCommission’, JournalofPoliticalEconomy ,91:765-800 Whitford,AndrewB.(2002),‘DecentralizationandPoliticalControlofthe Bureaucracy’, JournalofTheoreticalPolitics ,14(2):167-193.

35 Wilson,JamesQ.(1989). Bureaucracy.WhatGovernmentAgenciesDoandWhy TheyDoIt .NewYork:BasicBooks Wood,DanB.(1988),‘Principals,BureaucratsandResponsivenessinCleanAir Enforcement’, AmericanpoliticalScienceReview ,82:213-234. Wood,DanB.andRichardWaterman(1991),‘TheDynamicsofPoliticalControl oftheBureaucracy’, AmericanPoliticalScienceReview ,85:801-828.

36