Does the independence of independent directors matter

2021-02-24 来源：个人技术集锦

CORFIN-00757;NoofPages19

JournalofCorporateFinancexxx(2013)xxx–xxx

ContentslistsavailableatScienceDirect

JournalofCorporateFinance

journalhomepage:www.elsevier.com/locate/jcorpfin

Doestheindependenceofindependentdirectorsmatter?

RafelCrespí-Cladera1,BartoloméPascual-Fuster⁎

Departamentd'Economiadel'Empresa,UniversitatdelesIllesBalears,Cra.deValldemossakm7.5,07122Palma(IllesBalears),Spain

articleinfoabstract

Thispaperanalyzesthecharacteristicsoffirmsthatdeclareboarddirectorsasindependents,althoughthedirectorsarenotstrictlyindependent,andexaminestheconsequencesintermsofperformanceandcorporategovernanceoutcomes.Basedonpubliclyavailableinformation,eightcriteriaof“independence”usedtoexamineapanelofSpanishlistedfirmsclassify14.2%ofthedirectorsasstrictlyindependent,whereasthefirmsclassify32.5%oftheboardasindependentdirectors.Firmswithdispersedownershipstructuresmisclassifydirectorsmorefrequentlythandofirmswithlargecontrollingowners.Intermsofconsequences,wefindweakevidenceofanegativerelationbetweenmisclassificationandafirm'sfutureoperatingperformance.However,norelationisfoundbetweenindependents'misclassificationandseveralrelevantoutcomesoftheprimarydelegatedcommitteeswithmonitoringroles:theauditcommitteeandthenominationandremunerationcommittee.Thereisnosignificancewithregardtothenon-strictlyindependentmeasuresexplainingexecutivedirectors'compensation,CEOturnover,auditqualificationsorearningmanagementbehavior.

Articlehistory:

Received26September2013

Receivedinrevisedform26November2013Accepted9December2013AvailableonlinexxxxJELclassification:G30G34K22

Keywords:

BoardstrictindependenceCorporategovernanceExecutivecompensationCEOturnover

AuditqualificationsEarningsmanagement

1.Introduction

Independentdirectorsareperceivedasminimizingthepotentialopportunismofmanagers,orlargecontrollingowners,inaprincipalagentsetting.InthecontextofAnglo-Americandispersedownership,independentsarekeymembersoftheboardinavoidingmanagerialmisappropriation.IntheconcentratedownershipsettingofcontinentalEurope,theroleassignedtoindependentdirectorsistolimittheextractionofprivatebenefitsbythecontrollinglargeshareholders,whousuallyappointtheremainingmembersoftheboardofdirectors.Theunderlyingfundamentalconceptisthatthemonitoringactivityoftheboardroomdependsontheeffectivenessoftheindependentmembers.Thisview,whichiswidelyacceptedintheacademicworld(AdamsandFerreira,2007;Adamsetal.,2008;BhagatandBlack,2002;DyckandZingales,2004;HermalinandWeisbach,2003),isinthespiritofregulationssuchastheOECDPrinciplesofCorporateGovernanceof2004,theCommissionoftheEuropeanCommunitiesRecommendation,2thefinalcorporategovernancerulesoftheNewYorkStockExchangeof2009,andnearlyallexistingcorporategovernancecodesorguidelines.Effectively,theEuropeanCommissionRecommendationof2005assertsthatindependentdirectorsplayaroleinbothdispersedownershipcompanies,inwhichtheconcernishowtomakemanagersaccountabletoshareholders,andincompanieswithlargecontrollingshareholdersthatmustaccountfortheinterestsofminorityshareholders.

⁎Correspondingauthor.Tel.:+34971172652.

E-mailaddresses:rafel.crespi@uib.es(R.Crespí-Cladera),tomeu.pascual@uib.es(B.Pascual-Fuster).1Tel.:+34971171323.2TheCommissionoftheEuropeanCommunitiesRecommendationof15February2005ontheroleofnon-executiveorsupervisorydirectorsoflistedcompaniesandonthecommitteesofthesupervisoryboard.

Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

2R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Inlinewithbestpracticescodesandguidelines,themajorityofcompaniesreportincreasingproportionsofindependentsamongtheirboardmembers(Gordon,2007;Lincketal.,2009).Ratingagenciesalsoaccountforthepresenceofaqualifiednumberofindependentdirectorsasanelementinagencyratingoutputs,asSantellaetal.(2006)indicate.Onelong-termperspectiveregardingthechangingtrendofexecutivesandindependentsinUSboardroomscomesfromGordon(2007),whoindicatesthatfrom1950to2005,theaveragepercentageofindependentdirectorshasrisenfrom20%tolevelsabove70%.

Usingpubliclyavailableinformation,ourpapercomparesfirms'declarationsregardingindependentdirectorswithanunbiasedmeasureofstrictlyindependentboardmembersforapanelofSpanishlistedfirms.Inadditiontothecontributionofthedescriptiveinformationregardingthefirms'deviationbetweendeclaredandstrictlyindependentdirectors,thepaperalsoaddressesthefollowingtwomainquestions:

Question1:Whatarethecharacteristicsoffirmsthatmisclassifyindependentdirectorsmorefrequently?

Question2:Whataretheconsequencesoffirmsdeclaringdirectorsasindependentswhenthesedirectorsarenotstrictlyindependent?Thispaperwillexaminefirms'futureoperatingperformanceandtheoutcomesofthemainboardcommitteesintermsofmonitoring:theauditcommitteeandthenominationandremunerationcommittee.Anynegativeimpactonfirms'performanceoronthecorporategovernanceoutcomeswouldmeanthatthereisanagencyproblemthatshouldbecorrectedwithlawenforcement.Inaddition,thelackofeffectofindependentdirectors'misclassificationonfirms'performancewouldindicatethattheusualrecommendationsoflargeproportionsofindependentsshouldbereexamined.

Thepreviousempiricalliteratureaddressestherelevanceandtheroleofindependentboardmembers,usuallybyconsideringthenumberofdirectorsdeclaredasindependent(e.g.,Booneetal.,2007;Colesetal.,2008;Duchinetal.,2010;Kimetal.,2007;Lincketal.,2008).Noquestionregardingthequalityofthedirectors'independencewasexaminedinthesepapers.However,severalotherpapersaddressthequalitativeaspectoftheindependenceoftheboardofdirectors.IndependentdirectorswhojoinedtheboardaftertheCEOareassumedtobelessindependentinCoreetal.(1999)andinColesetal.(2010).Santellaetal.(2006)quantifytheextenttowhichcorporatedisclosureforthefinancialyear2003allowsforverificationoftheindependenceofdirectorsformallydeclaredasindependentsbythe40Italianbluechipfirms.Santellaetal.(2007)extendthepreviousdescriptiveresultscomparingfinancialandnon-financialfirmsintermsofownershipstructure.Byrdetal.(2009)examinesamorebehavioraldefinitionofdirectorindependencetoexcludeproblematicdirectors,whoareunlikelytostanduptomanagement,totestthedirectors'impactonfirmperformanceandmanagementcompensation.HwangandKim(2009)analyzesocialindependencebetweenindependentdirectorsandtheCEOandtheimplicationsforCEOcompensation,payperformancesensitivityandturnover.Cohenetal.(2012)analyzefirmsthatappointindependentdirectorswhoareoverlysympathetictothemanagement.Finally,Ansarietal.(2013)inthisspecialissue,analyzethelinksofindependentdirectorswiththecontrollingfamilytounderstandthedeterminantsofCEOsuccessorchoiceinfamilyfirms.

Ourcontributiontothebodyofknowledgeregardingstrictboardindependence,includingdeterminantsandconsequences,isthreefold.First,basedonthebestpracticesdefinitionsofindependentdirectors,weprovideatestablesetofeightformalindependencecriteriaandmakethesetoperationalforapanelofSpanishlistedcompanies.Weevaluatethestrictindependenceofeachdeclaredindependentdirectorandfindarelevantdifferencebetweenstrictindependenceanddeclaredboardindependence.Second,weestimateamodeltodisentanglethecharacteristicsofthecompaniesthatmisclassifytheirindependentdirectors.Wefindthattheproxiesofmanagerialcontrolincreasewiththepresenceofnon-strictlyindependentdirectors.Third,wetesttheimpactofthismisclassificationonfirms'futureoperatingperformanceandonfirms'corporategovernancepractices,withparticularattentionontheimpactofmisclassificationintheauditcommitteeandinthenominationandremunerationcommittee.Wefindonlyweakempiricalevidenceofanegativeeffectonfutureoperatingperformance.

Thispaperbeginswithadiscussionoftheacademicliteratureregardingtheroleofboardindependenceandrecommendationsregardingthecodesofgoodgovernance.ThenextsectionlinkstheinternationalrecommendationsandtheSpanishregulationwiththeproposedempiricalmeasureofindependence.Section4includesthedatadescriptionandthemostrelevantstatistics.Thegovernancecharacteristicsoffirmswithnon-strictlyindependentdirectorsarediscussedinSection5.Section6analyzestheconsequencesofthepresenceofnon-strictlyindependentdirectorsonfutureoperatingperformance,executivedirectors'compensation,CEOturnover,auditqualifications,andearningsmanagement.Section7presentsthediscussionofresultsandSection8concludes.

2.Independentdirectorsinthecodesofgoodgovernanceandthepreviousliterature

Anincreasingnumberofcorporategovernancecodesgloballyproposeexplicitrecommendationsregardingthestructureofboardsofdirectors,andmorespecifically,regardingthedesirableproportionthatindependentdirectorsshouldhold.AguileraandCuervo-Cazurra(2009)indicatethat46countriesissued196distinctcodesstemmingfromthefirstcodeof1978,includingtherelevant1992UKCadburyReportandleadinguptothemiddleof2008.Theseresearchersindicatethatthemajorityofcodeshaveexplicitorimplicitrecommendationsonsixgovernancepractices;thefirstpracticeexaminesthebalanceofexecutiveandnon-executivedirectors,whichincludestheindependentnon-executivedirectors.

ThecorporategovernancerulesoftheNewYorkStockExchangeapprovedbytheSECon2003(amendedon2009)andcodifiedinSection303AoftheNYSE'sListedCompanyManualdescribethefirstruleasfollows:“Listedcompaniesmusthaveamajorityofindependentdirectors”.Thejustificationisthat“effectiveboardsofdirectorsexerciseindependentjudgmentinPleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx3

carryingouttheirresponsibilities.Requiringamajorityofindependentdirectorswillincreasethequalityofboardoversightandlessenthepossibilityofdamagingconflictsofinterest”.

Indeed,theproportionofindependentdirectorshasincreasedovertime.Gordon(2007)indicatesthatthefocusonshareholdervalueandstockmarketpriceshavecontributedtosuchanincrease.Morerecently,Lincketal.(2009)notedthewayinwhichtheSarbanes–OxleyActintheUSsubstantiallyincreasedthepresenceofindependents,indicatingthatasubstantial(althoughsmaller)increasewasalsodetectedinfirmsforwhichthisactwasnotmandatory.However,althoughthepreviousliteratureanalyzingtheinfluenceofindependentdirectorsoncompanies'behaviororperformanceisextensive,theliteratureisnotconclusive.Ontheonehand,somestudiessupportpositiveoutcomes.Forexample,ByrdandHickman(1992)showapositiverelationbetweenindependentoutsidedirectorsandshareholders'interests.Cotteretal.(1997)findthatshareholders'wealthincreasedduringtenderofferswhenalargerproportionofindependentdirectorswereontheboard.WeirandLaing(2000)reportthatfortheUK,bestcorporategovernancepracticesoccurwhenindependentdirectorsplayanimportantrole.Benkeletal.(2006)foundthatalargeproportionofindependentdirectorsontheboard(andintheauditcommittee)reducedthelevelofearningsmanagement,particularlyforlargefirms.

Ontheotherhand,AgrawalandKnoeber(1996)foundempiricalevidenceofanegativerelationbetweentheproportionofindependentdirectorsandfirms'performance.Daltonetal.(1998)couldnotsupportasignificantrelationshipbetweenfinancialperformanceandboardcomposition.Inthesamedirection,BhagatandBlack(2002)findnocorrelationbetweenthemeasuresoflong-termperformanceandindependentdirectorsontheboard.FerrisandYan(2007)foundnorelationshipeitherbetweenboardindependenceandperformanceinmutualfundsorbetweenboardindependenceandscandalsinthesefirmsaftertheSECchangeditscorporategovernancerules.BhagatandBolton(2009)alsofoundthatlargerboardindependencenegativelyaffectedoperatingperformancebeforetheSarbanes–OxleyAct;however,afterpassageoftheact,therelationshipwasfoundtobepositiveandsignificant.

Severaltheoreticaldevelopmentshaveattemptedtoclarifytheroleplayedbyboardindependence,developingwhatcouldbecalledanoptimalboardindependencetheory(HermalinandWeisbach,1998;Raheja,2005;AdamsandFerreira,2007;HarrisandRaviv,2008;KumarandandSivaramakrishnan,2008).Thistheoryshowsthatboardindependenceisnotalwaysinshareholders'interest,partiallyexplainingtheapparentlyconflictingempiricalevidencefoundregardingtheroleofboardindependence.Forexample,HermalinandWeisbach(2003)showthatpastperformanceisadeterminantofoptimalboardindependence.Inwell-performingfirms,itisintheshareholders'interesttoallowboardswithlessindependentdirectors.Severalempiricalpapersonboardcompositionprovideevidenceconsistentwiththeoptimalboardindependencetheory(e.g.,Booneetal.,2007;Colesetal.,2008;Lincketal.,2008).Inaddition,thepreviousliteraturethatcontrolsforthisendogeneityproblemfindsthatboardindependenceisinshareholders'interestwhenindependencealignswiththeoptimalboardindependencetheory.ThesefindingsaresupportedbyDuchinetal.(2010),whoanalyzetheeffectofregulatorymodificationsregardingboardindependence,andbyNguyenandNielsen(2010),whoanalyzetheeffectofthesuddendeathsofindependentdirectors.Adoptingadifferentmethodologybyusinglaboratoryexperiments,Gilletteetal.(2003)indicatethatboardindependenceisintheshareholders'interestevenwhennon-informedindependentdirectorsarepresent.

However,toempiricallydeterminetheimpactofboardindependence,werequireanaccuratemeasureofthefirms'strictindependencelevel.Ourapproachreliesontheformalindependencecriteriareportedpubliclybythefirms.Thesecriteriadifferfromtheinformalindependencecriteriathatconsidersocialties,i.e.,HwangandKim(2009),orthatexcludesasindependentsthedirectorsinvolvedinscandalsinpreviousyears,i.e.,Byrdetal.(2009).UnlikeSantellaetal.(2006,2007),whoanalyzewhetherdisclosureissufficienttocorroboratethedeclaredindependencebyfirms,ourformalindependenceapproachappliestoalldirectorsinanexplicitandmeasurableway.Furthermore,weanalyzetherelationshipofindependentdirectorswithallcontrollingshareholders,notjustwiththecontrollingfamilyasinAnsarietal.(2013)inthisspecialissue.Thisapproachallowsustodetectthecharacteristicsoffirmsthataremorepronetomisclassifydirectorsasindependentsthanothers,andtheconsequencesofsuchmisclassification.3.Whatshouldanindependentdirectorbe?

Theobservednumberofindependentboarddirectorscomesfromthefirm'sself-classificationwhenfilingtherequiredformsaccordingtotheircountryorstockmarketregulationsordisclosingtheircorporategovernancereport.TheSpanishUnifiedGoodGovernanceCodeofListedCompaniesspecifiesthatindependentdirectorsare“directorsappointedfortheirpersonalorprofessionalqualitieswhoareinapositiontoperformtheirdutieswithoutbeinginfluencedbyanyconnectionwiththecompany,itsshareholdersoritsmanagement”.ThespiritofthisdefinitiondoesnotdifferfromthecurrentdefinitionsintheNYSE,theEuropeanUnionortheUKregulations.Nevertheless,thedetailsinthedefinitionthatinternationalregulatorybodiesproposeasindependentdirectors'characteristicsareheterogeneous.AguileraandCuervo-Cazurra's(2009)comparisonofcorporategovernancecodesremarksthatthedefinitionsofindependentdirectorchangeacrosscountriesandevenacrossfirms.Table1showsfourproposalsregardingwhatanindependentdirectorshouldbe,andmorespecifically,whatanindependentshouldnotbe.TheseproposalsarefoundintheNYSEListedCompanyManual(amendedNovember25,2009),theEuropeanUnionCommissionrecommendationof15February2005ontheroleofnon-executivedirectors,theUKCorporateGovernanceCodeform2012,andtheSpanishUnifiedGoodGovernanceCodeform2006.IntheSpanishcase,thedefinitionhasbeenmandatorysince2007.

Thecriteriatocheckeachdirector'sindependenceinasetofSpanishlistedcompaniesfitsintotheSpanishregulationthatissimilartothemajorityofinternationallocations,asTable1indicates.Furthermore,theSpanishregulatoryenvironmentallowsusPleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

4FinancePlease(2013),citethishttp://dx.doi.org/10.1016/j.jcorparticleas:Crespí-Cladera,Table1

DefinitionsofIndependentdirectors'requirementsaccordingtoNYSEListedCompanyManual,EuropeanUnioncommissionrecommendation,UKunifiedcorporategovernancecodeandSpanishunifiedgoodgovernanceR.code.

NYSEListedCompanyManual.Section303A.02EuropeanUnionCommissionrecommendationof15TheUKCorporateGovernanceCodeSeptember2012.UnifiedGoodGovernanceCodeoflistedcompanies.(AmendedNovember25,2009)

February2005ontheroleofnon-executiveorMay19,2006.Spain

R.,supervisorydirectorsoflistedcompaniesandontheCrespí-Cladera,committeesofthe(supervisory)board

B.ﬁAnindependentdirectoris

n.2013.12.009Pascual-Fuster,Nodirectorqualifiesas“independent”unlesstheAdirectorshouldbeconsideredtobeindependentTheboardshouldidentifyintheannualreporteachDirectorsappointedfortheirpersonalor

boardofdirectorsaffirmativelydeterminesthattheonlyifheisfreeofanybusiness,familyorothernon-executivedirectoritconsiderstobeindependent.professionalqualitieswhoareinapositiontodirectorhasnomaterialrelationshipwiththelistedrelationship,withthecompany,itscontrollingTheboardshoulddeterminewhetherthedirectorisperformtheirdutieswithoutbeinginfluencedbycompany(directlyorasapartner,shareholderorshareholderorthemanagementofeither,thatindependentincharacterandjudgmentandwhetheranyconnectionwiththecompany,itsshareholdersPascual-Fuster/B.,officerofanorganizationthathasarelationshipcreatesaconflictofinterestsuchastoimpairhistherearerelationshipsorcircumstanceswhichareoritsmanagement.

Doeswiththecompany).

judgment.

likelytoaffect,orcouldappeartoaffect,thedirector'sjudgment.TheboardshouldstateitsreasonsifitJournalofthedeterminesthatadirectorisindependentnotwith-standingtheexistenceofrelationshipsorcircumstancesindependencewhichmayappearrelevanttoitsdetermination.

CorporateAnindependentdirectorisnot

(a)Thedirectoris,orhasbeenwithinthelastthree(a)nottobeanexecutiveormanagingdirectorof(a)hasbeenanemployeeofthecompanyorgroup(a)Pastemployeesorexecutivedirectorsofgroupyears,anemployeeofthelistedcompany,oranthecompanyoranassociatedcompany,andnotwithinthelastfiveyears;

companies,unless3or5yearshaveelapsed,Financeimmediatefamilymemberis,orhasbeenwithinhavingbeeninsuchapositionforthepreviousfive(b)has,orhashadwithinthelastthreeyears,arespectively,fromtheendoftherelation.

xxxofthelastthreeyears,anexecutiveofficerofthelistedyears;

materialbusinessrelationshipwiththecompany(b)Thosewhohavereceivedsomepaymentorindependentcompany.

(b)nottobeanemployeeofthecompanyoraneitherdirectly,orasapartner,shareholder,directorotherformofcompensationfromthecompanyor(b)Thedirectorhasreceived,orhasanimmediateassociatedcompany,andnothavingbeeninsuchaorsenioremployeeofabodythathassuchaitsgroupontopoftheirdirectors'fees,unlessthe(2013)familymemberwhohasreceived,duringanypositionforthepreviousthreeyears,exceptwhenrelationshipwiththecompany;

amountinvolvedisnotsignificant.Dividendsorxxxtwelve-monthperiodwithinthelastthreeyears,thenon-executiveorsupervisorydirectordoesnot(c)hasreceivedorreceivesadditional

pensionsupplementsreceivedbyadirectorfor–morethan$120,000indirectcompensationfrombelongtoseniormanagementandhasbeenelectedremunerationfromthecompanyapartfromaprioremploymentorprofessionalservicesshallnotxxxthelistedcompany,otherthandirectorandtothe(supervisory)boardinthecontextofadirector'sfee,participatesinthecompany'ssharecountforthepurposesofthissection,providedsuchdirectorscommitteefeesandpensionorotherformsofsystemofworkers'representationrecognizedbyoptionoraperformance-relatedpayscheme,orisasupplementsarenoncontingent,i.e.thepayingdeferredcompensationforpriorservice(providedlawandprovidingforadequateprotectionagainstmemberofthecompany'spensionscheme;

companyhasnodiscretionarypowertosuspend,suchcompensationisnotcontingentinanywayonabusivedismissalandotherformsofunfair(d)hasclosefamilytieswithanyofthecompany'smodifyorrevoketheirpayment,andbydoingsocontinuedservice).

treatment;

advisers,directorsorsenioremployees;

wouldbeinbreachofitsobligations.

matter?,(c)(A)Thedirectorisacurrentpartneror(c)nottoreceive,orhavereceived,significant(e)holdscross-directorshipsorhassignificantlinks(c)Partners,noworonthepast3years,inthe

employeeofafirmthatisthelistedcompany'sadditionalremunerationfromthecompanyoranwithotherdirectorsthroughinvolvementinotherexternalauditororthefirmresponsiblefortheauditinternalorexternalauditor;(B)thedirectorhasanassociatedcompanyapartfromafeereceivedascompaniesorbodies;

report,overthesaidperiod,ofthelistedcompanyJ.immediatefamilymemberwhoisacurrentpartnernon-executiveorsupervisorydirector.Such

(f)representsasignificantshareholder;orhas

oranyotherwithinitsgroup.

Corp.ofsuchafirm;(C)thedirectorhasanimmediateadditionalremunerationcoversinparticularanyservedontheboardformorethannineyearsfrom(d)Executivedirectorsorseniorofficersofanotherfamilymemberwhoisacurrentemployeeofsuchaparticipationinashareoptionoranyother

thedateoftheirfirstelection.

companywhereanexecutivedirectororseniorfirmandpersonallyworksonthelistedcompany's

performance-relatedpayscheme;itdoesnotcover

officerofthecompanyisanexternaldirector.

FinancePleaseaudit;or(D)thedirectororanimmediatefamilythereceiptoffixedamountsofcompensationunder(e)Thosehavingmaterialbusinessdealingswithmemberwaswithinthelastthreeyearsapartneraretirementplan(includingdeferredcompensa-thecompanyorsomeotherinitsgrouporwhohaveoremployeeofsuchafirmandpersonallyworkedtion)forpriorservicewiththecompany(providedhadsuchdealingsintheprecedingyear,eitherononthelistedcompany'sauditwithinthattime.thatsuchcompensationisnotcontingentinanytheirownaccountorasthesignificantshareholder,(2013),cite(d)Thedirectororanimmediatefamilymemberis,wayoncontinuedservice);

directororseniorofficerofacompanythathasorthisorhasbeenwiththelastthreeyears,employedas(d)nottobeortorepresentinanywaythehashadsuchdealings.Businessdealingswill

anexecutiveofficerofanothercompanywhereanycontrollingshareholder(s)(controlbeing

includetheprovisionofgoodsorservices,includinghttp://dx.doi.org/10.1016/j.jcorparticleofthelistedcompany'spresentexecutiveofficersatdeterminedbyreferencetothecasesmentionedinfinancialservices,aswellasadvisoryorconsultancythesametimeservesorservedonthatcompany'sArticle1(1)ofCouncilDirective83/349/EEC(1));relationships.

as:compensationcommittee.

(e)nottohave,orhavehadwithinthelastyear,a(f)Significantshareholders,executivedirectorsor(e)Thedirectorisacurrentemployee,oransignificantbusinessrelationshipwiththecompanyseniorofficersofanentitythatreceivessignificantCrespí-Cladera,immediatefamilymemberisacurrentexecutiveoranassociatedcompany,eitherdirectlyorasadonationsfromthecompanyoritsgroup,orhasofficer,ofacompanythathasmadepaymentsto,orpartner,shareholder,directororsenioremployeeofdonesointhepast3years.Thisprovisionwillnotreceivedpaymentsfrom,thelistedcompanyforabodyhavingsucharelationship.Business

applytothosewhoaremerelytrusteesofaR.propertyorservicesinanamountwhich,inanyofrelationshipsincludethesituationofasignificantFoundationreceivingdonations.

thelastthreefiscalyears,exceedsthegreaterof$1supplierofgoodsorservices(includingfinancial,(g)Spouses,partnersmaintainingananalogousmillion,or2%ofsuchothercompany'sconsolidatedlegal,advisoryorconsultingservices),ofasignifi-affectiverelationshiporcloserelativesofoneoftheR.,grossrevenues.

cantcustomer,andoforganizationsthatreceivecompany'sexecutivedirectorsorseniorofficers.GeneralCommentarytoSection303A.02(b):Ansignificantcontributionsfromthecompanyorits(h)AnypersonnotproposedforappointmentorCrespí-Cladera,Pascual-Fuster,“immediatefamilymember”includesaperson'sgroup;

renewalbytheNominationCommittee.

B.ﬁspouse,parents,children,siblings,mothersand(f)nottobe,orhavebeenwithinthelastthree(i)Thosestandinginsomeofthesituationslistedinn.2013.12.009fathers-in-law,sonsanddaughters-in-law,brothersyears,partneroremployeeofthepresentorformera),e),f)org)aboveinrelationtoasignificantandsisters-in-law,andanyone(otherthandomesticexternalauditorofthecompanyoranassociatedshareholderorashareholderwithboard

employees)whosharessuchperson'shome.Whencompany;

representation.Inthecaseofthefamilyrelationssetapplyingthelook-backprovisionsinSection(g)nottobeexecutiveormanagingdirectorinoutinletterg),thelimitationshallapplynotonlyinPascual-Fuster/B.,303A.02(b),listedcompaniesneednotconsideranothercompanyinwhichanexecutiveormanag-connectionwiththeshareholderbutalsowithhisorDoesindividualswhoarenolongerimmediatefamilyingdirectorofthecompanyisnon-executiveorherproprietarydirectorsintheinvesteecompany.membersasaresultoflegalseparationorsupervisorydirector,andnottohaveothersignif-Proprietarydirectorsdisqualifiedassuchand

Journaldivorce,orthosewhohavediedorbecomeicantlinkswithexecutivedirectorsofthecompanyobligedtoresignduetothedisposalofsharesbytheoftheincapacitated.

throughinvolvementinothercompaniesorbodies;shareholdertheyrepresentmayonlybere-electedindependenceInaddition,referencestothe“listedcompany”or(h)nottohaveservedonthe(supervisory)boardasasindependentsoncethesaidshareholderhassold“company”includeanyparentorsubsidiaryinaanon-executiveorsupervisorydirectorformoreallremainingsharesinthecompany.AdirectorCorporateconsolidatedgroupwiththelistedcompanyorthanthreeterms(or,alternatively,morethanwithsharesinthecompanymayqualifyassuchothercompanyasisrelevanttoany12yearswherenationallawprovidesfornormalindependent,providedheorshemeetsallthedeterminationundertheindependentstandardstermsofaverysmalllength);

conditionsstatedinthisRecommendationandFinancesetforthinthisSection303A.02(b).

(i)nottobeaclosefamilymemberofanexecutivetheholdinginquestionisnotsignificant.

xxxoformanagingdirector,orofpersonsinthesituationsindependentreferredtoinpoints(a)to(h);

(2013)xxx–xxxdirectorsmatter?,J.Corp.56R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

tomeasurestrictindependencedirectorbydirectorusingpublicdataprovidedbythelistedfirms,comparedwiththeirdeclaredindependence.

UndertheSpanishregulation,since2004listedcompanieshavebeenrequiredtoreleaseastandardizedannualreportoncorporategovernancepractices(ARCG)totheSpanishsecuritiesandexchangecommission(CNMV).ThereportisfiledelectronicallyandispubliclyavailableattheCNMVwebpage.Thereportincludesinformationaboutboarddirectors'namesandtheirtypologyaccordingtothecompany'sjudgmentandspecificallydefinesexecutivedirectors,proprietarydirectors,independentdirectorsandothers.TheARCGalsoincludesinformationontheboardmembers'relationshipwithsignificantshareholders,thefirm,itsassociatesandsubsidiaries.

TheARCGallowsustoempiricallytesteightindependencecriteriaforalldirectorsdeclaredasindependent.Thesecriteria,showninTable2,canbeinterpretedasinternationalstandarddefinitionsspecificallyfocusedontheSpanishregulationasstatedinTable1.Theselectionofthesecriteria,basedontheSpanishmandatorydefinitionofindependentdirectorsandtheinternationalstandards,isadaptedtotheavailabilityofpublicinformationtocheckthemajorityoftheaspectsofthisdefinition.3Criterion1addressesthe2007Spanishcorporategovernancecodethatstatesthattheappointmentbythenominationcommitteeiscompulsoryforthedirectortobeclassifiedasindependent.Nevertheless,internationalbestpracticescodesincludetherecommendationofgivingthiscommitteetaskssuchastheappointmentoftheindependentdirectors.ShivdasaniandYermack(1999)foundempiricalevidencethatCEOs'involvementintheappointmentsofnewdirectorswhennonominationcommitteeexistedyieldedmore‘gray’,non-strictlyindependentdirectorswithconflictsofinterest.Limitedtenureofuptotwelveyearsisthesecondcriterion,whichisalsoincludedintheUKcombinedcodeandtheEUrecommendations.Thesixremainingcriteriaarestandardintheinternationalregulations,asTable2confirms.Thethirdcriterionrestrictsindependencetothosedirectorswhodonothaveasignificantbusinessrelationshipwiththecompany.TherelationshipwiththecontrollingshareholdersisanelementthattheSpanishcode,theNYSErulesandothercodesdefineasessentialfortheindependenceofdirectors.Theconditionofbeingadirector,amanageroremployeeofasignificantshareholder(fourthcriterion),havingothertypesofrelevantrelationship(i.e.,family)withasignificantshareholder(fifthcriterion),orreceivingcompensationbythecompany,itssubsidiariesoritsassociates,forotherfunctionsapartfromthedirectorship(sixthcriterion)mustexcludethedirectorfrombeingqualifiedasarealindependentmember.Companiescanalsobeformallyboardmembersthrougharepresentative,althoughtheseventhcriterionobviouslydoesnotacceptthistypeofdirectorasanindependent.Ourlastcriterion,theeighth,avoidsclassifyingasindependentanyrecentcompanyexecutive.Table2summarizesthesecriteria,showingthefitwiththecorrespondingruleinthecodesrevisedinTable1.

4.Dataanddescriptivestatistics

Thedatasetconsistsof752firms/yearobservationsrangingfrom2004to2009.Thissix-yeardatasetincludesalllistedfirmsintheSpanishstockmarket,rangingfrom118to135firmsdependingontheyear.Informationregardingboardcomposition,corporategovernancepracticesandindividualinformationonboardmembers,suchastenureortheirrelationshipwithsignificantshareholders,isobtainedfromthestandardizedARCGthatfirmsmustfulfill.StockmarketdataarefromtheThomsonFinancialdatabaseanddataonfirms'annualfinancialreportsarefromtheBureauVanDijkdatabaseonfinancialreportsforSpanishfirms.

PanelAofTable3describesthemostrelevantstatisticsofboardindependenceoftheSpanishlistedcompanies.Althoughthefirmsdeclaredanaverageof32.5%ofindependentdirectorsontheboard,thisfiguredecreasesto14.2%whenconsideringstrictlyindependentdirectorsoverboardsize;thatis,56.3%ofthedeclaredindependentdirectorsdonotcomplywithatleastoneoftheeightindependencecriteriadescribedinTable2.Theproportionofnon-strictlyindependentsisnearlyuniformacrossfirmsizeandindustriesanddecreasesalongtime(panelC),particularlyin2007whentheSpanishmandatorydefinitionofindependentdirectorsbecamecompulsory.

PanelBofTable3analyzesthepatternofnewindependentdirectorsappointedeachyear.Theproportionofnon-strictlyindependentnewdirectorssupportsthenotionthatthefindingsofpanelAarenotmerelydrivenbythestockofappointeddirectors,whenindependentdirectors'appointmentswerelessrigorous.Theshiftinthelevelin2007isconsistentwiththementionednewmandatorydefinitionofanindependentdirectorundertheSpanishregulation.TheSpanishbestpracticescodealsorecommendsthattheauditcommitteeandthenominationandremunerationcommitteeshouldbechairedbyanindependentdirectorandformedbyamajorityofindependents,withnoexecutivesinplace.PanelDofTable3showsthattheaveragepercentageofdeclaredindependentdirectorsinbothcommissionsissubstantiallyhigherthanthepercentageofindependentsontheboard(one-halfversusone-third,respectively).However,afterfilteringthenon-strictlyindependentdirectors,thesepercentagesdecreasesignificantly.Thedeclaredaverageof49.4%ofindependentmembersintheauditcommitteeandtheaverageof50.4%inthenominationandremunerationcommitteedecreaseto21.2%and24.1%ofstrictlyindependents,respectively.Themajorityoffirmsappointedadeclaredindependentdirectorforthechairofbothcommittees,althoughonaverageinlessthanone-thirdoffirmswerechairedbystrictlyindependentdirectors.PanelEindicatesthattherateofmisclassifiedindependentsinthesecommitteesisdecreasingovertime.Nevertheless,fortheyear2009theproportionofstrictlyindependentsisalwaysbelow50%whenone-thirdofthefirmsstillhavenon-strictlyindependentdirectorschairingthese

Thepotentialbiasofourempiricaltestofindependenceistowardtheoverestimationoftrueindependence,nevertowardanexcessofstrictlynonindependents.

3Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx7

Table2

EmpiricaldefinitionofindependentdirectorsofSpanishlistedfirms.IncludesthecorrespondingruleunderdifferentinternationallegalsettingsfromTable1.

NYSE

[1]ProposedforappointmentorrenewalbytheNominationCommittee[2]Tenureasindependentdirectorforuptotwelveyears

[3]Nothavingasignificantbusinessrelationshipwiththecompany

[4]Notholdingadirectorship,tobeamanageroranemployeeofsignificantshareholderorashareholderwithboardrepresentation

[5]Nothavingotherrelevantrelationship(differentthanthoseinpoint4)withsignificantshareholderorashareholderwithboardrepresentation

[6]Notbeingadirectororexecutiveinsubsidiariesorassociatedcompanies[7]Nottobeacompanyasboarddirector

[8]NotbeingexecutivedirectorofthefirminthepreviousyearbaaEU(h)(e)(g)(g)(a)(b)(a)

UK(f)(b)(f)(d)(e)(b)(a)

Spain(h)

(b)(d)(d)(c)(a)(a)

(e)(i)(i)(d)(c)(a)

In2007theCNMV(theSpanishSecuritiesandExchangeCommission)modifiedtheinformationrequirementsregardingdirectorproposals.Firmsmustcommunicatewhoproposedeverydirector,exceptforindependentdirectors.Therefore,after2006weassumethatallindependentdirectorshavebeenproposedbythenominationcommittee,exceptwhenthiscommitteedoesnotexist,orifthedirectorhasnotbeenformallyrenewedandwasnotpromotedbythiscommitteebefore2007.bAswehaveinformationstartingfrom2004,thiscriteriondoesnotoperateforthisyear.

committees.One-sixthofthefirmshavenon-strictlyindependentschairingbothcommittees.Thesefindingsremainrelevantinfirmsofallsizesandindustryactivities.

Table3capturesthevaluesofthedependentvariablesintheempiricalmodelofmisclassification.Inastepfurther,thesemeasuresofnon-strictlyindependentdirectorsbecomeapartofthesetofexplanatoryvariablesoftheoperatingperformancemodelandalsoofthemodelsofexecutivedirectors'pay,CEOturnover,auditqualificationandearningsmanagement.

Table4describesthebasicstatisticsofthevariablesincludedintheposterioranalysis,asperformancemeasures,corporategovernancepracticesorcontrolvariables.Thistablecontainsashortdefinitionofthesevariables.Thejustificationoftheirinclusionintheanalysisandthefulldefinitionwillbefoundinthesectionsanalyzingtheempiricalmodels.

Thedescriptivestatisticsshowthattheaverageownershipconcentrationishigh.Non-reportedPearsoncorrelationcoefficientsshowhighvaluesamongvariablessuchassalesandmarketcapitalization,bothproxiesoffirmsize,andarealsopositivelycorrelatedwithboardsize.

5.Thegovernanceofﬁrmswithnon-strictlyindependents

Firmscontrolledbymanagersmayappointnon-strictlyindependentdirectorstoachievetheregulator'srecommendedlevelofboardindependenceandtoreachthedesiredlevelofboardrealindependence(afriendlyboard).Infirmscontrolledbylargeblockholders,independentsaresupposedtoprotecttheinterestofminorityshareholders,andlargeshareholdersmayalsousenon-strictlyindependentstoachieveadesiredlowlevelofrealboardindependenceatthesametimeastherecommendedlevelofdeclaredboardindependence.Therefore,weanalyzewhetherownershipstructureandmanagerialcontrolproxiesarerelatedtothepresenceofnon-strictlyindependentdirectors.Weestimatethefollowinglogitmodel:

PNSIDi;t¼αþα1ÁLogSalesi;tþα2ÁC1i;tþα3CEOisboardchairi;tþα4Boardsizei;tþþα5ÁVotingCapi;tþα6Á%Busynon‐executivedirectorsi;tþþα7Á%Interlockedexecutivedirectorsi;tþXi;tβþεi;t

ð1Þ

wherePNSIDi,ttakesavalueof1iffirmihasatleastonenon-strictlyindependentdirectorinyeart.Weusetheownershipofthelargestshareholder(C1)asaproxyofownershipconcentration(itishighlycorrelatedwiththeaccumulatedownershipofthethreeandfivelargestshareholders,0.92and0.84,respectively).Asproxiesofmanagerialpower,weusethefollowing:adummyvariabletoidentifyfirmsinwhichtheCEOalsochairstheboardofdirectors;boardsize(todetectboardcoordinationproblemsinitsmonitoringfunctions,Yermack,1996);adummyvariabletodetectfirmswithvotingcaps(amaximumnumberofvotesthatashareholdermayfacilitate);thepercentageofbusynon-executivedirectors(thosewiththreeormoredirectorships,withnotimetoproperlymonitorexecutives,FichandShivdasani,2006);andthepercentageofinterlockedexecutivedirectors(servinginasecondfirmasnon-executiveswithanon-executivedirectorofthefirstfirmservingasexecutive,Hallock,1997).Finally,weaddsize(thelogofsales),andyearandindustrialsectorfixedeffects(inXi,t)todetectanysystematicdifferencesinthepresenceofnon-strictlyindependentsrelatedtosuchfactors.

Table5showstheresultsofthefollowingthreelogitspecifications:pooledlogitmodelswithrobuststandarderrorsclusteredbyfirm(Huber–White),paneldatarandomeffectslogitmodels,andgeneralizedestimatingequations(GEE)logitmodels(LiangandZeger,1986;ZegerandLiang,1986)toaccountforanypersistenceinthedecisiontohavemisclassifieddirectorswithinthesamefirm.4Thismethodisequivalenttofeasiblegeneralizedleastsquaresforlinealmodels.Thepaneldatamodelsshow50replicationbootstrapstandarderrorstoobtainrobuststatisticalinference(MooneyandDuval,1993).5Models1,4,and7include

45SeeBallinger(2004)foradescriptionofthismethodinorganizationalresearch.

Huber–Whitestandarderrorsarenotfeasibleforlogitpaneldatamodelswithrandomeffects.

Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

8R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Table3

Thestrictindependenceofdirectors.PanelA:“%declared”istheaveragepercentageofindependentdirectorsoverboardsizedeclaredbyfirmsintheirAnnualReportofCorporateGovernance(ARCG).Theremainingpercentagesrefertotheaverageofstrictlyindependentsovertheboardsizeoncepasstheeightjointindependencecriteria[1..8]andindividually[i].Initalics,thereisthepercentageoffirmsthatdonotcomplywiththeindependencecriteria.PanelB:Numberofnewappointedindependentdirectorsandtheproportionthatmeetthecriteria.PanelC:LongitudinalperspectiveofPanelAandPanelBdatacontents.PanelD:Theaveragepercentageofindependentdirectorsovercommittees'sizeasdeclaredandtheaveragepercentageofstrictlyindependentsintheaudit,andinthenominationandremunerationcommitteeovercommittees'size.%firmschairisthepercentageoffirmswithadeclaredindependentdirectorchairingthecommitteeandthepercentagesoffirmswherethereisastrictlyindependentsaftertheeightcriteria.PanelE:LongitudinalperspectiveofPanelDdatacontents.DataforapoolofalllistedfirmsinSpainfrom2004to2009:118firmsin2004,119firmsin2005,126in2006,135for2007,130during2008,and124in2009.Independencecriteria%declared

[1…8]

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

PanelA:%ofexistingdeclaredandstrictlyindependentdirector's32.5%

14.2%69.8%

20.3%43.0%

28.0%28.7%

30.6%10.5%

32.1%3.7%

32.1%2.8%

28.0%27.3%

31.6%7.0%

32.5%0.5%

PanelB:#Newappointeddeclaredand%strictlyindependentdirector's419

64.70%

83.80%

95.50%

95.90%

98.10%

99.30%

88.10%

95.50%

99.00%

PanelC:Declaredandstrictlyindependentdirector'salongtime

Existingdirectors%declared

200420052006200720082009

33.2%33.5%32.2%30.8%32.4%33.3%

[1…8]8.3%10.5%10.7%16.1%18.4%20.5%

Newappointed#New56851178279

[1…8]58.9%29.4%74.4%73.2%83.5%

PanelD:Declaredandstrictlyindependentdirector’sincommittees

Auditcommittee%IndependentsDeclared

After[1…8]

%firmsChairIndependentDeclared

After[1…8]

Nominationandremunerationcommittee%IndependentsDeclared50.4%48.7%49.7%50.3%48.3%52.6%52.6%

After[1…8]24.1%15.1%18.4%18.7%26.0%30.4%31.4%

%firmsChairIndependentDeclared64.0%52.8%56.0%56.9%61.9%72.7%78.0%

After[1…8]32.9%18.0%20.9%24.5%36.5%46.3%43.2%

49.4%21.2%71.8%30.9%

PanelE:declaredandstrictlyindependentsincommitteesalongtime200420052006200720082009

47.1%48.3%48.6%49.7%51.7%50.6%

9.8%13.5%17.0%26.3%28.7%30.5%

64.4%63.9%68.3%75.6%76.2%81.5%

10.2%15.1%20.6%40.7%45.4%50.0%

oneproxyofshareholderpower(C1)andmanagerialpower(CEOchairstheboard).Models2,5,and8includetheremainingcorporategovernancevariables.Becauseboardsizemightproxyfirmsize,wealsoestimatemodels3,6,and9withoutthisvariableasarobustnesstest.

ThemodelsinTable5clearlyshowthatthestrongerthecontrolbythelargestshareholderis,thelowertheproportionoffirmsmisclassifyingindependentsis.6Wealsoobtainweakempiricalevidence(forthepooledlogitmodels)ofapositiveinteractionamongfirmswithnon-strictlyindependentsandvotingcapslimitingthepowerofpotentialnewentrantshareholders.Boardsizeisalsopositivelyrelatedtonon-strictlyindependents,exceptintheGEEpaneldatalogitspecification.

Insum,theempiricalevidenceoftheselogitmodelssuggeststhatlargeshareholdersdonotusenon-strictlyindependentdirectorstodecreaseindependents'influence.Firmscontrolledbymanagers(withdispersedownershipstructures)tendtoreducerealboardindependencewithnon-strictlyindependents.6.Theimpactofnon-strictlyindependentsintheboardroom

Basedontheagencytheory,theapproachregardingtherelevanceofindependentdirectorsintheboardroomisthatthemonitoringofmanagerialpowerisakeyelementinpreventingshareholders'wealthexpropriation.Apotentialmanager's

Evenbasicstatisticsareconsistentwiththisresult:497observations(ﬁrm/year)belongtoﬁrmsinwhichthelargestshareholderownslessthanhalfofshares;in77.4%oftheseobservationstherearenon-strictlyindependentdirectors.Thelargestshareholderownsmorethanhalfofthesharesin255observationsandin54.9%ofthemtherearemisclassiﬁedindependentdirectors.

6Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx9

Table4

Descriptivestatisticsoffirms'characteristics:mean,medianandstandarddeviations.NCRreferstothenominationandremunerationcommittee,ACreferstotheauditcommittee.Percentagesofnon-strictlyindependentsarecomputedoverboardsizeandoverthecommitteesize,respectively.AverageExecutivecompensationisthetotalcompensation,ofanykind,ofexecutivedirectorsdividedbythenumberofexecutivedirectorsontheboard.CEOturnoverreferstothepercentageofcaseswherethereisachangeofaCEObetweentwoconsecutiveyears.Auditqualificationcapturesthe“notclean”opinionoftheexternalauditor(nodataforbanks).Lossisadummyvariableforthefirmsthatreportlossesintheearningsstatement.Leverageistheleveloflongtermdebtovertotalassets(nodataforbanks).Liquidityiscurrentassetsovercurrentliabilities(nodataforbanks).OperatingCycleisthesumofdaysreceivableanddaysinventory(nodataforbanks,neitherDayspayable).CapitalIntensityisfixedassetsovertotalassets(nodataforbanks).Executives'ownershipandNon-executives'ownershipisthesumoftheownershipstakesofexecutiveandnonexecutivedirectorsrespectively.C1isthepercentageofsharesownedbythelargestshareholder.CEOisboardChairisadummyvariabletoidentifyfirmswheretheCEOchairstheboardofdirectors.Boardsizeisthenumberofdirectors.Tenureistheaveragenumberofyearsthatboardexecutivesremaininboardsincetheirappointment.Operationsdirectors-firmisadummythatcapturesifthefirmreportshavingtransactionsorcommercialoperationswithitsdirectors.%busynon-executivedirectors(adirectorisbusyifsheholdsapositioninthreeormoreboardsofdirectors)iscomputedoverboardsize,asitisthepercentageofinterlockedexecutivedirectors(thosewhoarealsonon-executivedirectorsinafirmwhereanon-executivedirectorofthefirstfirmisanexecutivedirector).Goldenparachuteisadummyvariablethatcapturesfirmsprotectingexecutivesagainstdismissal.Votingcapisadummyvariableidentifyingfirmswithamaximumnumberofvotesthatashareholdermayexerciseindependentlyofthenumberofsharesshehas.

%Observ.

%non-strictlyindependents

%non-strictlyindependentsNRC%non-strictlyindependentsACNRCnon-strictlyindependentchairACnon-strictlyindependentchair

AverageExecutiveCompensation(1000€)CEOturnovereventsAuditQualificationSales(Mill€,annual)

MarketCapitalization(Mill€,yearend)ROA(ReturnonAssets,annual)StockReturn(annual)

MarkettoBook(equity,yearend)Loss

LeverageLiquidity

OperatingCycle(days)DaysPayable(days)CapitalIntensity

Executivesownership

Non-executivesownership

C1(OwnershipLargestShareholder)CEOisboardchairBoardsize

%Executivesontheboard

Tenureofboardexecutives(years)Operationsdirectors-firm

%Busynon-executivedirectors%InterlockedexecutivedirectorsGoldenparachutesVotingCap

Mean18.3126.3428.13

31.0740.96

969.12

14.0411.72

3,509.334,655.18

4.6214.113.26

13.16

0.273.63329.37213.990.5511.2112.5536.39

57.98

11.0420.168.56

37.77

9.6412.02

57.5812.77

0.000.00

14.0126.20

10.0019.387.02

3.8612.386.88

0.251.23118.08149.600.580.103.8829.54

0.1835.22573.86211.260.2322.2318.9826.05

618.80833.714.6811.182.13

9,262.8912,074.34

9.1359.134.48

554.00

1252.47

Median14.2925.0025.00

StdDev

17.8927.2029.53

expropriationbehaviorismoreinfluentialunderdispersedownershipstructures,whereshareholders'coordinationisdifficultandcostly,thanunderconcentratedownershipstructures.Forfirmswithoutsignificantlargeowners,ourfindingsshowthatincentivestoappointnon-strictlyindependentsarestronger.However,thisscenariodoesnotimplythatthemanagerialpoweristheultimatereasonforthewidespreadpresenceofnon-strictlyindependentdirectorsinoursample.Theoptimalboardindependencetheorymaybethisultimatereason.Forexample,ourresultsmaybeexplainedbyHermalinandWeisbach(2003),whofoundthatitisoptimalinshareholders'interesttoreduceboardindependenceforCEOswithpositiverecordsofpastperformance.Whenevertheoptimalboardindependencelevelofafirmisbelowtherecommendedlevelofboardindependence,firmshaveincentivestousenon-strictlyindependentdirectorstoachievetheoptimallevelofrealboardindependenceatthesametimeastherecommendedlevelofdeclaredindependence.Indeed,Ciceroetal.(2013)foundempiricalevidenceoffirmsdeclaringboardindependenceaboveitsoptimumlevel.

Therefore,wenextanalyzetheconsequencesofindependents'misclassificationintermsoffuturefirms'performanceandintermsofseveralrelevantoutputsofthemainboardcommitteeswithmonitoringroles,whichincludetheauditcommitteeandthenominationandremunerationcommittee.Negativeconsequenceswillbeinterpretedassupportforthemanagerialpowerapproach,whereaspositiveconsequencesoralackofconsequenceswillsupporttheoptimalboardindependencetheory.First,forthenominationandremunerationcommittee,wetestwhetherexecutivedirectors'compensationpracticesandCEOturnoveraredependentonthelevelofnon-strictlyindependentdirectors.Second,fortheauditcommittee,wetestwhetherthenumberofnon-strictlyindependentdirectorsaffectsfirms'auditqualificationandearningsmanagementbehavior.

NguyenandNielsen(2010)showthatthecontributiontothefirmvaluebyindependentdirectorsisgreaterwhenthedirectorsplayrelevantrolesontheboardcommittees,particularlyontheauditcommittee.UndertheSpanishSecuritiesandPleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

10R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Table5

Corporategovernanceandnon-strictlyindependents.Thedependentvariabletakesvalue1whenafirmhasadirectorthatdoesnotmeettheeightcriteriaforindependencedescribedinTable2.PanelAshowsPooledlogitmodelestimationswithHuber(1967)andWhite(1980,1982)robusttstatistics(inparenthesis),clusteringbyfirm(Petersen,2009).PanelBshowsPaneldatalogitestimationswithrandomeffectswheretstatisticsarecomputedbybootstraptechniqueswith50replications(MooneyandDuval,1993).PanelCpresentstheGEEpaneldatalogitmodelsallowingcorrelationamongtheerrortermsofthesamefirm,with50replicationsbootstrapstandarderrors.Theexplanatoryvariablesarethelogarithmofsales,thepercentageofsharesownedbythelargestshareholder,adummyvariabledetectingiftheCEOalsochairstheboardofdirectors,theboardsize,adummyvariableidentifyingfirmswithvotingcaps(amaximumnumberofvotesthatashareholdermayexerciseindependentlyofthenumberofsharesshehas),thepercentageofbusynon-executivedirectors(adirectorisbusyifsheholdsapositioninthreeormoreboardsofdirectors),thepercentageofinterlockedexecutivedirectors(thosewhoarealsonon-executivedirectorinafirmwhereanon-executivedirectorofthefirstfirmisanexecutivedirector),dummyvariablesforindustries(6)andyears.SampleofallfirmslistedinthemaintradingmechanismoftheSpanishStockExchange(SIBE).Chi2isaWaldtestofthestatisticalsignificanceofalltheexplanatoryvariables.Chi2-IIisaWaldtestofthejointstatisticalsignificanceofBoardsize,VotingCap,%Busynon-executivedirectors,and%Interlockedexecutivedirectors.Chi2-IIIisaWaldtestofthejointstatisticalsignificanceofVotingCap,%Busynon-executivedirectors,and%Interlockedexecutivedirectors.Likelihoodratiotestrho=0isatestonthestatisticalsignificanceofthevarianceoftheunit-specificresidual.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

PanelA:Pooledlogitmodels

PanelB:Paneldatalogitwithrandomeffects(4)0.161(0.778)−0.037***(−2.768)0.565(0.784)

(5)−0.111(−0.443)−0.030**(−2.236)0.563(0.997)0.263***(2.883)1.078(1.092)0.638(0.399)−0.648(−0.554)YesYes752

40.21***14.37***172.4***

(6)0.136(0.809)−0.036***(−3.007)0.571(1.098)

PanelC:GEEPaneldatalogitmodels

(1)

LogSalesC1

CEOisboardchairBoardsizeVotingCap

%Busynon-executivedirectors%InterlockedexecutivedirectorsYearfixedeffectsIndustryfixedeffectsObservationschi2chi2-IIchi2-III

Likelihoodratiotestrho=0

YesYes752

41.51***0.106(1.229)−0.022***(−3.830)0.341(1.127)

(2)−0.089(−0.824)−0.017***(−2.809)0.430(1.364)0.148***(2.877)1.134*(1.830)1.316(1.489)−0.001(−0.002)YesYes752

52.10***15.52***

(3)0.079(0.924)−0.020***(−3.368)0.312(1.044)

(7)−0.081(−1.448)−0.009**(−2.282)0.078(0.663)

(8)−0.092(−1.620)−0.008**(−2.079)0.079(0.645)0.010(0.480)0.262(1.117)0.120(0.327)−0.116(−0.427)YesYes752

95.72***1.68

(9)−0.083(−1.358)−0.008**(−2.109)0.081(0.691)

1.186**(1.993)1.769*(1.940)0.064(0.103)YesYes752

50.66***7.27*

YesYes752

36.53***

1.251(0.344)1.043(0.536)−0.376(−0.330)YesYes752

57.99***0.50

YesYes752

111.42***

0.264(1.107)0.138(0.372)−0.108(−0.302)YesYes752

112.91***1.3

197.23***

MarketLaw,alllistedfirmsmusthaveanauditcommittee,whereastheSpanishcorporategovernancecodemerelyrecommendsthepresenceofanominationandremunerationcommittee(86%ofobservations,firm/year).6.1.Futureoperatingperformance

Ifnon-strictlyindependentdirectorsareappointedtomaskanagencyproblemamongmanagersandowners(orlargeandminorityowners),weshouldexpectanegativeimpactonfirmfutureperformance.Asoperatingperformance,weevaluatereturnonassets,calculatedasthenetincomeplusinterestpayments,netoftaxeffects,overtheamountofthepreviousyear'stotalassets.Thefollowingfivemeasurescapturethepowerofnon-strictlyindependents:theproportionofnon-strictlyindependentdirectorsoverboardsize;theproportionoverthecommitteesizeofnon-strictlyindependentdirectorsinthenominationandremunerationcommitteeandintheauditcommittee;andtwodummyvariablestoidentifyfirmsinwhichthechairofthesecommitteesisanon-strictlyindependentdirector.Giventherelevanceoftheownershipstructurefoundintheprevioussection,thesemeasuresarealsointeractedwiththeownershipofthelargestshareholdertoinvestigateanydifferentimpactofnon-strictlyindependents.Thisinteractionisalsoconductedinourposterioranalyses.ThestructureofthemodelandthecontrolvariablesfollowHwangandKim(2009):

󰀁󰀃

ð2ÞPerformancei;tþ1¼αþα1ÁNSIDi;tþα2ÁC1i;t×NSIDi;tþXi;tβþεi;twherePerformancei,t+1istheROAoffirmiinyeart+1,NSIDi,tisthenon-strictlyindependentdirectorsmeasureoffirmiin

yeart,andXi,tisasetofcontrolvariablesincludingproxiesoffirmsize,growthopportunities,andcorporategovernancevariables.Themeasureoffirmsizeisthelogofsales,andgrowthopportunitiesaretakenasoneperiodlaggedmarkettobookratio(ofequity).Tocapturethepotentialimpactofanuncontrolledagencyproblemweusetheownershipstructureandagroupofboardofdirectorsandcorporategovernancecharacteristics.Ownershipstructureproxiesaretheownershipstakeofthelargestshareholder,theownershipofexecutivedirectorsandthesharesownedbynon-executivedirectors.Theboardofdirectors'characteristicsare;adummyvariabletoidentifyfirmsinwhichtheCEOchairstheboard,thepercentageofexecutivedirectors,Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx11

thepresenceof“goldenparachutes”toprotectexecutivedirectorsagainstdismissals(Bricketal.,2006),theboardsizetodetectboardcoordinationproblems(Yermack,1996),thepercentageofbusydirectorswithdedicationproblems(FichandShivdasani,2006),andthepercentageofinterlockedexecutivedirectorstodetectthepotentiallylowermonitoringeffortofnon-executivedirectors(Hallock,1997).Withadummyvariable,wealsocontrolwhetherdirectorsconductrelevanteconomictransactionswiththefirm(asdeclaredbyfirmsintheirARCG)asapotentialrentextractionmechanism;thisconditionisparticularlyrelevantinoursamplegiventhepredominanceofowner-controlledfirms(Faccioetal.,2001).Weestimatethesemodelswithtimeandfirmfixedeffectstocontrolforanysystematictimeeffectandcontrolfornon-observableconstantfirmcharacteristicsaffectingfuturefirmperformance.Weincludeoneperiodlaggedreturnonassetstoaddressanyendogeneityconcern;however,theresultsdonotchangewithoutthisvariable(availableonrequest).InferenceisbasedonHuber–Whiterobusttstatistics.

Theinitial752firm/yearobservationsdecreaseto667becauseofthelagsinthemarkettobookratio;missingvaluesaregeneratedbymergersandacquisitionsintheperiodandthenewlistingfirmsthatdonotreportlaggedvalues.7Wheninformationonthenominationandremunerationcommitteeisused,thesampledecreasesto580firm/yearobservationsbecause14%havenosuchcommittee.

Table6showsthatthepresenceofmisclassifiedindependentdirectorsisnotrelatedtofutureoperatingperformanceexceptwhenanon-strictlyindependentdirectorchairstheboardcommittees,althoughwithlowstatisticalsignificance.Inthiscase,futureoperatingperformancetendstobelower(models4and5)thanforfirmswithstrictlyindependentschairingthesecommittees.Wheninteractingthenon-strictlyindependentsvariableswiththeownershipstakeofthelargestshareholderwedonotdetectrelevantdifferences,exceptthehigherstatisticalsignificanceofthenegativeeffectofanon-strictlyindependentchairingtheauditcommittee.8TheresultsremainrobustusingdifferentspecificationsasthemodeloffutureoperatingperformanceusedinCoreetal.(1999),whichincludesthevarianceofpastoperatingperformanceandoneperiodlaggedsalesascontrolvariablesinsteadofthelaggedmarkettobookratioandofcontemporaneoussales,resultsareomittedforbrevitybutareavailableonrequest.

Criticismoftheperformancemodelsincorporatefinanceoccursprimarilyregardingthepotentialendogeneityofsomeoftheexplanatoryvariablesorregardingtheomissionofrelevantvariables.Althoughourpaneldataestimationstendtouncovertheseissues,additionalmeasuresoftheimpactofnon-strictlyindependentsontheoutcomesofboardcommitteesareanalyzed.9Theexpectedoutcomesofaproperlyperformingnominationandremunerationcommitteeinclude,amongothers,ascenarioinwhichthepresenceofnon-strictlyindependentsdoesnotdistortexecutivecompensationorCEOturnover.Similarly,fortheauditcommittee,giventhelackofagencyproblemsbecauseofnon-strictlyindependentdirectors,wedonotexpectthatthepresenceofnon-strictlyindependentsaffecttheexternalauditorqualificationsortheearningsmanagementbehaviorofthefirm.6.2.Outcomesofthenominationandremunerationcommittee

6.2.1.Executivedirectorcompensation

Inthissub-sectionweanalyzetherelationbetweennon-strictlyindependentdirectorsandtheaverageindividualcompensationofexecutivedirectors,givingspecialattentiontonon-strictlyindependentsinthenominationandremunerationcommittee.10Theconjectureinthiscasewouldbethatamidagencyproblemsrelatedtonon-strictlyindependents'controlinthenominationandremunerationcommittee,executivecompensationwouldbelargercomparedtofirmswithoutnon-strictlyindependentsintheboardroom(seeGoergenandRenneboog,2011,fortherelationbetweenexecutives'compensationandweakcorporategovernance).Theproposedmodelofcompensationincludesfixedpay,bonuses,cashfromexertedstockoptions,retirementbenefits,andanyadditionalremunerationfromthefirm.Weestimatethefollowingregression:

󰀁󰀃

Compensationi;t¼αþα1ÁNSIDi;tþα2ÁC1i;t×NSIDi;tþXi;tβþεi;t

ð3Þ

whereCompensationi,tisthelogoftheaverageexecutivescompensationoffirmiinyeart,NSIDi,tisthenon-strictlyindependentdirectorsmeasure,andXi,tisasetofcontrolvariables.ThestructureofcontrolvariablesfollowsCoreetal.(1999)withsize(thelogofoneperiodlaggedsales),growthopportunities(theaverageofthepastthreeyears'markettobookratio),pastperformance(oneperiodlaggedROAandstockreturn),andrisk(thevarianceofthepastthreeyears'ROAandstockreturn)astheeconomicdeterminantsofcompensationbasedonfirmcharacteristics(seealsoAggarwalandSamwick,1999;CoreandGuay,1999;Cyertetal.,1997;Engesaeth,2013inthisspecialissue;Jacksonetal.,2011;LambertandLarcker,1987).Themodelthenconsiderstheownershipstructureandothercorporategovernancecharacteristicsasproxiesofapotentiallyuncontrolledagencyproblem.Themeasureofownershipstructureis,aspreviously,theamountoftheownershipstakeofthelargestshareholder,theownershipofexecutivedirectorsandofnon-executivedirectors(seeBertrandandMullainathan,2001;ElstonandGoldberg,2003;Hartzelland

InSection6,ﬁrmsresultingfrommergersandacquisitionsareanalyzedasnewentities.

Table6doesnotshowcontrolvariablesduetospaceconstrains.Amongthesevariables,theoneperiodlaggedmarkettobookratioistheonlyonethathasastatisticallysigniﬁcantcoefﬁcientinallmodels(positive,with5%or1%statisticalsigniﬁcancedependingonthemodel).9Alternatively,Andresetal.(2013)andLiuetal.(2013),bothinthisspecialissue,alsousedifferencesindifferencesandinstrumentalvariablesrespectivelytoaddresstheendogeneityconcernwhenoperatingperformanceisthedependentvariable.10Asinothercountries,suchasGermanytill2006,theARCGinSpainjustreportsaggregatecompensationofexecutivedirectors.Indeed,withGermanydata,Andresetal.(2013)inthisspecialissuealsoanalyzetheaverageindividualcompensationofexecutivedirectors.

87Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

12R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Table6

Futurefirm'sperformanceandindependents’misclassification.FirmfixedeffectspaneldataregressionsoffutureReturnonAssetsoneyearaheadexplainedbythepercentageofmisclassifiedindependentsontheboardofdirectors,inthenominationandremunerationcommittee(NRC),intheauditcommittee(AC),andadummyvariabletoidentifyfirmswherethechairofthepreviouscommitteesisamisclassifiedindependentdirector.Models6to10alsocontainthepreviousvariablesinteractedwiththeownershipofthelargestshareholder(C1).Controlvariables(omittedtosavespace)arethelogofsales,thelogoflaggedmarkettobookratio,thelaggedreturnonassets,andthefollowingCorporateGovernancevariables;thepercentageofsharesownedbyexecutives,bynon-executives,andbythelargestshareholderofthefirm,adummyvariableidentifyingwhethertheCEOisalsothechairmanoftheboardofdirectors,thesizeofthisboard,thepercentageofexecutivesinthisboard,adummyvariableidentifyingwhetherdirectorshavedonecommercialtransactionswiththefirm,thepercentageofbusynon-executivedirectors(adirectorisbusyifsheholdsapositioninthreeormoreboardsofdirectors),thepercentageofinterlockedexecutivedirectors(thosewhoarealsoanon-executivedirectorinafirmwhereanon-executivedirectorofthefirstfirmisanexecutivedirector),adummyvariableidentifyingwheretherearegoldenparachutesprotectingtopexecutivesagainstdismissal,andadummyvariableidentifyingfirmswheretherearevotingcaps(amaximumnumberofvotesthatashareholdermayexerciseindependentlyofthenumberofsharesshehas).Yeardummyvariablesareintroduced.Huber(1967)andWhite(1980,1982)robusttstatisticsareinparenthesis.Models2,4,7and9havealowernumberofobservationssincejust86%ofobservations(firm/year)havenominationandremunerationcommittee.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

(1)

%non-strictlyindep.%non-strictlyindep.NRC%non-strictlyindep.ACNRCchairnon-strictlyindep.ACchairnon-strictlyindep.%non-strictlyindep.xC1%non-strictlyindep.NRC×C1%non-strictlyindep.AC×C1NRCchairnon-strictlyindep.×C1ACchairnon-strictlyindep.×C1YearfixedeffectsFirmfixedeffectsObservationsR2AdjustedR2YesYes6770.1210.094

YesYes5800.1350.104

YesYes6770.1210.095

YesYes5800.1390.108

YesYes6770.1320.106

YesYes6770.1210.093

YesYes5800.1350.103

YesYes6770.1210.093

YesYes5800.1400.107

1.571(0.469)

−0.533(−0.380)

−1.167(−0.60)

−1.486*(−1.812)

−2.269*(−1.947)

0.009(0.095)

0.026(0.498)

0.016(0.266)

0.021(0.838)

0.060(1.532)YesYes6770.1370.109

(2)

(3)

(4)

(5)

(6)1.302(0.264)

−1.367(−0.653)

−1.6717(−0.564)

−2.134*(−1.663)

−4.201**(−2.291)

(7)

(8)

(9)

(10)

Starks,2003;Sunetal.,2009).Theremainingboardandcorporategovernancecontrolvariablesarethoseoftheperformancemodel(Eq.(2)),exceptthatwealsoincludethetenureofexecutivedirectorsasadeterminantoftheircompensation.11TheresultsshowninTable7referto535firm/yearobservationsinsteadoftheinitial752becauseoftheexistenceoffirmswithnoexecutivesontheirboard(46observations),missinglaggedstockmarketdatabecauseofnewlistings,mergersandacquisitions(132observations),andseveralfirmsthatdonotreportexecutives'compensationintheAnnualReportofCorporateGovernance(39observations).Whenreferringtothenominationandremunerationcommittee,thefinalsampleisreducedto482firm/yearobservationsbecauseoffirmsthatdonotpossessthisparticularcommittee.TheempiricalmodelsareestimatedwithyearandfirmfixedeffectsandreportHuber–Whiterobusttstatistics.

Theresultsdonotdetectarelationbetweennon-strictlyindependentdirectorsandexecutivecompensation,eveninthenominationandremunerationcommittee.Theinteractionbetweenthemisclassificationmeasuresandtheownershipstructurealsogeneratesnostatisticallysignificantresults.Modelswiththelaggeddependentvariableasanadditionalcontrolvariabletoaddresspotentialendogeneityorestimationsdroppingcorporategovernancecontrolsexceptnon-strictlyindependentsmeasuresdonotbringanysignificancetothenon-strictlyindependentmeasures.12Therefore,wemustconcludethatnomalfunctionisdetectedinthefullboardofdirectorsandonthenominationandremunerationcommitteeregardingexecutivecompensationwhennon-strictlyindependentdirectorsarepresent.13TheremaybesigniﬁcantdifferencesbetweenCEOcompensationandotherexecutivecompensation,andthenumberofexecutivedirectorsmaydistortmechanicallytheaveragecompensationofexecutivedirectors,ourdependentvariable.Thenumberofexecutivedirectorsmaycontrolforthismechanicaleffect,anditisinourmodelthroughtheboardsizeandthepercentageofexecutivedirectors.Furthermore,theintroductionofthenumberofexecutivedirectorsasanadditionalexplanatoryvariabledoesnotaffecttheresults.Resultsareavailableonrequest.12Theseestimationsareavailableuponrequest.Unreportedresultsalsoshowthatnon-strictlyindependentdirectorsintheauditcommitteearealsonotstatisticallyrelatedtoexecutivedirectors'compensation.InTable7wepresenttheresultswithoutthelaggeddependentvariable,asanadditionalcontrol,becausethereisnodataoncompensationbefore2004andweloseoneyearofobservations.13Controlvariables'coefﬁcientsareomittedinTable7.Onlyﬁrmsize(positivecoefﬁcient),theownershipofthelargestshareholder(positivecoefﬁcient),andthepercentageofexecutivedirectors(negativecoefﬁcient)arestatisticallysigniﬁcant(5%,1%and1%ofstatisticalsigniﬁcancerespectively).Asexpected,compensationishigherinlargerﬁrms,thecontroloflargeshareholders,withtheirinvolvementinmanagement,mayexplainthepositiveeffectofownershipincompensation,andthenegativerelationbetweenthepercentageofexecutivedirectorsandtheircompensationisexplainedasfollows;thelargerthenumberofexecutivedirectorsis,thelowertheiraveragecompensationis.

11Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx13

Table7

Executives'compensationandnon-strictlyindependents.Firmfixedeffectspaneldataestimation.Dependentvariable:logofaverageindividualexecutiveboardmembers'remuneration.Thekeyexplanatoryvariablesarethepercentageofnon-strictlyindependentsontheboard,inthenominationandremunerationcommittee(NRC),andthechairingoftheNRCbyanon-strictlyindependent.Controlvariables(omittedforspaceconsiderations)areoneperiodlaggedlogofsales,stockreturn,andofreturnonassets(ROA),theaverageofthemarkettobookratiointhepreviousthreeyears,thevarianceofROAandthevarianceofstockreturnforthethreepreviousyears.Corporategovernancecontrolvariables(omitted)arethepercentageofsharesownedbyexecutives,bynon-executives,andbythelargestshareholderofthefirm,adummyvariableidentifyingwhethertheCEOisthechairmanoftheboardofdirectors,theboardsize,thepercentageofexecutivesontheboard,theaveragetenureofexecutivedirectors,adummyvariableidentifyingwhetherdirectorshavedonecommercialtransactionswiththefirm,thepercentageofbusynon-executivedirectors(adirectorisbusyifsheholdsapositioninthreeormoreboardsofdirectors),thepercentageofinterlockedexecutivedirectors(thosewhoarealsoanon-executivedirectorinafirmwhereanon-executivedirectorofthefirstfirmisanexecutivedirector),adummyvariableidentifyingwheretherearegoldenparachutesprotectingtopexecutivesagainstdismissal,andadummyvariableidentifyingfirmswheretherearevotingcaps(amaximumnumberofvotesthatashareholdermayexerciseindependentlyofthenumberofsharesshehas).Yeardummyvariablesareintroduced.Huber(1967)andWhite(1980,1982)robusttstatisticsareinparenthesis.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

(1)

%non-strictlyindep.%non-strictlyindep.NRCNRCchairnon-strictlyindep.%non-strictlyindep.×C1%non-strictlyindep.NRC×C1NRCchairnon-strictlyindep.×C1YearfixedeffectsFirmfixedeffectsObservationsR2AdjustedR2YesYes5350.24780.2124

YesYes4820.26460.226

YesYes4820.26090.2221

YesYes5350.24780.2109

YesYes4820.26470.2244

0.406(1.62)

0.296(1.375)

0.117(1.21)

−0.002(−0.177)

0.001(0.233)

−0.0008(−0.316)YesYes4820.26110.2205

(2)

(3)

(4)0.444(1.351)

0.259(1.09)

0.141(1.267)

(5)

(6)

6.2.2.CEOturnover

Theevidenceregardingtheuseofindependentandoutsidedirectorsasamethodofincreasingturnover-sensitivitytopastperformanceismixed,asindicatedbyFranksetal.(2001).CEOturnoverandboardindependencearetreatedassubstitutemechanismsinEasterwoodetal.(2012)subsequenttofirmunderperformance.TheHermalinandWeisbach(1998)modelshowsthattheCEO'sroleinselectingboardmembersdecreasesdisciplinewhenfirmsunderperform.Thisevidenceallowsustoindicatethatnon-strictindependenceintheboardroommightdecreasethesensitivityofCEOturnoverforunderperformingfirms,reflectinganuncontrolledagencyproblem.

BecauseofthelackofstandardizedAnnualReportsofCorporateGovernancebefore2004,wecanonlydetectCEOturnovereventsfrom2005to2009,resultinginadatabasewith634firm/yearobservations,1414%ofthemcorrespondingtoCEOturnoverevents(89observations).Thisnumberisobservedtoincreaseovertime,from14in2005to20in2009,andisregularlydistributedacrossindustries.TheCEOturnoverof4eventsoverlapswithamergeroracquisition.

Weanalyzetherelationshipbetweenthepresenceofnon-strictlyindependentdirectorsandCEOturnovereventswithanempiricallogitmodelinwhichthedependentvariableidentifiestheturnoverevents.AsinHwangandKim(2009)theexplanatoryvariablesareoneperiodlaggedstockreturn,agroupofcorporategovernancevariableslaggedoneperiodtoproxyanyuncontrolledagencyproblem,andtheinteractionwiththelaggedstockreturn:

Turnoveri;t

󰀁󰀃󰀁󰀃

¼αþα1ÁReti;t−1þα2ÁNSIDi;t−1þα3ÁReti;t−1×NSIDi;t−1þXi;t−1Áβ1þReti;t−1×Xi;t−1Áβ2þεi;t

ð4Þ

whereRetisstockreturn,NSIDisthenon-strictlyindependentdirectorsmeasure,andXisthesetcorporategovernancecontrolvariables,thesameasintheperformancemodel(Eq.(2)).15Giventhatexplanatoryvariablesarelaggedoneperiod,77observationsarelostbecauseofmissingpaststockreturnsinnewlistingsandasaresultofmergersandacquisitions,leavingafinalsampleof557observations.Thefirmsresultingfrommergersandacquisitionsareanalyzedasnewentities.Finally,firmswithnonominationandremunerationcommitteegenerate88additionalmissingobservationsintheestimationinwhichthiscommitteeisrelevant.

TheSpanishstandardizedAnnualReportofCorporateGovernancedoesnotexplicitlyidentifytheCEO.Ouridentiﬁcationprocedureincludesﬁvesteps.First,itiswhenaﬁrmdeclaresthatthetopexecutiveoftheﬁrmchairstheboard(441identiﬁcations).Second,forﬁrmswithoutchair-CEOduality,theCEOisidentiﬁedasthe“ConsejeroDelegado”amongboarddirectors(214CEOsidentiﬁed).Third,ifnoCEOisstillidentiﬁed,theﬁgureof“directorgeneral”amongthelistofnon-directortopexecutivesistakenasCEO(41CEOs).Fourth,iftheCEOisstillnotidentiﬁedthetopexecutivedirectorisselected(49CEOs).Insevenobservations(year/ﬁrm)noCEOisstillidentiﬁed;however,theCEOoftheposterioryearisthesameastheCEOofthepreviousyear,andshe/heisidentiﬁedastheCEO.15Thetenureofexecutivedirectorsisnotincludedbecauseseveralﬁrms(46observations)donothaveexecutivedirectors.

14Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

14R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Table8showstheestimationoftheCEOturnovermodelwithGEEpaneldatalogitestimationstoaccountforanyunobservedpersistenceintheresiduals,withrobust(Huber–White)standarderrors,andtimeandindustrialsectorfixedeffects.Modelsaccountforthepercentageofnon–strictlyindependentdirectorsoverboardsize,thepercentageofnon-strictlyindependentdirectorsinthenominationandremunerationcommittee,andadummyvariableforthenon-strictlyindependentschairingthiscommittee.CorporategovernancecontrolvariablesareomittedinTable8tomakeitshorter;thecompleteresultsareavailableonrequest.

TheresultsshowtheexpectednegativerelationbetweenpastperformanceandtheprobabilityofaCEOturnoverevent(seealsoFiordelisiandRicci,2013;Liu,2013inthisspecialissue),andnostatisticallysignificantrelationbetweennon-strictlyindependentdirectorsandCEOturnoverevents.Thisnon-significantrelationremainswhenpaneldatarandomeffectslogitmodelswithbootstrapstandarderrorsareestimated,notreportedforbrevity.TheunreportedresultsinTable8showthatCEOturnovereventsarelessprobablewhentheCEOchairstheboardofdirectors(1%statisticalsignificanceinmodels1and3,and5%inmodel2)andthehighertheproportionofexecutivedirectorsis(10%statisticalsignificanceinmodels2and3).Finally,thehigherprobabilityofCEOturnoverafterfirmunderperformanceisreducedthelargerthesizeoftheboardofdirectorsis(5%statisticalsignificanceinmodel1,10%inmodels2and3),consistentwithmonitoringcoordinationproblems.166.3.Outcomesoftheauditcommittee

6.3.1.Auditqualiﬁcations

Thereisevidencethatboardandauditcommitteeindependenceaffectsthecontrolofmanagersthroughthequalityofaccountinginformation.Andersonetal.(2004)suggestthatmoreindependentauditcommitteesprovideameasurableandsignificantbenefittofirmsbecausethesecommitteesprovidemorereliableaccountinginformationandaffectthecostofdebt.KnechelandWillekens(2006)arguethatindependentboardmembersapplypressuretoenhancetheexternalauditfunction.Independentboardmembersmaybemoreconcernedabouttheirpersonalexposureifmanagersmisbehave.Therefore,thesemembersaremoreinterestedinanextensiveauditactivitytominimizetheriskofmanagerialmisbehavior,whichcouldaffectthemembers'personalreliability.Forfirmswithnon-strictlyindependentdirectors,thepressureonreliableaccountinginformationisexpectedtobelowercomparedtoboardswithoutmisclassifiedindependents.

Theexternalauditoropinionsintheannualfinancialreport,availablefortheauditedfirms,allowustoanalyzewhetherthepresenceofnon-strictlyindependents,particularlyintheauditcommittee,isrelatedtotheprobabilityofaqualifiedopinion(notclean).WefollowtheempiricalmodelofauditqualificationbySánchez-BallestaandGarcía-Meca(2005)testedonaSpanishsamplefrom1999to2002:

󰀁󰀃

ð5ÞAuditQualificationi;t¼αþα1ÁNSIDi;tþα2ÁC1i;t×NSIDi;tþXi;tÁβþεi;twhereAuditQualificationi,tisadummyvariableequal1ifthereisaqualifiedopinion,NSIDi,tisthenon-strictlyindependent

directorsmeasure,andXi,tisthesetofcontrolvariables.Theaccountingliteratureshowsthatseveralcharacteristicsinfluencetheauditoropinion.Firmsizeinourpaperismeasuredasthelogofsales.Largerfirmsaremorelikelytoreceiveaqualifiedopinionbecausetheopportunitycostofauditorsincreasesifamisstatementisdetected.Althoughlargerfirmstendtohavemoredevelopedinternalfinancialcontrolsystems,andauditorsmayhesitatetoissueaqualifiedopinionbecauseofapotentiallossofsignificantfees.Highleverage(long-termdebtovertotalassets)asaproxyofcompanyriskmakestheauditorsstricter,expectingapositiverelationwithauditqualifications.Operatingperformance(ROA)andliquidity(currentassetsovercurrentliabilities)proxytheprobabilityoffirmfailureandareexpectedtobenegativelyrelatedwithauditqualifications.Lowerperformanceandliquiditymaketheauditorsstrictertoavoidlitigationcostsincaseoffirmfailure.Finally,inlinewithSánchez-BallestaandGarcía-Meca(2005),weincludeownershipstructure(C1andtheownershipofexecutiveandnonexecutivedirectors)andtheboardofdirectors'coordinationproblems(boardsize)asadditionalcontrolvariables.Whenownerscareabouttheauditorqualifiedopinion,theexpectedconsequencesareaneffectivemonitoringovermanagement.

Weeliminatebanksfromthesamplebecauseofthespecialregulationsintermsoffinancialreporting(78observations).Intheremaining674observations,wefind79auditqualifications.Thisnumberincreasesfrom10in2004to18in2009andsimilarlyaffectsallindustrialsectors.Thelowfrequencyofauditqualificationsanditspersistenceovertimereducesthepowerofstandardeconometricexplanatorytechniques.Therefore,followingaccountingliteraturestandards,weestimatethelogitempiricalmodelinareducedsampleofauditqualificationswithamatchingsample(seealsoDopuchetal.,1987).Foreachfirm,weusethefirstauditqualification(31observations).Thematchingsampleselectionisamongfirmswithnoauditqualificationinthesameindustryandyearwiththeclosestamountofsales.Afirmentersintothematchingsamplejustonetime.

Table9showstheestimationoftheempiricallogitmodelofauditqualificationsonthisreducedsamplewithrobust(Huber–White)standarderrorsandyearandindustrialsectorfixedeffects.17Models1to3inTable9measurenon-strictlyindependenceasthepercentageofnon-strictlyindependentdirectorsoverboardsize,thepercentageofnon-strictlyindependentdirectorsintheauditcommitteeoveritssize,andadummyvariabletoidentifyfirmsinwhichthechairoftheauditcommitteeisanon-strictlyindependentdirector.Theresultsshownostatistically

Non-tabulatedresultsalsoshownostatisticallysigniﬁcantrelationbetweennonstrictly-independentsintheauditcommitteeandCEOturnoverevents.Availableonrequest.17Noclusteringisintroducedtocomputetherobuststandarderrorsandnopaneldatatechniquesareusedbecausethereisjustoneobservationperﬁrminthisreducedsample.

16Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx15

Table8

CEOturnoverandindependentsmisclassification.GEEpaneldatalogitmodelsallowingpersistenceintheresidualswithHuber(1967)andWhite(1980,1982)robusttstatistics(inparenthesis),wherethedependentvariableisadummyvariabletoidentifyCEOturnoverevents.Thekeyexplanatoryvariablesareoneperiodlaggedstockreturn,oneperiodlaggedpercentageofnon-strictlyindependentsontheboardofdirectors,inthenominationandremunerationcommittee(NRC),adummyvariabletoidentifyfirmswherethechairofthepreviouscommitteeisamisclassifiedindependentdirector,anditsinteractionwiththelaggedstockreturn.Controlvariables(omittedtosavespace)arelaggedoneperiod;percentageofsharesownedbyexecutives,bynon-executives,andbythehighestshareholderofthefirm,adummyvariableidentifyingwhethertheCEOisalsothechairmanoftheboardofdirectors,thesizeofthisboard,thepercentageofexecutivesinthisboard,adummyvariableidentifyingwhetherdirectorshavedonecommercialtransactionswiththefirm,thepercentageofbusynon-executivedirectors(adirectorisbusyifsheholdsapositioninthreeormoreboardsofdirectors),thepercentageofinterlockedexecutivedirectors(thosewhoarealsoanon-executivedirectorinafirmwhereanon-executivedirectorofthefirstfirmisanexecutivedirector),adummyvariableidentifyingwheretherearegoldenparachutesprotectingtopexecutivesagainstdismissal,andadummyvariableidentifyingfirmswheretherearevotingcaps(amaximumnumberofvotesthatashareholdermayexerciseindependentlyofthenumberofsharesshehas),andtheinteractionbetweenthesevariablesandoneperiodlaggedstockreturn.Finally,industrialsectorandyeardummyvariablesareintroduced.Chi2isaWaldtestofthestatisticalsignificanceofalltheexplanatoryvariables.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

(1)

Stockreturnt−1%non-strictlyindep.%non-strictlyindep.

t−1(2)−3.644***(−2.262)

(3)−3.645***(−2.245)

t−1×Stockreturnt−1t−1−4.569***(−2.789)0.3715(0.446)1.8175(0.944)

%non-strictlyindep.NRC%non-strictlyindep.NRC

t−1×Stockreturnt−1t−10.448(0.804)0.312(0.280)

0.1448(0.445)0.0002(0.000)YesYes469

61.312***

NRCchairnon-strictlyindependentNRCchairnon-strictlyindependentYearfixedeffectsIndustryfixedeffectsObservationsChi2t−1×Stockreturnt−1YesYes557

70.026***

YesYes469

61.190***

significantrelationshipamongnon-strictindependenceandauditqualifications.Inmodels4to6ofTable9,theinteractiontermbetweentheindependents'misclassificationmeasuresandtheownershipofthelargestshareholderalsoshownostatisticallysignificantrelations.ControlvariablesareomittedinTable9,wherethemeasuresofoperatingperformanceandliquidityarestatisticallysignificant(bothat5%ofconfidencelevel)andbehaveasexpected:thelowertheoperatingperformanceandtheliquidity(currentassetsovercurrentliabilities)are,thehigherthefirmfailureriskis,andthehighertheprobabilityoffinancialstatementsauditqualificationis.

Additionalestimationstochecktherobustnessofthepreviousfindingsincludealogitmodelinthefullsample(674observations)withpaneldatarandomeffects.Inaddition,asetofGEEpaneldatalogitmodelshavebeentestedtoaccountforanyunobservedpersistenceintheresiduals,withnostatisticallysignificantrelationfoundbetweennon-strictlyindependentsandauditqualifications.186.3.2.Earningsmanagement

Klein(2002)suggeststhatboardsstructuredtobemoreindependentaremoreeffectiveinmonitoringthecorporatefinancialaccountingprocess.Thisassertionisbasedonevidencethatreductionsinboardorauditcommitteeindependenceareaccompaniedbylargeincreasesinabnormalaccruals.Xieetal.(2003)findaparticularlyrelevantroleoftheauditcommitteeinthepropensityofmanagerstoengageinearningsmanagement.Therefore,ifthepresenceofnon-strictlyindependentdirectorsistheresultofanuncontrolledagencyproblem,wemayexpectmoreearningsmanagement,particularlywithnon-strictlyindependentsontheauditcommittee.

Earningsmanagementmaybeindicativeofinformationprovisionorinformationmanipulation.Somepreviousaccountingstudiesuseddiscretionaryaccrualsmodelstodecomposetheproxiesofearningsmanagementintoaninformationcomponentandamanipulationcomponent(Jones,1991).Giventhecriticismofthesemodels(e.g.,Dechowetal.,1995;Guayetal.,1996),however,recentaccountingliteratureintroducedaregressionapproachtodecomposetheinformativecomponentofearningsmanagementfromthemanipulativecomponent(GopalanandJayaraman,2012).Wefollowthisapproachtodetectthemanipulationcomponentofearningsmanagementthroughthefundamentaldeterminantsoftheinformativeportion,asdifferencesinfirms'underlyingbusinessprocessorinthelengthoftheoperatingcycle(Dechow,1994).Weestimatethefollowingempiricalmodel:

󰀁󰀃

EarningsManagementMeasurei;t¼αþα1ÁNSIDi;tþα2ÁC1i;t×NSIDi;tþXi;tÁβþεi;t

ð6Þ

Statisticalinferenceobtainedwithbootstrapstandarderrors.Additionally,inthefullandinthereducedsample,wealsoestimatethelogitmodelonlywiththemisclassiﬁedindependentsmeasuresasexplanatoryvariablesandnostatisticallysigniﬁcantresultsarefound.

18Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

16R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Table9

Auditqualificationsandindependents'misclassification.LogitmodelswithHuber(1967)andWhite(1980,1982)robusttstatistics(inparenthesis),wheredependentvariableisadummyvariabletoidentifyaqualifiedopinionoftheexternalauditorintheannualfinancialreport.Thekeyexplanatoryvariablesarethepercentageofnon-strictlyindependentsontheboardofdirectors,ontheauditcommittee(AC),adummyvariabletoidentifyfirmswherethechairofthepreviouscommitteeisamisclassifiedindependentdirector,anditsinteractionwiththeownershipofthelargestshareholder.Controlvariables(omittedtosavespace)aretheownershipofthelargestshareholder,returnonassets,leverage(longtermdebtovertotalassets),thelogofsales,thesizeoftheboardofdirectors,theownershipofexecutivedirectorsandofnonexecutivedirectors,andLiquidity(currentassetsovercurrentliabilities).Finally,industrialsectorandyeardummyvariablesareintroduced.Allmodelsareestimatedwithasampleof31firstauditqualificationswithamatchingsamplebyyear,industrialsector,andsales.Chi2isaWaldtestofthestatisticalsignificanceofalltheexplanatoryvariables.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

(1)

%non-strictlyindep.%non-strictlyindep.ACACchairnon-strictlyindep.%non-strictlyindep.×C1%non-strictlyindep.AC×C1ACchairnon-strictlyindep.×C1YearfixedeffectsIndustryfixedeffectsObservationsChi2YesYes6218.512

YesYes6219.090

YesYes6217.220

YesYes6217.666

YesYes6217.236

−1.716(−0.688)

−0.621(−0.495)

−0.426(−0.498)

−0.083(−1.158)

−0.053(−1.188)

−0.024(−0.883)YesYes6216.540

(2)

(3)

(4)0.647(0.194)

1.071(0.554)

0.398(0.302)

(5)

(6)

wereXisthesetofproxiesofthedeterminantsoftheinformativeportionofearningsmanagement.FollowingGopalanandJayaraman(2012)andLeuzetal.(2003),earningsmanagementisdefinedastheabsolutevalueofaccruals(ACC)dividedbytheabsolutevalueofcashflowsfromoperations.Theuseofaccrualsisinterpretedasmanagerialdiscretiontoconcealtrueeconomicperformancefromoutsiders(HealeyandWahlen,1999).Cashflowfromoperationsisdefinedasoperatingincomeminusaccruals.Accrualsarecalculatedasfollows:

hihi

ð7ÞACCi;t¼ΔCAi;t−ΔCashi;t−ΔCLi;t−ΔSTDi;t−Depi;twhereΔCAi,tisthechangeintotalcurrentassetsforfirmiinyeart,ΔCashi,tisthechangeincash,ΔCLi,tisthechangeintotal

currentliabilities,ΔSTDi,tisthechangeindebtincludedincurrentliabilities,andDepi,tisthedepreciationandamortizationexpense.

Theseaccrualsarenaturallyaffected(withnomanipulationofinformation)bydifferencesinfirms’operatingcycle,thecreditofsuppliers,thevolatilityoftheoperatingenvironment,capitalintensity,andprofitability(Dechow,1994;DechowandDichev,2002;HribarandNichols,2007).Consequently,weincludethefollowingasexplanatoryvariablesofearningsmanagement(XinEq.(6)):thesumofdaysreceivableanddaysinventory(operatingcycle),dayspayable(thecreditofsuppliers),thevarianceofROAinthepreviousthreeyears(thevolatilityoftheoperatingenvironment),fixedassetsovertotalassets(capitalintensity),andadummyvariabletoidentifyfirmswithnegativenetincome(profitability).19AligningwithGopalanandJayaraman(2012),wealsoincludedifferencesingrowthopportunities(theaverageofthemarkettobookratiooverthepreviousthreeyears),insize(thelogofsales),andinleverage(long-termdebtovertotalassets)asadditionaldeterminants.

Fromtheinitialsampleof752observations,weeliminatebanks(78observations)giventhissector'sspecialregulationsonfinancialreportingand89additionalobservationsbecauseofmissingdataonstockmarketinformationfornewfirmsandnewlistingsinthestockmarket(firmsresultingfrommergersandacquisitionsareanalyzedasanewentity),whichresultsinasampleof585observations.

Table10showstheestimatedempiricalmodelinwhichthedependentvariableisthelogoftheearningsmanagementmeasure,andomitsthecoefficientsofcontrolvariables.Thekeyexplanatoryvariablesarethepercentageofnon-strictlyindependentdirectorsoverboardsize,thepercentageofnon-strictlyindependentsintheauditcommitteeoveritssize,adummyvariabletoidentifyfirmswithanon-strictlyindependentchairingthiscommittee,andtheinteractionofthesevariableswiththeownershipofthelargestshareholder.Controllingfortheabovementionedvariables,nosignificantrelationisfoundbetweenearningsmanagementandthepresenceofmisclassifiedindependentdirectors.ConsistentwithGopalanandJayaraman(2012)wefindthatreportinglossesandfinancialleveragearepositivelyrelatedwithearningsmanagementandthatgrowthopportunitiesarenegativelyrelatedwithearningsmanagement.

Theestimationmethod,pooledOLSwithyearandindustrialsectorfixedeffectsandrobusttstatistics(Huber–White)clusteredbyfirm,isselectedtobecomparablewithGopalanandJayaraman(2012).Alternativeestimationswithfirmfixed

Daysreceivableiscomputedbydividing360bytheratioofsalesdividedbytheaverageaccountsreceivable,daysinventorybydividing360bytheratioofcostofgoodssolddividedbytheaverageinventory,anddayspayablebydividing360bytheratioofthecostofgoodssolddividedbytheaverageaccountspayable.Averagesarecomputedwiththepreviousannualﬁnancialstatements.

19Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx17

Table10

Earningsmanagementandindependents'misclassification.PooledOLSregressionwithyearandindustryfixedeffectmodelswithHuber(1967)andWhite(1980,1982)robusttstatistics(inparenthesis),clusteringbyfirm(Petersen,2009).Dependentvariable:thelogofabsolutevalueofaccrualsoverabsolutevalueofoperatingcashflow.Thekeyexplanatoryvariablesarethepercentageofnon-strictlyindependentsontheboardofdirectors,ontheauditcommittee(AC),adummyvariabletoidentifyfirmswherethechairofthepreviouscommitteeisamisclassifiedindependentdirector,anditsinteractionwiththeownershipofthelargestshareholder(C1).Controlvariables(omittedtosavespace)areadummyvariabletoidentifyfirmsreportingnegativenetincome,OperatingCycleasthesumofdaysreceivableanddaysinventorytimes10−4,DaysPayablemultipliedby10−4,theaverageofthemarkettobookratiointhepreviousthreeyears,thelogofsales,thevarianceofROAusingtheannualdataofthethreepreviousyearstimes10−2,leverage(longtermdebtovertotalassets),andCapitalIntensity(fixedassetsovertotalassets).Yearandindustrialsectordummyvariablesarealsointroduced.***denotessignificanceatthe1%level;**denotessignificanceatthe5%level;*denotessignificanceatthe10%level.

(1)

YesYes5850.0620.030

YesYes5850.0620.031

YesYes5850.0630.030

YesYes5850.0620.029

−0.064(−0.190)

−0.017(−0.096)

0.050(0.474)

−0.007(−0.72)

−0.003(−0.46)

0.000(0.129)YesYes5850.0620.029

(2)

(3)

(4)0.129(0.295)

0.069(0.288)

0.035(0.235)

(5)

(6)

effectsalsoleavethemeasuresofindependentsmisclassificationasnon-significanttoexplainearningsmanagement.Finally,whenthemeasureoflaggedearningsmanagementisusedasanadditionalexplanatoryvariabletoaddressendogeneity,nodifferencesaredetected.Theresultsareavailableuponrequest.7.Discussion

Accordingtotheoptimalboardindependencetheory,non-strictlyindependentsintheboardroomcouldbeexplainedastheadjustmenttotherecommendationsofregulatorsandshareholderadvocatesregardingthedesirablelevelofindependentsontheboardthroughcorporategovernancecodes(thedeclaredlevel)whenthelevelishigherthantheoptimaldegreeofindependence(strictlyindependents).

Underthistheory,anendogenouslydeterminedhigherpowerofexecutivesoroftheCEOinfrontofshareholderswouldinduceastrongerpresenceofnon-strictlyindependents.Effectively,accordingtoHermalinandWeisbach(1998)andBooneetal.(2007),thehigherthispoweris,thelowertheoptimumlevelofboardindependenceforshareholdersinterestis.20ThefindingsinTable5indicatethatthestrongerthemanagerialpoweris,thehighertheprobabilityofnon-strictlyindependentsintheboardroomis.

However,thepresenceofnon-strictlyindependentdirectorswhenmanagershavestrongerpowerhastwoexplanations.First,thisscenariocanindicateanuncontrolledagencyproblemthatisharmfultothefirm,therebyincreasingagencycostsandnegativelyaffectingfirmperformance.Alternatively,thispresencecanbeawaytoadjusttheoversizingofdeclaredindependentstotheoptimalindependencelevel,hencewithoutnegativeimpactonthefirmperformance.

OurfindingsinTable6showweakevidenceoftheimpactofnon-strictlyindependentsontheboardonfutureoperatingperformance.Effectively,theproportionofnon-strictlyindependentsontheboardofdirectors,onthenominationandremunerationscommitteeorontheauditcommitteedoesnotaffectfutureoperatingperformance.Nevertheless,whennon-strictlyindependentdirectorspresideovereithercommittee,weobtainweakempiricalevidenceofanegativeimpactonfutureoperatingperformance.

Thesupervisoryroleofbothcommitteesisreinforcedbyregulatorsinthebestpracticescodesofgoodgovernance,whichrecommendthatcommitteesshouldbechairedbyanoutsideindependentdirectorandhaveasignificantproportionofindependentsamongtheirmembers.Thesupervisoryroleofthenominationandremunerationcommitteeisparticularlyvaluablewithregardtotwodecisions:executivecompensationandCEOremovalwhenthefirmunderperforms.OurresultsinTable7showthatthepresenceofnon-strictlyindependentsintheboardroomorinthenominationandremunerationcommitteedoesnotaffectboardexecutives'compensationandalsodoesnotaffectwhetherthiscommitteeischairedbyanon-strictlyindependentdirector.Similarly,theusualimpactofpoorperformanceontheCEOrateofturnoverisneitherenhancednorreducedbythepresenceofnon-strictlyindependentsontheboard,onthecommitteeorchairingthecommittee,asTable8indicates.

Forexample,intheHermalinandWeisbach(1998)model,positivepastperformancegeneratestheCEO'spowerwithalessindependentboardasanoptimaloutputforshareholders'interest.

20Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

18R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx

Weaddressthesupervisorybehavioroftheauditcommitteeinthefollowingtwospecificactions:thelevelofauditqualificationsandthefirms'earningsmanagement.InTable9,weshowthatnoneofthemeasuresofalackofstrictindependenceexplaintheobservedlevelofauditqualifications.Regardingtheimpactonthemeasureofearningsmanagementandthelevelofaccruals,themodelsfromTable10behaveasusualintheaccountingliteratureandthemeasuresofnon-strictindependencedonotaffectthisrelationship.

Basedonthesefindings,theappointmentofnon-strictlyindependentsintheboardroomappearstocomplywiththerequiredproportionofindependentssuggestedbythecodesofgoodgovernance.Becausethesenon-strictlyindependentsonthemainboardcommitteesdonotnegativelyaffectdecisions,thisapproachprovidessupportfortheoptimalboardindependencetheory.Themanagerialpowerapproachreceivesweakempiricalsupportbasedonthefactthatthechairingofthesecommitteesbyanon-strictindependentdirectordecreasesoperatingperformance.8.Conclusion

Insum,thispaperprovidesasetofeightcriteriatodeterminethestrictnessoftheindependenceofthedeclaredindependentdirectorsaccordingtotheSpanishregulation,whichwedemonstrateisinlinewithinternationalstandardsonthedefinitionofindependentdirectorsincludingNYSEstandards,theEuropeanUnionCommission,andtheUKcorporategovernancecode.TheempiricalmeasureoftheseeightcriteriaonapaneloflistedfirmsintheSpanishstockmarketshowsthatthelevelofmisclassificationsofindependentdirectorsismorethan50%.Thislevelofnon-strictindependenceisdecreasingovertimeandaffectsfirmsofallsizesandindustries.Firmswithmoredispersedownershipappointnon-strictlyindependentsmorefrequentlythanfirmswithlargesignificantshareholders.However,thepresenceofnon-strictlyindependentdirectorsdoesnotaffectseveralrelevantoutputsofthemainboardcommitteesintermsofmonitoringandonlyweakempiricalsupportisfoundforanegativerelationwithfutureoperatingperformance.Therefore,lowsupportisfoundforthemanagerialpoweroriginofnon-strictlyindependents.Strongersupportisfoundfortheoptimalboardindependencetheory,inwhichfirmsappointnon-strictlyindependentstosimultaneouslyachievetheoptimalrealindependencelevelandtherecommendedindependencelevel.Thelargeamountofindependentsrecommendedincorporategovernancecodesdoesnotnecessarilyimprovecorporategovernancepractices,andconversely,mayprovidefirmswithincentivestoappointnon-strictlyindependents.Acknowledgments

Wethankthescientificcommitteeofthisspecialissueandtheanonymousreviewerfortheirvaluablefeedback.Wealsothanktheparticipantsofthe2013EFMAconferenceheldinReading(UK)forhelpfulcomments,andtheFundacióndeEstudiosFinancieros(www.fef.es)whofacilitatedusseveralmeetingswithcorporateboarddirectorsoflistedcompanies.WeacknowledgethefinancialsupportoftheSpanishministryprojectsECO2010-18567andECO2010-21393-C04-02.Weareresponsibleforallerrors.References

Adams,R.,Ferreira,D.,2007.Atheoryoffriendlyboards.J.Financ.62,217–250.

Adams,R.,Hermalin,B.,Weisbach,M.,2008.Theroleofboardsofdirectorsincorporategovernance:aconceptualframeworkandsurvey.NBERWorkingPaper

14486(http://www.nber.org/papers/w14486).

Aggarwal,R.,Samwick,A.,1999.Theothersideofthetrade-off:theimpactofriskonexecutivecompensation.J.Polit.Econ.107,65–105.

Agrawal,A.,Knoeber,C.,1996.Firmperformanceandmechanismstocontrolagencyproblemsbetweenmanagersandshareholders.J.Financ.Quant.Anal.31,

377–396.

Aguilera,R.,Cuervo-Cazurra,A.,2009.Codesofgoodgovernance.Corp.Gov.Int.Rev.17,376–387.

Anderson,R.,Mansi,S.,Reeb,D.,2004.Boardcharacteristics,accountingreportintegrity,andthecostofdebt.J.Account.Econ.37,315–342.

Andres,C.,Fernau,E.,Theissen,E.,2013.ShouldIstayorshouldIgo?FormerCEOsasmonitors.J.Corp.FinanceXX,XXX-XXX(thisspecialissue).Ansari,I.,Goergen,M.,Mira,S.,2013.ThedeterminantsoftheCEOsuccessorchoiceinfamilyfirms.J.Corp.FinanceXX,XXX-XXX(thisspecialissue).Ballinger,G.,2004.Usinggeneralizedestimatingequationsforlongitudinaldataanalysis.Organ.Res.Methods7,127–150.

Benkel,M.,Mather,P.,Ramsay,A.,2006.Theassociationbetweencorporategovernanceandearningsmanagement:theroleofindependentdirectors.Corp.

Ownersh.Control.3,65–75.

Bertrand,M.,Mullainathan,S.,2001.AreCEOsrewardedforluck?Theoneswithoutprincipalsare.Q.J.Econ.116,901–932.

Bhagat,S.,Black,B.,2002.Thenon-correlationbetweenboardindependenceandlong-termfirmperformance.J.Corp.Law27,1231–1273.Bhagat,S.,Bolton,B.,Sarbanes-Oxley,J.,2009.Governanceandperformance.SSRNWorkingPaper(http://ssrn.com/abstract=1361815).

Boone,A.,Field,L.,Karpoff,J.,Raheja,C.,2007.Thedeterminantsofcorporateboardsizeandcomposition:anempiricalanalysis.J.Financ.Econ.85,66–101.Brick,I.,Palmon,O.,Wald,J.,2006.CEOcompensation,directorcompensation,andfirmperformance:evidenceofcronyism?J.Corp.Finance12,403–423.Byrd,J.,Hickman,K.,1992.Dooutsidedirectorsmonitormanagers?Evidencefromtenderofferbids.J.Financ.Econ.32,195–207.Byrd,J.,Coopermand,E.,Wolfe,G.,2009.Anotherlookatdirectorindependence.IRABF1,1–16.

Cicero,D.,Wintoki,M.,Yang,T.,2013.Howdopubliccompaniesadjusttheirboardstructures?J.Corp.Finance23,108–127.

Cohen,L.,Frazzini,A.,Malloy,C.,2012.Hiringcheerleaders:boardappointmentsof“independent”directors.Manag.Sci.58,1039–1058.Coles,J.,Daniel,N.,Naveen,L.,2008.Boards:doesonesizefitall?J.Financ.Econ.87,329–356.

Coles,J.,Daniel,N.,Naveen,L.,2010.Co-optedboards.SSRNWorkingPaper.(http://ssrn.com/abstract=1699272).

Core,J.,Guay,W.,1999.Theuseofequitygrantstomanageoptimalequityincentivelevels.J.Account.Econ.28,151–184.

Core,J.,Holthausen,R.,Larcker,D.,1999.Corporategovernance,chiefexecutiveofficercompensation,andfirmperformance.J.Financ.Econ.51,371–406.Cotter,J.,Shivdasani,A.,Zenner,M.,1997.Doindependentdirectorsenhancetargetshareholderwealthduringtenderoffers?J.Financ.Econ.43,195–218.Cyert,R.,Kang,S.,Kumar,P.,Shah,A.,1997.CorporategovernanceandthelevelofCEOcompensation.WorkingPaper.CarnegieMellonUniversity.

Dalton,D.,Daily,C.,Ellstrand,A.,Johnson,J.,1998.Meta-analyticreviewsofboardcomposition,leadershipstructure,andfinancialperformance.Strateg.Manag.J.

19,269–290.

Dechow,P.,1994.Accountingearningsandcashflowasmeasuresoffirmperformance:theroleofaccountingaccruals.J.Account.Econ.18,3–42.

Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

R.Crespí-Cladera,B.Pascual-Fuster/JournalofCorporateFinancexxx(2013)xxx–xxx19

Dechow,P.,Dichev,I.,2002.Thequalityofaccrualsandearnings:theroleofaccrualsestimationerrors.Account.Rev.77,35–59.Dechow,P.,Sloan,R.,Sweeney,A.,1995.Detectingearningsmanagement.Account.Rev.70,193–225.

Dopuch,N.,Holthausen,R.,Leftwich,R.,1987.Predictingauditqualificationswithfinancialandmarketvariables.Account.Rev.62,431–454.Duchin,R.,Matsusaka,J.,Ozbas,O.,2010.Whenareoutsidedirectorseffective?J.Financ.Econ.96,195–214.Dyck,A.,Zingales,L.,2004.Privatebenefitsofcontrol:aninternationalcomparison.J.Financ.59,537–600.

Easterwood,J.,İnce,Ö.,Raheja,C.,2012.TheevolutionofboardsandCEOsfollowingperformancedeclines.J.Corp.Finance18,727–744.Elston,J.,Goldberg,L.,2003.ExecutivecompensationandagencycostsinGermany.J.Bank.Financ.27,1391–1410.

Engesaeth,E.,2013.Improvingthedecision-makingofthecompensationcommitteeonCEOcompensationrisk.J.Corp.FinanceXX,XXX-XXX(thisspecialissue).Faccio,M.,Lang,l.,Young,L.,2001.Dividendsandexpropriation.Am.Econ.Rev.91,54–78.

Ferris,S.,Yan,X.,2007.Doindependentdirectorsandchairmenmatter?Theroleofboardsofdirectorsinmutualfundgovernance.J.Corp.Finance13,392–420.Fich,E.,Shivdasani,A.,2006.Arebusyboardseffectivemonitors?J.Financ.61,689–724.

Fiordelisi,F.,Ricci,O.,2013.CorporatecultureandCEOturnover.J.Corp.FinanceXX,XXX-XXX(thisspecialissue).

Franks,J.,Mayer,C.,Renneboog,L.,2001.Whodisciplinesthemanagementofpoorlyperformingcompanies?J.Financ.Intermed.10,209–248.

Gillette,A.,Noe,T.,Rebello,M.,2003.Corporateboardcomposition,protocols,andvotingbehavior:experimentalevidence.J.Financ.58,1997–2031.Goergen,M.,Renneboog,L.,2011.Managerialcompensation.J.Corp.Finance17,1068–1077.

Gopalan,R.,Jayaraman,S.,2012.Privatecontrolbenefitsandearningsmanagement:evidencefrominsidercontrolledfirms.J.Account.Res.50,117–157.

Gordon,J.,2007.TheriseofindependentdirectorsintheUnitedStates,1950–2005:ofshareholdervalueandstockmarketprices.ColumbiaLawandEconomics

WPNº.301.(http://ssrn.com/abstract=928100).

Guay,W.,Kothari,S.,Watts,R.,1996.Amarket-basedevaluationofdiscretionary-accrualmodels.J.Account.Res.34,83–105.Hallock,K.,1997.Reciprocallyinterlockingboardsofdirectorsandexecutivecompensation.J.Financ.Quant.Anal.32,331–344.Harris,M.,Raviv,A.,2008.Atheoryofboardcontrolandsize.Rev.Financ.Stud.21,1797–1832.

Hartzell,J.,Starks,L.,2003.Institutionalinvestorsandexecutivecompensation.J.Financ.58,2351–2374.

Healey,P.,Wahlen,J.,1999.Areviewoftheearningsmanagementliteratureanditsimplicationsforstandardsetting.Account.Horiz.13,365–383.Hermalin,E.,Weisbach,M.,1998.EndogenouslychosenboardsofdirectorsandtheirmonitoringoftheCEO.Am.Econ.Rev.88,96–118.

Hermalin,E.,Weisbach,M.,2003.Boardsofdirectorsasanendogenouslydeterminedinstitution:asurveyoftheeconomicliterature.EconomicPolicy

Review.FederalReserveBankofNewYork7–26.

Hribar,P.,Nichols,D.,2007.Theuseofunsignedearningsqualitymeasuresintestsofearningsmanagement.J.Account.Res.45,1017–1053.

Huber,P.,1967.Thebehaviourofmaximumlikelihoodestimatesundernon-standardconditions.ProceedingsoftheFifthBerkeleySymposiumonMathematical

StatisticsandProbability,1.UniversityofCaliforniaPress,Berkeley,CA,pp.221–233.Hwang,B.,Kim,S.,2009.Itpaystohavefriends.J.Financ.Econ.93,138–158.

Jackson,S.,Lopez,T.,Reitenga,A.,2011.AccountingfundamentalsandCEObonuscompensation.J.Account.PublicPolicy27,374–393.Jones,J.,1991.Earningsmanagementduringimportreliefinvestigations.J.Account.Res.29,193–228.

Kim,K.,Kitsabunnarat-Chatjuthamard,P.,Nofsinger,J.,2007.Largeshareholders,boardindependence,andminorityshareholderrights:evidencefromEurope.

J.Corp.Finance13,859–880.

Klein,A.,2002.Auditcommittee,boardofdirectorcharacteristics,andearningsmanagement.J.Account.Econ.33,375–400.

Knechel,W.,Willekens,M.,2006.Theroleofriskmanagementandgovernanceindeterminingauditdemand.J.Bus.Finan.Acc.33,1344–1367.

Kumarand,P.,Sivaramakrishnan,K.,2008.Whomonitorsthemonitor?Theeffectofboardindependenceonexecutivecompensationandfirmvalue.Rev.Financ.

Stud.21,1371–1401.

Lambert,R.,Larcker,D.,1987.Ananalysisoftheuseofaccountingandmarketmeasuresofperformanceinexecutivescompensationcontracts.J.Account.Res.25,

85–125.

Leuz,C.,Nanda,D.,Wysocki,P.,2003.Earningsmanagementandinvestorprotection:aninternationalcomparison.J.Financ.Econ.69,505–527.Liang,K.Y.,Zeger,S.L.,1986.Longitudinaldataanalysisusinggeneralizedlinearmodels.Biometrika73,13–22.Linck,J.,Netter,J.,Yang,T.,2008.Thedeterminantsofboardstructure.J.Financ.Econ.87,308–328.

Linck,J.,Netter,J.,Yang,T.,2009.TheeffectsandunintendedconsequencesoftheSarbanes–Oxleyactonthesupplyanddemandfordirectors.Rev.Financ.Stud.

22,3287–3328.

Liu,Y.,2013.OutsideoptionsandCEOturnover:thenetworkeffect.J.Corp.FinanceXX,XXX-XXX(thisspecialissue).

Liu,Y.,Wei,Z.,Xie,F.,2013.DowomendirectorsimprovefirmperformanceinChina?J.Corp.FinanceXX,XXX-XXX(thisspecialissue).Mooney,C.Z.,Duval,R.D.,1993.Bootstrapping:ANonparametricApproachtoStatisticalInference.Sage,NewburyPark,CA.Nguyen,B.,Nielsen,K.,2010.Thevalueofindependentdirectors:evidencefromsuddendeaths.J.Financ.Econ.98,550–567.Petersen,M.,2009.Estimatingstandarderrorsinfinancepaneldatasets:comparingapproaches.Rev.Financ.Stud.22,435–480.Raheja,C.,2005.Determinantsofboardsizeandcomposition:atheoryofcorporateboards.J.Financ.Quant.Anal.40,283–306.

Sánchez-Ballesta,J.,García-Meca,E.,2005.AuditqualificationsandcorporategovernanceinSpanishlistedfirms.Manag.Audit.J.20,725–738.Santella,P.,Drago,C.,Paone,G.,2006.Howindependentareindependentdirectors?ThecaseofItaly.Mimeo.March2006.Santella,P.,Drago,C.,Paone,G.,2007.WhoCaresAboutDirectorIndependence?Mimeo.March2007.

Shivdasani,A.,Yermack,D.,1999.CEOinvolvementintheselectionofnewboardmembers:anempiricalanalysis.J.Financ.54,1829–1853.

Sun,J.,Cahan,S.,Emanuel,D.,2009.Compensationcommitteegovernancequality,chiefexecutiveofficerstockoptiongrants,andfuturefirmperformance.

J.Bank.Financ.33,1507–1519.

Weir,C.,Laing,D.,2000.Theperformance–governancerelationship:theeffectsofCadburycomplianceonUKquotedcompanies.J.Manag.Gov.4,265–281.White,H.,1980.Aheteroskedasticity-consistentcovariancematrixestimatorandadirecttestforheteroskedasticity.Econometrica48,817–830.White,H.,1982.Maximumlikelihoodestimationofmisspecifiedmodels.Econometrica50,1–25.

Xie,B.,DavidsonIII,W.,DaDalt,P.,2003.Earningsmanagementandcorporategovernance:theroleoftheboardandtheauditcommittee.J.Corp.Finance9,

295–316.

Yermack,D.,1996.Highermarketvaluationofcompanieswithasmallboardofdirectors.J.Financ.Econ.40,185–211.Zeger,S.,Liang,K.,1986.Longitudinaldataanalysisfordiscreteandcontinuousoutcomes.Biometrics42,121–130.

Pleasecitethisarticleas:Crespí-Cladera,R.,Pascual-Fuster,B.,Doestheindependenceofindependentdirectorsmatter?,J.Corp.Finance(2013),http://dx.doi.org/10.1016/j.jcorpﬁn.2013.12.009

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文

全部栏目

Does the independence of independent directors matter