(重点文献)POWERINTOPMANAGEMENTTEAMSDIMENSIONS,

®Academy0/Management/ournaJ1992.Vol.35,No.3.505-538.POWERINTOPMANAGEMENTTEAMS:DIMENSIONS,MEASUREMENT,ANDVALIDATIONSYDNEYFINKELSTEINUniversityofSouthernCaliforniaTopmanagers'powerplaysakeyroleinstrategicdecisionmaking.!However,althoughnumerousscholarshaverecognizeditsimportance,'veryfewhaveattemptedtomeasurethephenomenon.Inthisarticle,Ipresentasetofdimensionsmeasuringtopmanagers'powerandsuggest[ameasurementmethodologytofacilitateempiricalinquiry.Datafromiagroupof1,763topmanagersinthreeindustrieswereusedtoassess;thevalidityandreliahilityofthepowerdimensionsinthreestudies.|Resultsdemonstratestrongsupportfortheproposedpowerdimen-sions.Thetopicofexecutiveleadershiphasrecentlyreceivedsignificantat-tentionfromscholarsinstrategyandorganizationtheory.Theyhavecon-centratedonsuchissuesasthecompositionoftopmanagementteams,ex-ecutivesuccession,managerialstyles,board-managementrelations,andfit-tingexecutiveteamstoenvironmentsandstrategies.However,oneimportantareathathasreceivedlittleattention(EisenhardtandBourgeois[1988]isanexception)istopmanagers'power.Thislackissurprising,giventheimportanceofpowerrelationshipstostrategicchoice.Powerisequallycentraltoresearchontopmanagementteams.Infact,thechoiceofunitofanalysisinresearchontopmanagersandtheissueofmanagerialpoweraretwosidesofthesamecoin.Thatis,adoptionofaunitofanalysisrestsonanimplicitassumptionaboutthedistributionofpoweramongtopmanagers.Forexample,inanorganizationinwhichthechiefexecutiveofficer(CEO)wieldsdominantpower,studyingonlytheCEOmayprovidesufficientinformationwithwhichtotestpropositions.However,inorganizationsinwhichpowerislesspolarized,considerationofacoalitionoftopmanagersisnecessarytofullycapturetherangeofmanagerialorien-tationsprevailing.Hence,considerationofthedistributionofpoweramongtopmanagersseemsanessentialingredientforresearchontopmanagementteams.IIwouldliketothankRichardD'Aveni,JamesFredrickson,DonaldHambrick,SamHari-haran,IanMacMillan,MichaelTushman,BorisYavitz,andtheanonymousreview'ersofthisjournalfortheirvaluablecommentsonearlierdraftsofthispaper.ThegenerousresearchsupportoftheFacultyResearchandInnovationFundoftheUniversityofSouthernCalifornia,theExecutiveLeadershipResearchCenteratColumbiaUniversity,theIrwinFoundation,andtheSocialSciencesandHumanitiesResearchCouncilofCanadaaregratefullyacknowledged.505506AcademyofManagementJournalAugustInthisarticle,Ifocusonthemostsenioroftopmanagers,thedominantcoalitionsoffirms(Cyert&March,1963).Althoughmostlargefirmshavemanyofficers,typicallyonlyasmallsubsetofmanagersismostresponsibleforsettingpolicy(Thompson,1967).Itisthisinnercircle,ordominantcoalition,thatwasthefocusofthisresearch.ThedominantcoalitionofafirmtypicallyconsistsoftheCEOandseveralofhisorhermostseniormanagers.However,althoughtheCEOisusuallythemostpowerfulmemberofthisgroup,suchisnotalwaysthecase(Mintzberg,1983).Forexample,managerswithlargeshareholdingsmaybemorepowerfulthanaCEO.Exceptinthemostextremecases,managementisasharedeffortinwhichadominantcoalitioncollectivelyshapesorgan-izationaloutcomes.ThelimitedempiricalresearchcomparingexplainedvarianceusingCEOsorawidergroupoftopmanagershasconsistentlyfoundthatthelatterunitofanalysisyieldedsuperiorresults(Bantei&Jack-son,1989;Finkelstein,1988;Hage&Dewar,1973;Tushman,Virany,&Ro-manelli,1985).ThesefindingssupportthenotionthatCEOssharepowerwithotherseniorexecutivesinmanyfirms.Hence,tomorefullyunderstandhowtopmanagersinfluenceorganizationaldirection,itisimportanttodif-ferentiatemanagersintermsoftheirpower.Theresearchreportedherehadthreeprimarypurposes:(1)toarguethatmanagerialpowerisacentralelementinstrategicchoice,(2)toconceptu-alizemajorpowersourcesindominantcoalitions,and(3)tosuggestandvalidatespecificmeasuresofpowerthatarereadilyavailabletootherre-searchers.Thefollowingsectionexaminespowerandstrategy,illustratinghowtheyinteractandhence,whystudyingmanagerialpowerissoimpor-tantforresearchersinstrategicmanagement.Next,Ipresentbothasetoftopmanagerialpowerdimensionsandameasurementmethodologytoaidre-searchers,andfinally,reportthreestudiesthattestthevalidityofthesedimensions.POWERINSTRATEGICCHOICEPowerisdefinedhereasthecapacityofindividualactorstoexerttheirwill.Thisdefinitionisconsistentwiththoseofotherscholars(Hickson,Lee,Schneck,&Pennings,1971;MacMillan,1978;Pfeffer,1981)andreadilylendsitselftoananalysisofpoweramongtopmanagersinorganizations.Althoughpowermaybeexercisedinnumeroussettings(Pfeffer,1981),thisarticleconcentratesonitsroleinstrategymaking.Child(1972)recognizedthatpoweriscentraltostrategicchoice.Herecommendedthatinvestigatorsstudypowertounderstandwhatstrategicchoicesaremade.Bysodoing,theycanmakeconfidentpredictionsabouttheimpactofmanagerialorientationsonstrategy.AsChildargued,onlywhenpowercanbeadequatelymeasuredishighpredictivecertaintylikelytobeachieved.Thereisconsiderablesupportforthisviewfromotherscholars.Strate-gicdecisionsareunstructuredandrepletewithambiguities(Mintzberg,1992Finkelstein507Raisinghani,&Theoret,1976).Hence,theyinvitetheuseofpower(Mintz-berg,1983),withdifferentexecutivesfavoringtheirpr