Collaborative Governance in Theory and Practice

JPART18:543–571CollaborativeGovernanceinTheoryandPracticeChrisAnsellAlisonGashUniversityofCalifornia,BerkeleyABSTRACTOverthepastfewdecades,anewformofgovernancehasemergedtoreplorativegovernance,asithascometobeknown,bringspublicandprivatestakeholderstogetherincollectiveforarticle,weconductameta-analyticalstudyoftheexistingliteratureoncollaborativegovernancewiteview-ing137casesofcollaborativegovernanceacrossarangeofpolicysectors,weidentifycriticalvariablesthatwillinfluencewheariablesincludethepriorhistoryofconflictorcooperation,theincentivesforstakeholderstoparticipate,powerandresourcesimbalances,leadership,identifyasactorsincludeface-to-facedialogue,trustbuilding,dthatavirtuouscycleofcollaborationtendstodevelopwhencollaborativeforumsfocuson‘‘smallwins’’thatdeepentrust,commitment,icleconcludeswithadiscus-sionoftheimplicationsofourcontingencymodelasttwodecades,anewstrategyofgoverningcalled‘‘collaborativegovernance’’deofgovernancebringsmultiplestakeholderstogetherincommonforarticle,weconductameta-analyticalstudyoftheexistingliteratureoncollaborativegovernancimategoalistodevelopacontingencyapproachtocollaborationthatcanhighlightconditionsunderwhichcollaborativegovernancewillbemoreorlesseffectiveasanEarlyversionsofthisarticlewerepresentedattheConferenceonDemocraticNetworkGovernance,Copenhagen,theInterdisciplinaryCommitteeonOrganizationsattheUniversityofCalifornia,Irvine,ktheparticipantsoftheseeventsfortheirusefulsuggestionsandMarthaFeldman,inparticular,scorrespondencetotheauthoratcansell@ligash@:10.1093/jopart/mum032AdvanceAccesspublicationonNovember13,2007ªhedbyOxfordUniversityPressonbehalfoftheJournalofPublicAdministrationResearchandTheory,missions,pleasee-mail:sions@

544JournalofPublicAdministrationResearchandTheoryapproachtopolicymakingandpublicmanagement.1Inconductingthismeta-analyticstudy,weadoptedastrategywecall‘‘successiveapproximation’’:weusedasampleoftheliteraturetodevelopacommonlanguageforanalyzingcollaborativegovernanceandthensuccessively‘‘tested’’thislanguageagainstadditionalcases,refiningandelaboratingourmodelofcollaborativegovernanceasweevaluatedadditionalcases.´,Althoughcollaborativegovernancemaynowhaveafashionablemanagementcachetheuntidycharacteroftheliteratureoncollaborationreflectsthewayithasbubbledupfrommanylocalexperiments,o-rativegovernancehasemergedasaresponsetothefailuresofdownstevelopedasanalternativetotheadversarialismofinterestgrouppluralismandtotheaccountabilityfailuresofmanageri-alism(especiallyastheauthorityofexpertsischallenged).Morepositively,onemightarguethattrendstowardcollaborledgebecomesincreasinglyspecializedanddistributedandasinstitutionalinfrastructuresbecomemorecomplexandinterdependent,monmetricforallthesefactorsmaybe,asGray(1989)haspointedout,theincreasing‘‘turbulence’’ghSusskindandCruikshank(1987),Gray(1989),andFungandWright(2001,2003)havesuggestedmoregeneraltheoreticalaccountsofcollaborativegovernance,kofthecollaborativegovernanceliteratureiscomposedofsingle-casecasestudiesfocusedonsector-specificgovernanceissueslikesite-basedmanagementofschools,communitypo-licing,watershedcouncils,regulatorynegotiation,collaborativeplanning,communityhealthpartnerships,andnaturalresourcecomanagement(thespecies).2Moreover,anum-berofthemostinfluentialtheoreticalaccountsofthisphenomenonarefocusedonspecifi(1996,2003)andInnesandBooher(1999a,1999b),forexample,providefoundationalaccountsofcollaborativeplanning,asFreeman(1997)doesforregulationandadministrativelawandWondolleckandYaffee(2000)listobuildonthefindingsofthisrichliterature,butalsotoderivetheoreticalandempiricalclaimsaboutthegenusofcollaborativegovernance—NGCOLLABORATIVEGOVERNANCEWedefinecollaborativegovernanceasfollows:Agoverningarrangementwhereoneormorepublicagenciesdirectlyengagenon-statestakeholdersinacollectivedecision-makingprocessthatisformal,consensus-oriented,anddeliberativeandthatfinitionstressessiximportantcriteria:(1)theforumisinitiatedbypublicagenciesorinstitutions,(2)participantsintheforumincludenonstateactors,(3)participantsengagedirectlyindecisionmakingandarenotmerely‘‘consulted’’bypublicagencies,(4)the1Thomas(1995)developsacontingencyperspectiveonpublicparticipation,thoughitaimsmorebroadlyandisdevelopedfromtheperspectiveofpublicmanagers.2Asmallergroupofstudiesevaluatesspecifictypesofcollaborativegovernanceatamoreaggregatedlevel(forexample,seeBeierle[2000],Langbein[2002],andLeach,Pelkey,andSabatier[2002]).

AnsellandGashCollaborativeGovernanceinTheoryandPractice545forumisformallyorganizedandmeetscollectively,(5)theforumaimstomakedecisionsbyconsensus(evenifconsensusisnotachievedinpractice),and(6)amorerestrictivedefir,thewide-ranginguseofthetermhas,asImperialnotes,beenabarriertotheorybuilding(Imperial2005,286).Sinceourgoalistocompareappleswithapples(totheextentpossible),wehavedefiticalcomponentofthetermcollaborativegovernanceis‘‘governance.’’Muchresearchhasbeendevotedtoestablishingaworkabledefinitionofgovernancethatisboundedandfalsifiable,tance,Lynn,Heinrich,andHill(2001,7)construegovernancebroadlyas‘‘regimesoflaws,rules,judicialdecisions,andadministrativepracticesthatconstrain,prescribe,andenabletheprovisionofpubliclysupportedgoodsandservices.’’Thisdefinitionprovidesroomfortraditionalgovernmentalstructuresaswellasemergingformsofpublic/,ontheotherhand,argues:Asabaselinedefinitionitcanbetakenthatgoverefocusisondecision-makinginthecollectiveimpliesthatgovernanceisnotaboutoneindividualmakingadecisionbutratheraboutgroupsofindividualsororganisationsorsystemsoforganisationsmakingdecisions(2004,3).Healsosuggeststhatamongthevariousinterpretationsoftheterm,thereis‘‘baselineagreementthatgovernancereferstothedevelopmentofgoverningstylesinwhichbound-ariesbetweenandwithinpublicandprivatesectorshavebecomeblurred’’(Stoker1998,17).ewithLynn,Heinrich,andHillthatgovernr,weadoptStoker’sclaimthatgovernanceisalsoaboutcollectivedecisionmaking—andspecificallyabouorativegovernanceisthereforeatypeofgovernanceinwhichpublicandprivateactorsworkcollectivelyindistinctiveways,usingparticularprocesses,ghtherearemanyformsofcollaborationinvolvingstrictlynonstateactors,ourdefinitionstipulatesaspecifigtheterm‘‘publicagency,’’ourintentionistoincludepublicinstitutionssuchasbureaucracies,courts,legislatures,andothergovernmentalbodiesatthelocal,state,typicalpublicin-stitutionamongourcasesis,infact,anexecutivebranchagency,andtherefore,theterm‘‘publicagency’’blicagenciesmayinitiatecollaborativeforumseithertofulfilltheirownpurposesortocomplywithamandate,includingcourtorders,legislation,mple,theWorkforceInvestmentActof1998stipulatesthatallstatesandlocalitiesreceivingfederalworkforcedevelop-mentfundsmustconveneaworkforceinvestmentboardthatcomprisedpublicandprivateactorsinordertodevelopandoverseepoliciesatthestateandlocallevelconcerningjobtraining,ingtoourdefinition,theseworkforcghpublicagenciesaretypicallytheinitiatorsorinstigatorsofcollaborativegovernance,ourdefighthereisnothinginherentlywrongwithusingtheterminthisway,muchoftheliteratureoncollaborative

546JournalofPublicAdministrationResearchandTheorygovernanceusesthistermtosignaladifferen(1998,61),forexample,arguesthatcollaborativesin-volve‘‘representationbykeyinterestgroups.’’ConnickandInnes(2003,180)definecollaborativegovernanceasincluding‘‘representativesofallrelevantinterests.’’Reilly(1998,115)describescollaborativeeffortsasatypeofproblemsolvingthatinvolvesthe‘‘sharedpursuitofgovernmentagenciesandconcernedcitizens.’’Weusetheterm‘‘stakeholder’’toreferbothtotheparticipatiovenience,wewillalsohere-afterusetheterm‘‘stakeholder’’torefertobothpublicagenciesandnonstatestakeholders,thoughwebelievethatpublicfinitionofcollaborativegovernanceaevethatcollaborativegovernanceisnevermerelyconsultative.3Collaborationimpliestwo-waycommunicationandinfluencebe-tweenagenciesandstaesandstrwords,asdescribedabove,tativetechniques,suchasstakeholdersurveysorfocusgroups,althoughpossiblyveryusefulmanagementtools,arenotcollaborativeinthesenseimpliedherebecausetheydonotpermittwo-wayflorationalsoimpliesthatore,weimposethn(1997,22),forexample,arguesthatstakeholdersparticipate‘‘inallstagesofthedecisionmakingprocess.’’ThewatershedpartnershipsstudiedbyLeach,Pelkey,andSabatier(2002,648)makepolicyandimplementationdecisionsonarangeofongoingwatermanagementissuesregardingstreams,rivers,teauthoritymayliewiththepublicagency(aswithregulatorynegotiation),bu,advisorycommitteesmaybeaformofcollaborativetice(andbydesign),however,sethecriteriaofformalcollaborationtodistinguishcollaborativegover-nancefrommple,thetermcollaborativegovernancemightbethoughttodescribetheinforma,interestgroupsandpublicagencieshavealwaysengagedintwo-wayflowsofinflferencebetweenourdefinitionofcollaborativegovernanceandconventionalinterestgroupinfluenceisthattheformerimpliesanexplicitandpublicstrategyoforganizingthisinflandPetr(2000,495),forexample,describecollaborativegovernanceasaformalactivitythat‘‘involvesjointactivities,jointstructuresandsharedresources,’’andPadillaandDaigle(1998,74)prescribethedevelopmentofa‘‘structuredarrangement.’’onsincollaborativeforumsareconsensusoriented(ConnickandInnes2003;Seidenfeld2000).Althoughpublicagenciesmayhavetheultimateauthoritytomake3SeeBeierleandLong(1999)foranexampleofcollaborationasconsultation.

AnsellandGashCollaborativeGovernanceinTheoryandPractice547adecision,thegoalofcollaborhetermconsensusorienter,thepremiseofmeetingtogetherinadeliberative,multilateral,andformalforumistostrivetowardconsensusor,atleast,y,usonpublicissuesdistinguishescollaborativegovernancefromotherformsofconsensusdecisionmaking,ghagenciesmaypursuedisputeresolutionormediationtoreducesocialorpoliticalconflict,thesetechniquesareoftenusedtodealwithstrictlyprivateconfler,publicdisputercknowledgingtheambiguityoftheboundarybetweenpublicandprivate,werestricttheuseoftheterm‘‘collaborativegovernance’’finitionofcollaborativegovernanceismeanttodistinguishcollaborativegov-ernancefromtwoalternativepatternsofpolicymaking:adversarialismandmanagerialism(Busenberg1999;Futrell2003;WilliamsandMatheny1995).Bycontrastwithdecisionsmadeadversarially,collaborativegovernanceisnota‘‘winner-take-all’’aborativegovernance,stakeholderswilloftenhaveanadversarialrelationshiptooneanother,butthegrsarialpolitics,groupsr,thiscooperationisadhoc,andadversarialpoliticsdoesnotexplicitlyseektotransformconflgerialism,publicagenciesmakedecisionsunilaterallyorthroughclosedde-cisionprocesses,typicallyrelyingonagencyexpertstomakedecisions(Futrell2003;WilliamsandMatheny1995).Althoughmanagerialagenciesmaytakeaccountofstake-holderperspectivesintheirdecisionmakingandmayevengosofarastoconsultdirectlywithstakeholders,collaborativegovernancerequir-ample,‘‘corporatism’’iscertainlyaformofcollaborativegovernanceaswedeficdefinitionsofcorporatism(likeSchmitter’s)emphasizetripartitlly,thesepeakassocia-tionshavearepresentationalmonopolyintheirsector(theyare‘‘encompassing’’).Ifwestartwiththisnarrowerdefinitionofcorporatism,orativegovernanceoftenimpliestheinclusionofabroaderrangeofstake-holdersthancorporatism,andtm‘‘associationalgovernance’’issometimesusedtorefertothemoregenericmodeofgoverningwithassociations,teAlegreproject,forexample,isaformofcollabo-rativegovernancethatincludesindividualcitizensinbudgetarydecisionmaking(FungandWright2001).Sometimestheterm‘‘policynetwork’’er,policynetworkstypicallyimplycooperative,thetermspor,collaborativegovernancereferstoanexplicitandformalstrategyofincorporatingstakeholdersinto

548JournalofPublicAdministrationResearchandrast,theco-operati,unacknowledged,unstated,nondesigned).Moreover,itmayoperatethroughinfor-malpatternsofbrokerageorativegovernanceandpu-privatepartnershipstypicallyrequirecollaborationtofunc-tion,buttheirgoalisoftentoachic-privatepartnershipmaysimplyrepresentanagreementbetweenpubltivedecisionmakingisthereforesecondarytothedefirast,theinstitutionalizationofacollectivedecision-makingprocessiscentraltothedefiy,arangrmsincludeparticipatorymanagement,interactivepolicymaking,stake-holdergovernance,erthetermgovernancetomanagementbecauseitisbroaderandencompassesvariousaspectsofthegoverningprocess,includingplanning,policymaking,mcollaborativeisalsomoreindicativeofthedeliberativeandconsensus-orientedapproachthatwecontrastwithadverOFCOLLABORATIVEGOVERNANCEArmedwithaworkingdefinitionofcollaborativegovernance,hisinthetypicalfashion:wesystematicallyreviewedjournalsacrossawiderangeofdisciplines,includingspecialistjournalsinpublichealth,education,socialwelfare,internationalrelations,conductedkeywordelectronicsearchesusingawidevarietyofsearchterms,,‘‘comanagement,’’‘‘publicparticipation,’’‘‘alternativedisputeres-olution’’).Ofcourse,tely,ghinternationalinscope,oursearchwasrestrictedtoliteratureinEnglish,andthus,ursoryexaminationofourcasesalnotduetoanysamplingbiasonourpartbutratherreflectstheimportanceofcolthestudieswereviewedwerecasestudiesoightimagine,theuniverseofcaseswecollectedwasquitediverseandthecasesdifferedinquality,methodology,ghourdefinitionwasrestrictivesoastofacilitatecomparisonofappleswithapples,eivedexperimentswithcollaborativegovernancebubblingupinmanydifferentpolicysectors,,wefelt,ginalintentiontotreatthesecasesasalarge-Ndatasetstisusefulforbothscholars

AnsellandGashCollaborativeGovernanceinTheoryandPractice549andpractitionerstounderstandhowwearrivedatourconclusions,webrieflttemptsatsystematiccodingwerefrustrating,ghscholarsstudyingcollaborativegovernancehadalreadymadesomeimportanttheoreticalstatements,dourselvesgropingtofindacommonlanguageofdescriptionandevaluationevenasweweretryingto‘‘code’’hischallengeasevereproblemof‘‘missingdata’’—areflectionofthehighlyvariedmotivationsoftheresearchers—tely,wemovctedasubsetofourcasesandusedthemtodevelopacommon‘‘model’’ondsubsetwasusedto‘‘test’’themodeldevelopedinthefirstroundandthentofurther‘‘refine’’sampleofcaseswasusedtotestthesecond-roundmodel,endixprovidesalisiveapproximationhastheadvantageofbothrefiningtheconceptualmodelwhileprovidingsomeoftheevaluative‘‘discipline’’-ever,weareundernoillusionthatthisprocessyielded‘‘theone’’asoceeded,lesandcausalrelationshipsproliferatedbeyondwhore,ourmodelrepresentsaconsciousattempttosimplifyasmucalofsimplificationledustostresscommonandfrequentfiproachstrengthensthegeneralityofourfindingsbutdis-countslessuniversalorfrequentlymentionedfitheendofouranalysis,thefierimportantclarificationneedstobemadebeforeweintroduceourfiveyofthecasesquicklydisabusedusofthenotionthatwecoulduseouranalysistoanswerthequestion:‘‘Iscollaborativegovernancemoreeffectivethanadversarialormanagerialgovernance?’’Venottosaythatthecomparisonbetweencollaborative,adversarial,mentswithcollaborativegovernanceweretypicaststudiesdidtrytodowasunyengageingoodfaithnegotiation?Didtheypursuemutualgains?Didtheyachieveconsensus?Weretheysatisfiedwiththeprocess?Inotherwords,moststudiesinthecollaborativegovernanceliteratureevaluate‘‘processoutcomes’’1providesavisualrepresentationofourcentralfielhasfourbroadvariables—startingconditions,institutionaldesign,leadership,andcollaborativeToavoidrecreatingthewheel,ourfirstsubsetwasnotrandomlyselectedbutincludedmanyofthemostprominenttheoreticalstatementsaboutcollaborativegovernance.4

550JournalofPublicAdministrationResearchandTheoryFigure1AModelofCollaborativeGovernanceParticipatory Inclusiveness,Forum Exclusiveness, ClearGround Rules, ProcessTransparencyInstitutional DesignStarting ConditionsPower-Resource-KnowledgeAsymmetriesCollaborative ProcessTrust-BuildingIncentives for andConstraints onParticipationFace-to-Face Dialogue-Good Faith NegotiationCommitment to Process-Mutual recognition ofinterdependence-Shared Ownership ofProcess-Openness to ExploringMutual Gains

Shared Understanding-Clear Mission

-Common ProblemDefinition-Identification ofCommon ValuesOutcomesPrehistory ofCooperation orConflict (initialtrust level)InfluencesIntermediate Outcomes-“Small Wins”-Strategic Plans-Joint Fact-FindingFacilitative Leadership(including empowerment)thesebroadvariablescanbedisaggregatedintomorefiorativeprocessvariablesaretreatedasthecoreofourmodel,withstartingconditions,institutionaldesign,andleadershipvariablesrepresentedasengconditionssetthebasicleveloftrust,conflict,andsoutiona,leadershipprovlaborativeprocessitselfishighlyiterativeandnonlinear,andthus,werepresentit(withconsiderablesimplification)ainderofthearticledescribeseachofthesevariablesinmoredetailanddrawsouNGCONDITIONSTheliteratureisclearthatconditionspresentattheoutsetofcollaborationcaneitherfacilitateordiscourag,thestakeholdershaveahistoryofbitterdivisionoversomeemotionallychather,thestakeholdershaveasharedvisionforwhattheywouldliketoachievetcases,collaborationmaybedifficult,butthefirstcasemustover-comeproblemsofdistrust,disrespect,owedthecriticalstartingconditionsdowntothreebroadvariables:imbalancesbetweentheresourcesor

AnsellandGashCollaborativeGovernanceinTheoryandPractice551powerofdifferentstakeholders,theincentivesthatstakeholdershavetocollaborate,andthepasthistoryofconfl/ResourceImbalancesPowerimbalancesbetweenstakeholdersareacommonlynotedproblemincollaborativegovernance(Gray1989;ShortandWinter1999;SusskindandCruikshank1987;Tett,Crowther,andO’Hara2003;Warner2006).Ifsomestakeholdersdonothavethecapacity,organization,status,orresourcestoparticipate,ortoparticipateonanequalfootingwithotherstakeholders,thecollabormple,Bradford(1998)demonstratesthatattemptsbytheGovern-mentofOntariotomakejobtrainingandoccupationalhealthandsafetypolicythroughcollaborativemeanswerethwartedbytheprivilegedstatusoffirmswho,through‘‘informalchannels,’’wereabletogainaccesstosenioroffitely,suchimbal-ancesproducedistrustorweakcommitment(Gray1989,119;Warner2006).Americanenvironmentalgroupsarenotablyskepticalaboutcollaborativegovernancebecausetheyfeelthatitisadvantageoustoindustrygroups(McCloskey2000).Echeverria(2001),forexample,criticizesthePlatteRiverCollaborativeWatershedPlanningProcessbecausehearguesthattesthatdevelopmentetheirconstituencyissolargeanddiffuse,conservationadvocatesareroutinelyatadisadvantageincontestswithrepresentativeststrongcountermeasurestorepre-sentlesspowerfulvoicesandwithout‘‘neutral’’agencyleadership,Schuckman(2001)arblemofpowerimbalancesisparticularlyproblematicwhereimportantstake-holdersdonothavetheorganizationalh(2000),forexample,arguesthatthemorediffusetheaffectedstakeholders,andthemorelongtermtheproblemhorizon,themoredifficases,theproblemisthatorganizedstakeholdergroupsdonotexisttorepresentindividualstakeholderscollec-tively(Buanesetal.2004;Rogersetal.1993).Anothercommonproblemisthatsomestakeholdersmaynothavetheskillandexpertisetoengageindiscussionsabouthighlytechnicalproblems(GuntonandDay2003;LaskerandWeiss2003;Merkhofer,Conway,andAnderson1997;Murdock,Wiessner,andSexton2005;Warner2006).Athirdcom-monproblemisthatsomestakeholdersdonothavethetime,energy,orlibertytoengageintime-intensivecollaborativeprocesses(YaffeeandWondolleck2003).entsofcollaborationhavepointedtoarangeofstrategiesthatcanbeusedtoempowerweakerorunderrepresentedgroups(Fawcettetal.1995;LaskerandWeiss2003;Merkhofer,Conway,andAnderson1997;Mitchell2005;Schuckman2001).5Intermsofacontingencytheoryofcollaborativegovernance,wedrawthefollowingconclusion:(1)Iftherearesignificantpower/resourceimbalancesbetweenstakeholders,suchthatimportantstakeholderscannotparticipateinameaningfulway,theneffectiveScholarsofregulationworry,however,thatempowermentmightleadtoagencyco-optationofstakeholders(Seidenfeld2000).5

552JournalofPublicAdministrationResearchandTheorycollaborativegovernancerequiresacommitmenttoapositivestrategyofivestoParticipateGiventhelargelyvoluntarynatureofparticipation,itiscriticaltounderstandtheincentivesthatstakeholdershavetoengageincollaborativegovernanceandthefactorsthatshapethoseincentives(Andranovich1995;ChrislipandLarson1994;Gray1989;NelsonandWeschler1998;SusskindandCruikshank1987).Thisincludesanalysisomple,Ebrahim(2004)comparesthedifferentincentivesIndianforestandirrigationagenciesfaceandshowshowpositivefinancialrsofcollaborativegovernancehaverecognizedthatpowerandresourceimbal-anceswillaffecttheincentivesofgroupstoparticipateincollaborativeprocesses(GuntonandDay2003;Imperial2005).Gray(1989)arguesthatpowerdifferencesamongplayersinflnmentalistspreferthetraditionalcongressionalhearingprocess,shepointsout,ore,shearguesthattimingconsiderationswillbeimportant:partiesthatbelievethattheirpowerivestoparticipatedependinpartuponstakeholderexpectationsaboutwhetherthecollaborativeprocesseswillyieldmeaningfulresults,particularlyagainstthebalanceoftimeandenergythatcollaborationrequires(Bradford1998;GeogheganandRenard2002;Rogersetal.1993;Schneideretal.2003;Warner2006).Incentivesincreaseasstakeholdersseeadirectrelationshipbetweentheirparticipationandconcrete,tangible,effectualpolicyoutcomes(Brown2002).Buttheydeclineifstakeholdersperceivetheirowninputtobemerelyadvisoryorlargelyceremonial(Futrell2003).Althoughcollaborativeapproachesmaybemandatedbycourtsorlegislatures,uently,theincentivesthatstakeholdershavetoenterintocollaborationwillloomlargeasafivestoparticipatearelowwhenstakeholderswhoviewthemselvesashavingstrongalliesinthecourtsorinlegis-latures,forexample,suchstakeholderstentativelydecidetoengageinthecollaborativeprocess,theymaytaketheirclaimstoanalternativevenueiftheybecomedisgruntledwiththeprocessoritsoutcomes(KhademianandWeber1997).Conversely,theincentiveforstakeholderstoparticipateislikelytoincreaseevadaturtlecase,describedbyReilly(2001,133),successfulcollaborationensuedafterthecourtrefusedtoivestoparticipateincollaborativegovernancewillalsoincreaseifstakeholdersperceiveachievementoftheirgoalstobedependentoncooperationfromotherstake-holders(Logsdon1991).Forexample,theprevalenceofcollaborativegovernanceinlocalresourcemanagementdisputesisprobablyrelatedtothejointdependenceoflocalgroupsonacommonresource(HeikkilaandGerlak2005).The,highlyantagonisticstakeholders

AnsellandGashCollaborativeGovernanceinTheoryandPractice553whoarealsohighlydependentuponeachothermaymovetowardasuccessfulcollabora-tiveprocess(Imperial2005;YaffeeandWondolleck2003).Reilly(2001),forexample,describesthe‘‘balanceofterror’’thatkeptrivalstakeholatively,stakeholderswithadeepfoundationoftrustandsharedvaluesmayfailatcollaborationbecausestakeholdersfitionsofinterdependence,ofcourse,,incentivestoparticipateareoftenshapedbythe‘‘shadowofthestate,’’suchasthreatsofregulationorcourt(Bentrup2001;Brown2002;ShortandWinter1999).Intheareaofendangeredspeciesprotection,forexample,itiscommonforcollaborationtobeativevenueswillbeparticularlyattractivetoproposetwoadditionstoourcontingencymodel:(2)Ifalternativevenuesexistwherestakeholderscanpursuetheirgoalsunilaterally,thencollaborativegovernancewillonlyworkifstakeholdersperceivethemselvestobehighlyinterdependent.(3)Ifinterdependenceisconditionaluponthecollaborativeforumbeinganexclusivevenue,thensponsorsmustbewillingtodotheadvanceworkofgettingalternativeforums(courts,legislators,andexecutives)toryofAntagonismandCooperationTheliteratureindicatesthattheprehistoryofantagonismorcooperationbetweenstakeholderswillhinderorfacilitatecollaboration(Andranovich1995;Gray1989;um2002).However,wenotethatwhenstakeholdersarehighlyinterdependent,ahighlevelofconflicberofcases,policydeadlockscanactuallycreateastrongimpetusforcollaborativegovernance(Futrell2003).Suchsituationsoftenoccurinresourcemanage-mentcontextswescribestheoriginsofalocalcollaborativeasfollows:‘‘Exhaustedandfrustratedfromconstantbattlingoverthedispositionofnaturalresourcesandlandmanagementapproaches,BrownandSwenson[leadersofthetworivalgroups]decidedtositdownandseeiftherewasanalternative,moreamicablemethodforreconcilingtheirdifferences’’(Weber2003,59).Therefore,itisclearthathighconflofthesuccessfulcollaborationsdescribedintheliterature,stake-holdershavecometoseethattheycannotachievetheirgoalswithoutengaginginacollab-orativeprocessorativegovernance,however,oftenbuildsonahistoratorsarekeenlyaware,‘‘usversusthem’’storyofconflictislikelytoexpressitselfinlowlevelsoftrust,whichinturnwillproducelowlevelsofcommitment,strategiesofmanipulation,rwords,aprehistoryofconflictcreatesaviciouscircleofsuspicion,distrust,therhand,ahistoryofsuccessfulpastcooperationcancreatesocialcapi,therefore,suggestthefollowingcontingency:(4)Ifthereisaprehistoryofantagonismamongstakeholders,thencollaborativegovernanceisunlikelytosucceedunless(a)thereisahighdegreeofinterdependenceamongthe

554JournalofPublicAdministrationResearchandTheorystakeholdersor(b)positivestepsaretakentor,however,thatstrongtrustandinterdependenceamongsubsetsofstakeholdersmayworksurveyoftheBayAreaenvironmentalmovement,Ansell(2003)foundthatcliquesofTATIVELEADERSHIPLeadershipiswidelyseenasacriticalingredientinbringingpartiestothetableandforsteeringthemthroughtheroughpatchesofthecollaborativeprocess(Burgeretal.2001;ChrislipandLarson1994;Frame,Gunton,andDay2004;Gilliametal.2002;GuntonandDay2003;HeikkilaandGerlak2005;HuxhamandVangen2000;Imperial2005;LaskerandWeiss2003;Margerum2002;Murdock,Wiessner,andSexton2005;Reilly1998,2001;RoussosandFawcett2000;Saarikoski2000;Smith1998;VangenandHuxham2003a).Although‘‘unassisted’’negotiationsaresometimespossible,theliteratureover-whelminglyfindsthatfacilitativeleadershipisimportantforbringingstakeholderstogetherandgettingthemtoengageeachotherinacollaborativespirit(ChrislipandLarson1994;Ozawa1993;Pine,Warsh,andMaluccio1998;Reilly2001;SusskindandCruikshank1987).Indescribingthreeformsof‘‘assistednegotiation,’’SusskindandCruikshank(1987)suggestincreasinglymoreinterventionistmediationtechtationistheleastintrusiveonthemanagementprerogativesofstakeholders;afacilitator’ionincreasestheroleofthethirdpartyin-terventioninthesubstantivedetailsofthenegotiatiy,ifstakeholderscannotreachaconsensuswiththehelpofmediation,thethirdpartymaycraftasolution(nonbindingarbitration).VangenandHuxham(2003a)arguethattomovecollaborationforward,shipiscrucialforsettingandmaintainingcleargroundrules,buildingtrust,facilitatingdialogue,andHuxham(2003a)arguethatleadershipisimportantforembracing,empowering,andinvolipandLarson(1994,125)describethecollaborativeleaderasastewardoftheprocess(transforming,servant,orfacilitativeleadership)whoseleadershipstyleis‘‘...characterizedbyitsfocusonpromotingandsafeguardingtheprocess(ratherthanonindividualleaderstakingdecisiveaction).’’Schol-arsassertthatcollaborativegovernancerequiresspecifi(2001,241),forexample,identifiesthreecomponentsof‘‘effective’’collaborativeleadership:adequatemanagementofthecollaborativeprocess,maintaining‘‘technicalcredibility,’’andensuringthatthecollaborativeisempoweredto‘‘makecredibleandconvincingdeci-sionsthatareacceptabletoall.’’LaskerandWeiss(2001,31)arguethatcollaborativeleadersmusthavetheskillsto(1)promotebroadandactiveparticipation,(2)ensurebroad-basedinfluenceandcontrol,(3)facilitateproductivegroupdynamics,and(4)sfulcollaborationsmayalsousemultipleleaders,formallyandinformally,ratherthanrelyingononeleader(Bradford1998;LaskerandWeiss2003).HuxhamandVangen(2000)emphasizethateffectivecollaborativeleadershipislikelytobetime,resource,andskillintensive.

AnsellandGashCollaborativeGovernanceinTheoryandPractice555Leadersh(1993),forexample,describeswhathecalls‘‘transformative’’techniquesinwhichmediationprocedureshelpstobringabouta‘‘balanceofpower’’yleoffacilitativeleaandWeiss(2003,31–3)arguethatfacilitativeleadersmust‘‘givemeaningfulvoicetoparticipants’’sshouldstimulatecreativityby‘‘synthesiz[ing]theknowledgeofdiverseparticipantssothegroupcancreatenewideasandunderstanding.’’Whereincentivestoparticipateareweak,powerandresourcesareasymmetricallydistributed,andpriorantagonismsarehigh,ethatstake-holdersfundamentallydistrusteachother,r,whenincentivestoparticipateareweakorwhenpowerisasym-metrical,theleadermustofteentiontoempowerweakeractors,forexample,mayupsettheperceptionthattheleaderisanhonestbroker(Warner2006).Moreover,therearesomnflictishigh,theroleofhonestbrokerisoftengiventoanoutsiutsidemediatormayalsohavelittleinflvethefollowingconclusionsfromthislogic:(5)Whereconflictishighandtrustislow,butpowerdistributionisrelativelyequalandstakeholdershaveanincentivetoparticipate,thencollaborativegovernancecansuc-cessfullyproceedbyrelyingontheservhonestbrokerwillalsobeabletodeveloptrustduringthecollaborativeprocessbyremainingabovethefrayandbymaintaininr,(6)Wherepowerdistributionismoreasymmetricorincentivestoparticipateareweakorasymmetric,thencollaborativegovernanceismorelikelytosucceedifthereisastrong‘‘organic’’leaderwhocommandstherespectandtrustofthevariousstakeholdersattheoutsetoftheprocess.‘‘Organic’’ilabilityicationofthiscontingencyisthatthepossibilityforeffUTIONALDESIGNInstitutionaldesignrefersheretothebasicprotocolsandgroundrulesforcollaboration,whictotheculdbeincluded?Itisnosurprisetofindthattheliteratureoncollaborativegovernanceempha-sizesthattheprocessmustbeopenandinclusive(Andranovich1995;Burgeretal.2001;ChrislipandLarson1994;Gray1989;GuntonandDay2003;LaskerandWeiss2003;

556JournalofPublicAdministrationResearchandTheoryMargerum2002;Martin,Tett,andKay1999;Murdock,Wiessner,andSexton2005;PlummerandFitzgibbon2004;Poweretal.2000;Reilly1998,2001)becauseonlygroupsthatfeeltheyhavehadalegitimateopportunitytoparticipatearelikelytodevelopa‘‘com-mitmenttotheprocess.’’AsChrislipandLarson(1994)write,‘‘Thefirstconditionofsuccessfulcollaborationisthatitmustbebroadlyinclusiveofallstakeholderswhoareaffectedbyorcareabouttheissue.’’Thisincludespotentially‘‘troublesome’’(1989,68)observes,disputesoverthelegitimacyofincludingspecificstake-holdersarecertaintoarise,but‘‘...successfulcollaborationdependsonincludingabroadenoughspectrumofstakeholderstomirrortheproblem.’’Inthecoalcollaborationshestudied,theattempttoexcludecertainstakeholdersultimatelythreatenedthelegitimacyoftheprocess(Gray1989,155).(2001),forexample,foundthatsuccessfulcollaborativespayconsiderableattentiontogettingstakeholderstoparticiptudyoftheelectricindustry,Koch(2005,601)foundthatcollaborativegovernancerequiredtheinclusionof‘‘smallfirmsandpublicpowerorganizations’’tha-basedinclusionisnotsimplyarefltheheartofalegitimationprocessbasedon(1)theopportunityforstakeholderstodeliberatewithothersaboutpolicyoutcomesand(2)noninclusiverepresentation,therefore,threatenstounderminethelegitimacyofcollaborativeoutcomes(BeierleandKonisky2001;GeogheganandRenard2002;Smith1998).6Proactivestrategiesofmobilizinglesswell-representedstakeholdersarethusoftenseenasimportant(Weech-MaldonadoandMerrill2000).Yetaswesawearlier,stakeholdersmaynothaveanincentivetoparticipate,eraturesuggeststhatinclusivenessisthereforecloselylinkedtotheexclusivenessofthecollaborativeforum(Schuckman2001;Tett,Crowther,andO’Hara2003).Whenthecollaborativeforumis‘‘theonlygameintown,’’itiseasiertomotivatestakeholderstoparticipate;conversely,whentheyareexcluded,mple,KraftandJohnson(1999,136)foundthatenvironmentalgroupscreatedan‘‘alternativeforum’’se,theexistenceofalternativeforumsly(2001,71)putsit,‘‘Whenalter-nativeavenuesexistforresolution,itistheorizedthatacollaborativemethodofresolutionisnotoptimal.’’FungandWright(2001,24)notethat‘‘participantswillbemuchmorelikelytoengageinearnestdeliberationwhenalternativestoit—suchasstrategicdominationorexitfromtheprocessaltogether—aremadelessattractivebyroughlybalancedpower.’’Theliteraturealsosuggeststhatcleargroundrulesandprocesstransparencyareimportantdesignfeatures(Busenberg1999;GeogheganandRenard2002;GlasbergenandDriessen2005;GuntonandDay2003;Imperial2005;Murdock,Wiessner,andSexton2005;Rogersetal.1993).Bothcanbeunderstoodintermsofprocedurallegitimacy6Franklin(2001)describestheprocessofexclusionusedduringstrategicplanningfor15federalagencies.

AnsellandGasareaskingstakeholderstoengageingoodfaithkeholdesensitivetoissuesofequity,concernedaboutthepowerofotherstakeholders,itimacyoftheprocessdepends,inpart,uponstakeholders’perceptionsthattheyhavegottena‘‘fairhearing.’’Clearandconsistentlyappliedgroundrulesreassurestakeholdersthattheprocessisfair,equitable,andopen(Murdock,Wiessner,andSexton2005).Processtransparencymeansthatstakeholderscanfeelcon-fidentthatthepublicnegotiationis‘‘real’’aefinitionofrolescanalsobeimportant(Alexander,Comfort,andWeiner1998).Forexample,inhisstudyofanOntariocollaboration,Bradford(1998,565)arguesthatitwasnotcleariftheroleofstateofficialswastoprovide‘‘directiontothesocialpartners,clarif[y]expectationsaboutacceptableoutcomes[orlead]theplanningprocess.’’Formalizationofgovernancestructuresisthereforesometimesseenasanimportantdesignfeature(FungandWright2001,2003;Imperial2005;Weech-MaldonadoandMerrill2000).Theliteralreadydefinedcollaborativegovernanceas‘‘consensusoriented,’’uehereiollaborativesstudiedbyMargerum(2002),consensuswasseenaspromotingrer,consensusrulesareoftencriticizedforleadingto‘‘leastcommondenominator’’outcomes(CoglianeseandAllen2003;GuntonandDay2003).Theycanalsoleadtodecisionstalemates(CoglianeseandAllen2003),thoughitispossibleforcollaborativeprocessestobeginwithconsensusproceduresandthentoreverttootherproceduresinthecaseofstalemate(TillandMeyer2001).Afighsomeauthorspointtotheimportanceofdeadlines(GlasbergenandDriessen2005),particularlybecausecollaborativemeetingscanbeendless,Freeman(1997)blem,shewrites,isthatdeadlinesmayundercuttheongoingnatureofthecollaboration,ndandCruikshank(1987)andGuntonandDay(2003)suggestthattimetables,whenused,mustbe‘‘realistic.’’THECOLLABORATIVEPROCESSProcessmodelsofcollaboratimple,SusskindandCruikshank(1987,95)describetheconsensus-buildingprocessashavingaprenegotiationphase,anegotiationphase,andanimplemen-tationphase;Gray(1989)definesathree-stepcollaborativeprocess:(1)problemsetting,(2)directionsetting,and(3)implementation;andEdelenbos(2005,118)identifiesathree-stepprocessthatincludespreparation,policydevelopment,anddecisionmaking,modelofcollaborationisimportantforcallingurreadingoftheliterature,wewerestruckatthewayinwhichthecollaborativeprocessiscyclical

55orationoftenseemedtodependonachievingavirtuouscyclebetweencommunication,trust,commitment,understanding,andoutcomes(Huxham2003;Imperial2005).Thiscyclical—orifyouprefer,iterative—dthecollaborativeprocessdifficulttorepresentandwesusresentationofthecollab-orationprocessasacycleisclearlyitselfagreatsimplifiallsattentiontothewayinwhichfeedbacksfromearlycollaborationcanpositivelyornegativelyinflendiffir,sincecommunicationisattheheartofcollaboration,-to-FaceDialogueAllcollabosensus-orientedprocess,the‘‘thickcommunication’’allowedbydirectdialogur,thecoreoftheprocessofbreakingdownstereotypesandotherbarrierstocommunicationthatpreventexplorationofmutualgainsinthefirstplace(Bentrup2001).Itisattheheartofaprocessofbuildingtrust,mutualrespect,sharedunderstanding,andcommitmenttotheprocess(Gilliametal.2002;LaskerandWeiss2003;PlummerandFitzgibbon2004;Schneideretal.2003;TompkinsandAdger2004;Warner2006).Wearguethatface-to-facedialogueisanecessarybutnotsuffimple,itispossibleforface-to-facedialoguetoreinforcestereotysdiffieratureoncollabo-rationaboundswithexamplesofthewuildingThelackoftrustamongstakeholdersisacommonstartingpointforcollaborativegov-ernance(Weech-MaldonadoandMerrill2000).Theliteraturestronglysuggeststhatthecollaborativeprocessisnotmerelyaboutnegotiationbutalsoaboutbuildingtrustamongstakeholders(Alexander,Comfort,andWeiner1998;BeierleandKonisky2001;Brinkerhoff1999;GlasbergenandDriessen2005;Imperial2005;Murdock,Wiessner,andSexton2005;ShortandWinter1999;Tett,Crowther,andO’Hara2003;VangenandHuxham2003b).Infact,whentherehasbeenaprehistoryofantagonismamongstakeholders,wefoundthattrustbuildingoftenbecomesthemostprominentaspectoftheearlycollaborativeprocessandcanbequitedifficulttocultivate(Murdock,Wiessner,andSexton2005).Thisisnottosaythattrustbuildingisdcollaborativeleadersrec-ognizethattheymustbuildtrucomesevidentinthecasestudiesisthattrustbuildingis

AnsellandGashCollaborativeGovernanceinTheoryandPractice559atime-consumingprocessthore,(7)Iftheprehistoryishighlyantagonistic,thenpolicymakerscannotjustifythenecessarytimeandcost,menttotheProcessAlthoughtheterminologyusedvariesratherwidelyintheliterature,casestudiessuggestthatstakeholders’levelofcommitmenttocollaborationisacriticalvariableinexplain-ingsuccessorfailure(Alexander,Comfort,andWeiner1998;GuntonandDay2003;Margerum2001;Tett,Crowther,andO’Hara2003).InasurveyofAmericanandAustraliancollaborativegroups,Margerum(2002)foundthat‘‘membercommitment’’kcommitmentofpublicagenciestocollaboration,particularlyattheheadquarterslevel,isoftenseenasaparticularproblem(YaffeeandWondolleck2003).Commitmentiscloselyrelated,ofcourse,keholdersmaywishtoparticipateinordertomakesuretheirperspectiveisnotneglectedortosecurelegitimacyfortheirpositionortofulfillalegalobligation,rast,commitmenttotheprocessmeansdevelopingabeliefthatgoodfaithbargainingformutualgainsisthebestwaytoachievedesirablepolicyoutcomes(Burgeretal.2001).opermaybelievethatthebestwaytogethishousesbmitmenttocollaborationcanstillrequireaverysignificantpsychologicalshift,particularlyamongthosewhoregardtheirpositionsinabsoluteterms(Putnam2004;Putnam,Burgess,andRoyer2003).Asafirststep,suchashiftrequireswhatissometimescalled‘‘mutualrecognition’’(Saarikoski2000)or‘‘jointappreciation’’(Gray1989;PlummerandFitzgibbon2004).menttothecollaborativeprocessrequiresanup-frontwillingnesstoabidebytheresultsofdeliberation,evense,theconsensus-orientedbasidynamicsofbargainingcanleadinunexpecteddirections,andstakeholderscanexpe-riencepressuretoconformtopositionstheydonotfullyembrace(Saarikoski2000).mentdependsontrsoeasytoseehowclear,fair,com-mittingtoaprocessthatcouldgoinunpredictabledirections,stakeholdersmustfeelconfiofcom-mitmentandownershipcanbeenhancedasinvolvementincreases(Gilliametal.2002).Anadditionaldimensionofcommitmentissometimescalled‘‘ownershipofthepro-cess.’’Inthetypicaladversarialormanagerialprocess,yseektolobby,pressure,orinfluencepublicagencydecisionmakers,borativegovernanceshifts‘‘ownership’’o,thisimpliesatrickydilemma.

560JournalofPublicAdministrationRew‘‘own’’thedecision-makingprocesscollectivelywithotherstakeholderswhomayholdopposingviews(ElAnsari2003;GeogheganandRenard2002;Weech-MaldonadoandMerrill2000).sponsibilityrequiresstakeholderstoseetheirrelationshipwithotherstakeholdersinanewlight,scriticalbecausewhywouldyoushareresponsibilitywithpeopleyoudon’ttrust?Ifyouadopta‘‘responsible’’perspectivetowardtheprocess,whatistoguaranteethatyouropponentwillnottakeadvantageofyourwillingnesstoactingoodfaith?Sharedownershipmaybehinderedbypowerinterviewswithstake-holdersinvolvedinseaurchinharvesting,forexample,Warner(1997)foundthatfisherypersonnelanddivershaddifferenviewedthemselvesasassistingthefisherystaff,whereasthefisheedformsofcollaborationmaybecriticalwhereincentivestoparticipateareweak,butmandatedcooperatiore,(8)Evenwhencollaborativegovernanceismandated,achieving‘‘buyin’’terdependenceamongthestakeholdersislikelytoenhancecommitmenttocollab-oration,emptationsareprobablycheckedwheratureoncollectiveaction,ofcourse,suggeststhore,(9)CollaborativegovernancestrategiesUnderstandingAtsomepointinthecollaborativeprocess,stakeholdersmustdevelopasharedunderstand-ingofwhattheycancollectivelyachievetogether(Tett,Crowther,andO’Hara2003).Sharedunderstandingisvariouslydescribedintheliteratureas‘‘commonmission’’(Alexander,Comfort,andWeiner1998;RoussosandFawcett2000),‘‘commonground’’(WondolleckandYaffee2000),‘‘commonpurpose’’(Tett,Crowther,andO’Hara2003),‘‘commonaims’’(Huxham2003),‘‘commonobjectives’’(PadillaandDaigle1998),‘‘sharedvision’’(ManringandPearsall2004;WalterandPetr2000;WondolleckandYaffee2000),‘‘sharedideology’’(Waage2001),‘‘cleargoals’’(GlasbergenandDriessen2005;RoberstonandLawes2005),‘‘clearandstrategicdirection’’(Margerum2002),orthe‘‘alignmentofcorevalues’’(HeikkilaandGerlak2005).Sharedunderstandingcanalsoimplyagreementonadefinitionoftheproblem(Bentrup2001;North2000;Pahl-WostlandHare2004).Or,itmightmeelopmentofsharedunderstandingcanbeseenaspartofalarger‘‘collabo-rativelearningprocess’’(DanielsandWalker2001).Blatneretal.(2001)havedevelopedausefulsurveystrategyforassessingtheextentofcollectivelearningthatresultsfromcollaboration.

AnsellandGashCollaborativeGovernanceinTheoryandPractice561IntermediateOutcomesAnumberofthecasestudiessuggestthatcollaborationismorelikelytoensuewhenthepossiblepurposesandadvantagesofcollaborationarerelativelyconcreteandwhen‘‘smallwins’’fromcollaborationarepossible(ChrislipandLarson1994;RoussosandFawcett2000;Warner2006;Weech-MaldonadoandMerrill2000).Althoughtheseintermediateoutcomesmayrepresenttangibleoutputsinthemselves,werepresentthemhereascriticalprocessoutcomesthatareessemallwinscanfeedbackintothecollaborativeprocess,encouragingavirtuouscycleoftrustbuildingandcommitment(Rogersetal.1993;VangenandHuxham2003b).Theseconsiderationsleadustodrawthefollowingconclusions:(10)Ifpriorantagonismishighandalong-termcommitmenttotrustbuildingisnecessary,the,underthesecircumstances,stakeholdersorpolicymakerscannotanticipatethesesmallwins,actfindingisatypeofintermediateoutcomethatanumberofauthorsmen-tionedinapositivelight(Saarikoski2000).WealsonotetheargumentofVangenandHuxham(2003b)thatsmallwinsmaynotbeanappropriatestrategyfortrustbuildingwherestakeholdershavemoreggestthatinthissituation,trustcanSION:TIME,TRUST,ANDINTERDEPENDENCETheterm‘‘collaborativegovernance’’stopromisethatifwegoverncollaboratively,wemayavoidthehighcostsofadversarialpolicymaking,expanddemocraticparticipation,rofthestudiesreviewedherehavepointedtowardthevalueofcollaborativestrategies:bitteradversarieshavesometimeslearnedtoengageinproductivediscussions;publicmanagershavedevelopedmorefruitfulrelationshipswithstakeholders;andsophis-tictudies,however,pointtotheproblemsthatcollaborativestrategiesencounterastheypursuethesevaluedoutcomes:powerfulstakeholdersmanipulatetheprocess;publicagen-cieslackrealcommitmenttocollaboration;poseinthisarticlehasbeentodrawpositiveandnegativefindingstogetherintoacommonanalyticalframeworkthatcanbegintospecifytheconditionsunderwhichwecanexpectcollaborativegovernancetowork(atleastintermsof‘‘processoutcomes’’)nameta-analysisof137studiesofcollaborativegovernanceacrossarangeofpolicyareas,ourfindingsarelargelyempiricallyinductive—thewingtheseempiricalandtheoret-icalstudies,ourgoalhasbeentoidentifythecontinoringthesecontingentconditions,ourgoalhasbeentomovebeyondasituationinwhichcollaborativegovernanceisregardedas

562JournalofPublicAdministrationResearchandTheoryinherently‘‘good’’or‘‘bad.’’Wewantscholarsandpractitionerstoaskthemselvesaboutthecontextualconditionslikelytofacevethatthis‘‘contingency’’approachisusefulbothforpracti-tionerswhomaybeconsideringtheadoptimaryofthecriticalvariablesourmeta-analysisfoundtobeimportantincollaboration,figtionerscanusethisframeworeseriousdifferencesinthepowerofstakeholders?Doallstakeholdershavetheorganizationalcapacitytoparticipateinameaningfulway?Istheresufficientleadershiptoguidetheprocessthroughdifficultpatches?Howmuchremedialtrustbuildingisnecessary?Thesequestionsandmanymorearesuggestedbyfirdthisarticleasofferingacontingency‘‘theory’’inthesensethatitoffersaframeworkfororganizingaimisnotthatitisacompleteorfullyworkedoutsetofpropositionsorcausalrelationshipsbutrathertha1,forexample,specifiescausalrelationshipsbetecificationiribedearlier,weadoptedameta-analyticstrategyofsuccessiveapproximationinlieuofamoreambitiousquasi-experimentalstrategybecausekeyconceptsinthisliteraturewereweaklyspecifiedandour‘‘data’’purposesoffutureresearch,figure1canbetreatedsiblestrategiesforaqu,asurveyofindividualstakeholdersmightbeutilizedtooperationalizekeybehavioralvariables,suchas‘‘commitmenttotheprocess.’’GoodexamplesoftheuseofsurveysincollaborativegovernanceresearchincludeMargerum(2001)andFrame,Gunton,andDay(2004).Pre-andpostcollaborationsurveysmightbeaparticularlyusefulstrategyforassessingattitudinalchange(Blatneretal.2001).Second,researchmightbedesignedtotakeadvantageof‘‘naturalexperiments’’incollaboration:situationswheretherearemultipleindependentcasesofcollaborationoperatingundertheauspicesofasimilarregulatoryprogram,publicagency,esincludeMurdock,Wiessner,andSexton’s(2005)studyoftheEnvironmentalProtectionAgency’sProjectXLorSchneideretal.’s(2003)nt,however,isnotthataggregatestatisticalanaudiesareparticularlyvaluablewheretheinteractionbetweenvariablesisnonlinear,andwebelieveintensiveethnographicresearchmightbethemostsuccessfulstrategyfordevelopingudyresearchintotrustbuilding,thedevelopmentofsharedunderstanding,ludebyemphasizingthreecorecontingenciessuggestedbyouranalysis:time,trust,1doesnotfullyrepresenttheprominenceofthesecontingenciesbecausetheirinflctitionersoughttoconsidereachofthesegeneralcontingenciesbeforeembarkingonacollaborativestrategy.

AnsellandGashCollaborativeGovernanceinTheoryandPractice563Manyofthecasestudiesnotethatcollaborativegovernanceisatime-consumingprocess(GuntonandDay2003;Imperial2005;Margerum2002;RoussosandFawcett2000;TillandMeyer2001;Warner2006).Consensusbuilding,inparticular,requirestimeandcannotberushed(CoglianeseandAllen2003;YaffeeandWondolleck2003).Whenremedialtrustbuildingiscritical,thetimenecessaryforincreasingtrustislikelytoaddsignifiore,collaborativegovernanceisprobablynotagoodstrategyr,itneedstobepointedoutthatup-frontinvestmentineffectivecollaborationcansakeholdersachieveaworkingconsensus,,policymakersmightbemorefavorabletocollaborativegover-nancewheretheyexpectadiffircingLogsdon’s(1991)argumentaboutinterdependenceandthemodelbyVangenandHuxham(2003b)oftrustbuilding,ouranalysissuggeststhatagenciesoughttoconsidertheinted,forinstance,thathighconflictsituationscharacterizedbylowtrustcouldstillbeependencefostersadesiretoparticipateandacommitmenttomeaningfulcollaboration,rast,whereinterde-pendenceisweaker,itwillbediffiolderswillengageincollaborationwithoneeyeonalternative(noncollaborative)takeholderisthreateningtodefectfromcollaboration,thecommitmentofallstakeholdersislikelytosuffer,anditwillbedifficulttodevelopasenseofownership,understanding,portanttopointoutthatbothtrustandinterdependencearepartlyendoge-nous—theya,stakeholdersenteringintoacollaboratioughdialoguewithotherstakeholdersandthroughachievementofsuccessfulintermediateoutcomes,theymaycometoanewunderstandingoftheirrelationship(HeikkilaandGerlak2005;2006).Manyofthecaseswereadsuggestthatstakeholderscomrcollaborativegovernanceisapassingfancy,fi-dentlypredict,however,thatthedemandforbettercooperationandcoordinationIXFirstapproximation(32):Andranovich(1995),Beierle(2000),BooherandInnes(2002),BrysonandCrosby(1992),ChrislipandLarson(1994),Coggins(1999),DanielsandWalker(2001),Echeverria(2001),Fawcettetal.(1995),Freeman(1997),FungandWright(2001),Gray(1989),Healey(1996,2003),InnesandBooher(1999a,1999b),KraftandJohnson(1999),Langbein(2002),Lee(2003),Lober(1997),NelsonandWeschler(1998),Ozawa(1993),Reilly(1998,2001),SchedlerandGlastra(2001),Schuckman(2001),Smith(1998),SusskindandCruikshank(1987),TakahashiandSmutny(2002),Thomas(1995),Weber(2003),andWondolleckandYaffee(2000).Secondapproximation(30):Ansell(2003),BeierleandKonisky(2001),Coglianese(1997),ConleyandMoote(2003),Ebrahim(2004),Ekoko(2000),Elliotetal.(1999),

564JournalofPublicAdministrationResearchandTheoryEnglish(2000),Fung(2001),Gebhardt,Kaphingst,andDeJong(2000),Hamalainenetal.(2001),Imperial(2005),InnesandBooher(2003),Leach,Pelkey,andSabatier(2002),Logsdon(1991),Manring(1998,2005),McCloskey(2000),Meyer(1996),MizrahiandAbramson(2000),Murdock,Wiessner,andSexton(2005),Phillips(2001),PlummerandFitzgibbon(2004),Schneideretal.(2003),Seidenfeld(2000),Springer,StokesSharp,andFoy(2000),Waage(2001),WalterandPetr(2000),Weech-MaldonadoandMerrill(2000),YaffeeandWondolleck(2003).Thirdapproximation(33):Alexander,Comfort,andWeiner(1998),Borrini-Feyerabend(1996),BouwenandTaillieu(2004),Bradford(1998),Brinkerhoff(1999),Brown(2002),CoglianeseandAllen(2003),Coughlinetal.(1999),FarringtonandBoyd(1997),Franklin(2001),GeogheganandRenard(2002),Gilliametal.(2002),GuntonandDay(2003),HeikkilaandGerlak(2005),LaskerandWeiss(2003),Lasker,Weiss,andMiller(2001),Lindell(2004),ManringandPearsall(2006),Margerum(2002),Merkhofer,Conway,andAnderson(1997),Mitchell(2005),Mutimukuru,Nyirenda,andMatose(2002),PlummerandFitzgibbon(2004),Saarikoski(2000),ShortandWinter(1999),Tett,Crowther,andO’Hara(2003),TillandMeyer(2001),TompkinsandAdger(2004),Verstraetenetal.(2003),Warner(1997),Warner(2006),WeaverandMoore(2004),andWeible,Sabatier,andLubell(2004).Fourthapproximation(42):Abdelhadietal.(2004),Blatneretal.(2001),Bryson,Cunningham,andLokkesmoe(2002),Buanesetal.(2004),Burgeretal.(2001),Busenberg(1999),Carteretal.(2003),Edelenbos(2005),ElAnsari(2003),Frame,Gunton,andDay(2004),Futrell(2003),Geldenhuys(2004),GemmillandBamidele-Izu(2002),GlasbergenandDriessen(2005),HeikkilaandGerlak(2005),Huxham(2003),HuxhamandVangen(2000),Innesetal.(2006),KlijnandKoopenjan(2000),Lee(2003),Mahonetal.(2003),Margerum(2001),Martin,Tett,andKay(1999),Matta,Kerr,andChung(2005),MitchellandShortell(2000),North(2000),Pahl-WostlandHare(2004),Pelletieretal.(1999),Pokornyetal.(2004),Poweretal.(2000),Redpathetal.(2004),Rhoadsetal.(1999),RoberstonandLawes(2005),Rogersetal.(1993),RoussosandFawcett(2000),Rummery(2006),Ryan(2001),Selman(2004),Sjoberg(2003),VangenandHuxham(2003a,2003b),Waage(2001),andWarner(2006).REFERENCESAbdelhadi,A.W.,,,ipatorymanagement:WoulditbeaturningpointinthehistoryoftheGeziraScheme?IrrigationandDrainage53:429–der,JefferyA.,t,anceinpublic-privatecommunityhealthpartnerships:AsurveyoftheCommunityCareNetwork:fitManagement&Leadership8:231–ovich,ingconsensusinpublicdecisionmaking:ApplyinginterestbaselofAppliedBehavioralResearch31:429–,ityembeddednessandcollaalmovementsandnetworks:Relationalapproachestocollectiveaction,ianiandDougMcAdam,123–:e,lityofstakeholder-baseddecisions:sionPaper00-56,ResourcesfortheFuture.

AnsellandGashCollaborativeGovernanceinTheoryandPractice565Beierle,ThomasC.,ewegainingfromstakeholderinvolvement?nmentandPlanningC:GovernmentandPolicy19:515–e,ThomasC.,ngcollaboration:TheFederalAdvisoryComnmentalLawReporter29:p,tionofacollaborativemodel:nmentalManagement27:739–r,Keith,MatthewCarroll,StevenDaniels,tingtheapplicationofcollaborativelearningtotheWenatcheefinmentalImpactAssessmentReview21:241–,DavidE.,lofPlanningEducationandResearch21:221–i-Feyerabend,orativemanagementofprotectedareas:,Switzerland:IUCN–,Rene,-partycollaborationassociallearningforinterdependence:DevlofCommunity&AppliedSocialPsychology14:137–rd,ctsforassociativegovernance:LessonsfromOntario,cs&Society26:539–rhoff,ingstate-civilsocietycollaboration:fitandVoluntarySectorQuarterly28(Suppl.1):59–,orativegovernanceversusconstitutionalpolitics:DecisionrulesforsustainabilityfromAustralia’nmentalScienceandPolicy5:19–,JohnM.,ncisco,CA:,JohnM.,GaryCunningham,dowhenstakeholdersmatter:ThecaseofproblemformulationfortheAfricanAmericanMenProjectofHennepinCounty,AdministrationReview62:568–,Arild,SveinJentoft,GeirRunarKarlsen,AnitaMaurstad,andSiriSøeinterest?An&CoastalManagement47:207–,Joanna,MichaelGochfeld,,LynnWaishwell,CamillaWarren,e,policy,stakeholders,andfishconsumptionadvisories:DevelopingafinmentalManagement27:501–erg,Sciences32:1–,Jane,BriekeSteenhof,EstherHaldimann,orativeforestmanagementinKyrgyzstan:eperSeriesNo.108,ip,David,orativeleadership:ncisco,CA:s,tingfederalnaturalresources:yLawQuarterly25:nese,ingconsensus:wJournal46:nese,Cary,ngsector-basedconsensus:AreviewoftheEPA’gPaperRWPO3,,Alexander,yandNaturalResources16:371–86.

566JournalofPublicAdministrationResearchandTheoryConnick,Sarah,esofcollaborativewaterpolicymaking:lofEnvironmentalPlanningandManagement46:177–in,C.,,pf,or:UniversityofMichigan,s,Steven,gthroughenvironmentalconflict:rt,CT:m,utionalpreconditionstocollaboration:stration&Society36:208–rria,esslikefailure:mandMaryEnvironmentalLawandPolicyReview25:bos,utionalimplicationsofinteractivegovernance:ance:AnInternationalJournalofPolicy,AdministrationandInstitutions18:111–,ingpolitics,economics,andconservation:pmentandChange31:131–ri,ionalpartnershipsforpublichealth:Dostakeholdersperceivesimilaroutcomes?JournalofPublicHealthManagementPractice9:136–,Susan,,PaulKrueger,NancyVoorberg,andSarahWakefierofperception:Healysis19:621–h,thestakeholdersinenvironmentalriskdecisions?Risk:Health,Safety&Environment11:243–gton,John,pmentPolicyReview15:371–t,StephenB.,AdriennePaine-Andrews,sco,z,r,,ms,mpowermentthanJournalofCommunityPsychology23:677–,TanisM.,ThomasGunton,eofcollaborationinenvironmentalmanagement:lofEnvironmentalPlanningandManagement47:59–in,gthepublicinterest?anReviewofPublicAdministration31:126–n,wReview45:,tableautonomy:cs&Society29:73–,Archon,ingdemocracy:cs&Society29:5–41.———.ingdemocracy::l,e,Technology,&HumanValues28:451–dt,Thomas,KimberlyKaphingst,lofAmericanCollegeHealth48:211–huys,rvestingfortraditionalmedicine:navianJournalofForestResearch19(Suppl.4):103–l,Barbara,alenvironmentalgovernance:Options&opportunities,a,1–en,CT:gan,Tighe,communityinvolvement:12(2):16–26.

AnsellandGashCollaborativeGovernanceinTheoryandPractice567Gilliam,Aisha,DavidDavis,TraceyBarrington,RomelLacson,GaryUhl,ucationandPrevention14(Suppl.A):5–rgen,Pieter,ctiveplanningofinfrastructure:nmentandPlanningC:GovernmentandPolicy23:263–,orating:ncisco,CA:,ThomasI.,oryandpracnments31(2):5–inen,Raimo,EeroKettunen,HarriEhtamo,tingaframewecisionandNegotiation10:331–,sus-buildingacrossdifficultydivisions:ngPracticeandResearch11:207–16.———.ngTheory2:101–la,Tanya,mationoflarge-scalecollaborativeresourcemanagementinstitutions:Clarifyingtherolesofstakeholders,science,StudiesJournal33:583–,ManagementReview5:401–,Chris,shipintheshapingandimplementationofcollaborationagendas:Howthingshappenina(notquite)yofManagementJournal43:1159–al,ollaborationasagovernancestrategy:stration&Society37:281–,Judith,loftheAmericanPlanningAssociation65:412–23.———.susbuildingasroleplayingandbricolage:loftheAmericanPlanningAssociation65:9–26.———.orativepolicymaking:berativepolicyanalysis:Understandinggovernanceinthenetworksociety,ar,33–dge:,JudithE.,SarahConnick,LauraKaplan,orativegovernanceintheCALFEDprogram:gPaperSeriesPaperWP-2006-01,ian,Anne,itationtocollaboration:AdministrationReview57(5):396–410Klijn,E.H.,ciansandinteractivedecisionmaking:Administration78:365–,CharlesH.,restLawReview40:,Michael,aterandthepromiseofcollaborativedecisionmaking:rdsustainablecommunities:Transitionandtransformationsinenvironmentalpolicy,,113–,MA:in,sivebureaus,equity,andregulatorynegotiation:lofPolicyAnalysisandManagement21:449–,RozD.,ningparticipationincommunityproblem-solving:AlofUrbanHealth:BulletinoftheNewYorkAcademyofMedicine80:14–,RozD.,,rshipsynergy:ApracticbankQuarterly,79(2):179–205.

568JournalofPublicAdministrationResearchandTheoryLeach,William,,olderpartnershipsascollaborativepolicymaking:EvaluatilofPolicyAnalysisandManagement21:645–,ingcollaborativegovernancefornaturalareamanagement:ationalPlanningStudies8:35–l,nflictingintereststocollectiveconsentinadvancedvocationaleducation:lofEducationandWork17:257–,ningtheformationofbusiness-environmentalistcollaboration:Science30:1–n,stsandintlofAppliedBehavioralScience27:23–,LawrenceE.,ch,inggovernance:gton,DC:,Robin,SharonAlmerigi,PatrickMcConney,ChristopherParker,&CoastalManagement46:1–g,orativeresourcemanagement:Organizationalbenefistration&Society30:274–90.———.stration&Society37:57–g,SusanL.,nganadaptiveecosystemmanagementnetworkamongstakeholdersoftheLowerRoanokeRiver,NorthCarolina,yandSociety10(2)::///vol10/iss2/art16/(accessedOctober31,2007).Margerum,zationalcommitmenttointegratedandcollaborativemanagement:nmentalManagement28:421–31.———.orativeplanning:lofPlanningEducationandResearch21:237–,Jane,andLynTett,pingcollaborativepartnerships:Limitsandpossibilitiesforschools,ationalStudiesinSociologyofEducation9:59–,JagannadhaRao,JohnKerr,restregulationtoparticipatoryfacilitation:ForestemplofEnvironmentalPlanningandManagement48:475–key,aisoUniversityLawReview34:fer,MileyW.,RarileeConway,ttributeutilityanalysisasaframeworknmentalManagement21:831–,denvironmentalpublicgoods:pmentandChange27:453–ll,ipatorypartnerships:Engagingandempoweringtoenhanceenvironmentalmanagementandqualityoflife?SocialIndicatorsResearch71:123–ll,Shannon,ernanceandmanagementofeffectivecommunityhealthpartnerships:Atypologyforresearch,kQuarterly78:241–i,Terry,orationbetweensocialworkersandphysicians:WorkinHealthCare31(3):1–k,Barbara,CarolWiessner,olderparticipationinvoluntaryenvironmentalagreements:e,Technology&HumanValues30:223–kuru,T.,da,ngamongstourselves:Towardsadaptivenessstakeholdersinforestmanagementthroughsociallearning,ted

AnsellandGashCollaborativeGovernanceinTheoryandPractice569atTheCommonsinanAgeofGlobalisation,the9thConferenceoftheInternationalAssociationfortheStudyofCommonProperty,VictoriaFalls,Zimbabwe,:///archive/00000882/.Nelson,Lisa,utionalreadinessforintegratedwatershedmanagement:ScienceJournal35:565–,espacefororganisationfrombelowwithintheUKgovernment’sactionzones?Atestof‘collaborativeplanning’.UrbanStudies37:1261–,ingcitizenparticipationinenvironmentaldecisionmaking:nmentandPlanningC:GovernmentandPolicy11:103–a,YolandaC.,strationinSocialWork22(1):65–-Wostl,Claudia,lofCommunity&AppliedSocialPsychology14:193–ier,David,VivicaKraak,ChristineMcCullum,UllaUusitalo,pingofcollectivevaluesthroughdeliberativedemocracy:AnempiricalstudyfromNewYork’Sciences32:103–ps,anstakeholders:lofCanadianStudies35:182–,Barbara,RobinWarsh,ipatorymanagementinapublicchildwelfareagency:strationinSocialWork22(1):19–r,Ryan,-managementofnaturalresources:nmentalManagement33:876–y,Benno,RaviPrabhu,CynthiaMcDougall,takeholders’participationlofForestry102(1):35–,James,JohnMcKenna,d,,pingintegratedparticipatorymanagementstrategiesforAtlanticdunesystemsinCountyDonegal,29:143–,ationsJournal20:275–,LindaL.,GuyBurgess,’tgoonlikethis:FramechangesinintractableconflnmentalPractice5:247–h,S.M.,,,,field,rez,ecisionmodelingwithstakeholderstoreducehuman-wildlifeconflict:vationBiology18:350–,itiesinconflict:ResolvingdifferencesthroughcollofSociologyandWelfare25:115–42.———.orationinaction:strationinSocialWork25(1):53–,BruceL.,DavidWilson,MichaelUrban,ctionbetweenscientistsandnonscientistsincommunity-basedwatershedmanagement:nmentalManagement24:297–ton,John,rceptionsofconservationandparticipatorymanagementofiGxalingenwaForest,nmentalConservation32:64–,Todd,BethHoward-Pitney,ry,,,teristicsEducationResearch,Theory&Practice8:345–s,StergiosTsai,wofReviewofPublicHealth21:269–y,rshipsandcollaborativegovernanceinwelfare:Policy&Society5:293–303.

570JournalofPublicAdministrationResearchandTheoryRyan,shipincollaborativepolicy-making:Sciences34:221–oski,nmentalimpactassessment(EIA)nmentalImpactAssessmentReview20:681–er,Petra,icatingpolicyinlatemodernsociety:&Politics29:337–der,Mark,JohnScholz,MarkLubell,DenisaMindruta,ngconsensualinstitutions:anJournalofPoliticalScience47:143–man,hardchoices:gtonUniversityLawQuarterly79:feld,ringstakeholders:LimitsoncollaborationforflmandMaryLawReview41:,lofEnvironmentalPlanningandManagement47:365–,Christopher,blemofcommonland:lofEnvironmentalPlanningandManagement42:613–g,alysis23:739–,orativeapproachestoPacificNorthwestfisheriesmanagement:etteJournalofInternationalLawandDisputeResolution6:er,David,DeborahStokesSharp,natedservicedeliveryandchildren’swell-being:lofCommunityPractice8(2):39–,anceastheory:ationalSocialScienceJournal50:17–28.———.inginstitutionsforgovernanceincomplexenvironments:Normativeratind,Lawrence,ngtheimpasse:k:shi,Lois,orativewindowsandorganizationalgovernance:fitandVoluntarySectorQuarterly31:165–,Lynn,JimCrowther,andPaulO’lofEducationPolicy18:37–,ncisco,CA:,JohnE.,Physics80:370–ns,EmmaL.,aptivemanagementofnaturalresourcesenhanceresiliencetoclimatechange?EcologyandSociety9(2)::///vol9/iss2/art10(accessedOctober31,2007).Vangen,Siv,ngleadershipforcollaborativeadvantage:DilehJournalofManagement14:S61–76.———.ingcollaborativerelations:lofAppliedBehavioralScience39:5–aeten,Gert,JeanPoesen,GerardGovers,KatleenGillijns,AntonVanRompaey,atingscience,policyandfarmerstoreducesoillossandsedimentdeliveryinFlanders,nmentalScience&Policy6:95–,S.A.2001.(Re)calGeography20:839–57.

AnsellandGashCollaborativeGovernanceinTheoryandPractice571Walter,Uta,esinSociety:TheJournalofContemporaryHumanServices81:494–,ipatorymanagement,popularknowledge,andcommunityempowerment:cology25:29–,stainableparticipation?esourcesDevelopment22(1):15–,Mark,resentedattheAnnualMeetingoftheRuralSociologicalSociety,Sacramento,CA,August11–,ngsocietybackin:Grassrootsecosystemmanagement,accountability,dge,MA:-Maldonado,Robert,ngpartnershipswiththecommunity:lofHealthcareManagement45:189–,Christopher,er,risonofacollaborativeandtop-downapproachtotheuseofscienceinpolicy:StudiesJournal32:187–ms,Bruce,acy,dialogue,andenvironmentaldisputes:en,CT:leck,JuliaM.,collaborationwork:gton,DC:,StevenL.,orativeecosystemplanningprocessesintheUnitedStates:nments31(2):59–72.