《the dynamics of political，embeddedness in china》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
AdministrativeScienceQuarterly2017,Vol.62(1)67–104TheDynamicsÓTheAuthor(s)2016Reprintsandpermissions:ofPoliticalsagepub.com/journalsPermissions.navEmbeddednessDOI:10.1177/0001839216657311journals.sagepub.com/home/asq*inChina123HeatherA.Haveman,NanJia,JingShi,2andYongxiangWangAbstractEconomictransitionsincountriesthatmovefromstateplanningandredistribu-tiontomarketexchangecreatebusinessopportunitiesbutalsouncertainty,becausemanyinterdependentfactors—modesofexchange,typesofproducts,andformsoforganizations—areinflux.Uncertaintyisevengreaterwhenthecountry’spoliticalinstitutionsremainauthoritarianbecausetheruleoflawisweakandstatebureaucratsretainpowerovertheeconomy.ThisstudyoflistedfirmsinChina,whichhasrecentlyseeneconomictransitionbutpersis-tentauthoritarianism,showsthatinsuchcontexts,firmscanreduceuncer-taintybydevelopingrelationshipswithstatebureaucrats,whichhelpfirmslearnhowstatebureaucraciesoperateandengendertrustbetweenfirmsandbureaucrats.Together,knowledgeandtruststabilizeoperationsandhelpper-suadebureaucratstolightenregulatoryburdens,grantfirmsaccesstostate-controlledresources,andimprovegovernmentoversight.Ourresultsshowthataseconomictransitionsproceedanduncertaintyincreases,business–statetiesincreasinglyimprovefirmperformance.Wealsoinvestigatetwolikelycon-tingencies,industryandfirmsize,andtwoimportantcausalmechanisms,accesstobankloansandprotectionfromrelated-partyloans,andshowthatthevalueofbusiness–staterelationsvariesovertime,dependingonthetrajec-toryofbotheconomicandpoliticalinstitutions.Keywords:firmperformance,corporategovernance,topmanagement,socialnetworks,economictransition,China,politicalembeddedness,marketdevelopmentEconomictransitionsfromstateplanningandredistributiontomarketexchangecantakemanypaths(StarkandBruszt,1998;Whitley,1999;Przeworskietal.,*Theauthorsarelistedinalphabeticalorder.1HaasSchoolofBusiness,UniversityofCalifornia,Berkeley2MarshallSchoolofBusiness,UniversityofSouthernCalifornia3SchoolofEconomics,FinanceandMarketing,RMITUniversity,andInternationalInstituteforFinancialStudies,JiangxiUniversityofFinanceandEconomics 68AdministrativeScienceQuarterly62(2017)2000),butallsuchtransitionsgenerateuncertaintyforeconomicactorsasnewvenuesforeconomicexchange,newkindsofgoodsandservices,andnewformsoforganizationsariseandoldonesdisappearorevolve.Uncertaintydur-ingeconomictransitionisalsoaffectedbypoliticalinstitutions,notablywhetherornotgovernmentsbecomedemocratic.Democracysupportsthecreationoftheruleoflaw(AcemogluandRobinson,2012;AssiotisandSylwester,2015),whichinturnsupportsmarketfunctioningbyreducinguncertaintyforeconomicactors.Incontrast,persistentauthoritarianismunderminestheruleoflawandheightensuncertainty,thushinderingmarketfunctioning.Becauseeconomictransitionsareinherentlydynamicphenomena,theyrequiredynamicanalysis,withattentiontochangesovertime(orthelackthereof)inbotheconomicandpoliticalinstitutions.Althoughtheuncertaintygeneratedbyeconomictransitionisgreaterunderpersistentlyauthoritarianregimes,itcanbereducedbycreatingrelationshipswithstatebureaucrats(PfefferandSalancik,1978;Burt,1982),becausepersis-tentlyauthoritarianregimesnotonlygenerateweakruleoflawbutalsogivestatebureaucratsconsiderablepowerovertheeconomy.Asmarketsdevelopinauthoritarianregimes,relationshipsbetweenfirmsandstateauthoritieshelpfirmslearnhowstatebureauswork,reducinguncertaintyforfirmsandengen-deringtrustbetweenfirmsandstateauthorities.Thishelpsbusinessesgraspnewopportunitiesbystabilizingoperationsandmakingiteasiertopersuadebureaucratstoallowaccesstoresourcesandcustomers,reduceregulatoryburdens,lowerfeesandtaxes,andprovidebetterregulatoryoversight.Therefore,asmarketsdevelopinauthoritarianregimes,business–statetiesincreasinglyimprovefirmperformance.Inthispaper,westudybusiness–statetiesinChina,whichsawgreateco-nomicchangecoupledwithpoliticalstability.Startingin1978,Chinamovedfromstateplanningandredistributiontomarketexchange:transactionsatmar-ketprices(ratherthanthroughstatemandate)rosefrom6percentin1978to97percentin2003forfarmcommodities,from0to87percentforproducergoods,andfrom3to96percentforretailpurchases(Dougherty,Herd,andHe,2007).YetChina’spoliticalinstitutionsremainedstable.TheCommunistPartyretainedabsolutecontroloverpolitics,thestatebureaucracy,andthelegalsys-tem,solawsconcerningproperty,contracts,andcompetitionwereslowtodevelopandnotenforcedimpartially(Clarke,Murrell,andWhiting,2008;McGregor,2012).Politicalinertiagavestateauthoritiespersistentpowerovertheeconomythroughtheirpowertoissuepermitsandlicensesandtolevyfees,fines,andtaxes(e.g.,Y.Lin,2001;Yang,2005)andtheircontrolofstate-ownedenterprises,land,andcapital(e.g.,Shih,2008;Choo,2014).Thecombi-nationofeconomicchangeandpoliticalinertiacreateduncertaintyforbusi-nessesandalteredthenatureoftheirrelationshipswiththestate.Previousresearchonbusiness–staterelationsinChina,however,haspro-videdlittledirectevidenceofthecausalmechanismsunderpinningtheireffectsandhasseldomconsideredcontingenciesthatlimitthoseeffects(forexcep-tions,seeY.Lin,2001;NeeandOpper,2010,2012).Moreover,mostpreviousresearchhasbeencross-sectionalandsocouldnotcapturethedynamicsofbusiness–stateties,andthefewlongitudinalstudieshavegenerallyreliedonrepeatedcross-sections,againrenderingdynamicanalysisimpossible(forexceptions,seeKeister,2004;Wang,Guthrie,andXiao,2012;LiandLiang, Havemanetal.692015).Incontrasttopreviouswork,weconsiderbotheconomicandpoliticalinstitutionsandconductadynamicanalysis.Westudylistedfirms,whichareamongthelargestandmostimportantinChina,andfocusonacommonformofbusiness–statetie:havingstatebureau-cratsserveaslistedfirms’executivesordirectors,whichembedspoliticalinsti-tutionsineconomicinstitutions(ChenandDickson,2010).Business–statetiesaffecttwoimportantcausalmechanismsthathaveanimpactonlistedfirms’performance.Thefirstisaccesstobankloans,whichfuelsgrowthandimprovesperformance.Thesecondisprotectionfrompressuresbycontrollingshareholderstosiphonfundsfromlistedfirmstosupportothermembersofcontrollingshareholders’businessgroups,whichimpairslistedfirms’perfor-mance.Webeginbyexplainingtransitionsfromstateredistributiontomarketexchange,ingeneralandthenforChinainparticular,andthenanalyzehow,inthefaceofeconomictransitionwithoutpoliticalchange,relationshipsbetweenbusinessesandthestateincreasinglyimprovefirmperformanceinanuncertaineconomicenvironment.ECONOMICTRANSITIONSGENERATEUNCERTAINTYDuringeconomictransitions,elementsofoldsystemsdisappearandarereplacedbyelementsofnewsystems.Newkindsofeconomicexchange,suchasmarket-basedwaystobuyandsell,andnovelvenuesfordoingso,developandoldones,suchasstateplanningandredistribution,dissolveorevolveintonewforms.Newformsofeconomicactors,suchasprivatelyownedbusi-nesses,tradeandindustryassociations,financialserviceinstitutions,regulatoryagencies,andlaborunions,ariseandoldones,suchasstate-ownedproductionandservicefacilities,andministerialplanningandredistributiondepartments,disappearoraretransformed.Newtypesofgoodsandservicesappear,includ-ingmorevariedconsumergoods,financialservices,andprivatehousing,andreplaceolder,standard-issueconsumergoodsandenterprise-basedhousingandsocialservices.Thesechangesarehighlyinterdependent:changeinoneelement,suchasthelaunchofanewmarketvenue,impelschangesinotherelements,forexample,thecreationofanewkindofbusiness,regulatoryagency,orindustryassociation,andthedevelopmentofanewtypeofgoodorservice.Thuseconomictransitionsareinstancesofcomplexsystemchange:becauseelementsofeconomicsystemsareinterdependent,changeinanysin-gleelementhasimplicationsformanyothers;moreover,theseimplicationscannotbepredictedinadvancebecauseofboundedrationality(Simon,1957)butmustbelearnedthroughtrialanderror(Simon,1962).Theupshotisthateconomictransitionsgeneratemuchuncertaintyforbusinesses:firms’decisionmakerscannotpredictwhatactionsareorarenotpossible,acceptable,andeffective,orwhatkindsofactorsareorarenotacceptableandcapable.Uncertainty,inturn,reducestheabilityoffirms’decisionmakerstotakeadvantageofthenewopportunitiescreatedbyeconomictransitions,makesitharderforfirmstodecidehowtoinvestresources,andincreasestheirlikeli-hoodofmakingmistakes.Thegreatertheuncertainty,themoredifficultitisforfirmstobesurewhichinvestmentswillbegood,muchlessbest,sothemoremistakesfirmswillmakeandtheworseandmorecostlythosemistakeswillbe. 70AdministrativeScienceQuarterly62(2017)Economictransitions,andtheuncertaintytheyunleash,aredeeplyinflu-encedbypoliticalinstitutions.The‘‘rulesofthegame’’putinplacebygovern-mentsarefundamentaltomarketfunctioning.Strongruleoflaw—notablytransparentandimpartiallyenforcedlawsconcerningpropertyrights,contracts,andcompetition—determineshow,andhowwell,marketexchangefunctions(North,1990;AcemogluandRobinson,2012;Walder,Isaacson,andLu,2015).Thestrongertheruleoflaw,themorepoliticalinstitutionssupportmarketexchangeandthemoretheuncertaintycausedbyeconomictransitionisreduced.Strongruleoflawiscreatedandmaintainedbydemocraticgovern-ments,whichareaccountableandresponsivetocitizens(AcemogluandRobinson,2012;AssiotisandSylwester,2015).Notalleconomictransitionshavebeenaccompaniedbypoliticaltransitionstodemocracyandthustoastrongruleoflaw.IncountrieslikePolandandRomania,democraticinstitutionshavedeveloped,whileincountrieslikeChinaandRussia,politicalinstitutionshaveremainedauthoritarian(Whitley,1999;Przeworskietal.,2000).Thelevelofuncertaintygeneratedbyeconomictransitionsdependsonthenatureofpoliticalregimes:itisgreaterwhentheyremainauthoritarianandlesswhentheybecomedemocratic.Andthepowerofstateauthoritiesovertheeconomy—specifically,theirabilitytodevelopapredatoryordevelopmentalstate(Johnson,1982;Evans,1995)—dependsonthenatureofpoliticalregimes:statepredationandstate-leddevelopmentarebothmorelikelywhenpoliticalregimesremainauthoritarianandlesslikelywhentheybecomedemo-cratic;thebalancebetweenthemdependsonthecontentofstatepoliciesandtheincentivestheycreateforstatebureaucrats.Inturn,greaterpowerovertheeconomygivesstatebureaucratsmorecapacitytoinfluencetheuncertaintygeneratedbyeconomictransition.EconomicTransitioninChinaIn1978,Chinabeganagradualandas-yetincompleteeconomictransition,withnewmodesofexchange,goodsandservices,andeconomicactorsappearing,andoldonesdisappearingorevolving.Morethanathousandstate-ownedenterpriseswerereorganizedasjoint-stockcorporationsandlistedonthestockexchanges,whilemillionsofprivatelyownedenterpriseswerelaunched.Theportionofcapitalinvestedinnon-state-ownedenterprisesrosefrom2percentin1992to41percentin2007,whiletheportionofthelaborforceworkinginnon-state-ownedenterprisesrosefrom16to68percent(ChinaStatisticalYearbook,variousyears).Ledfirstbymanufacturingandlaterbyservices,GDPpercapitaexplodedfrom$316(incurrentU.S.dollars)in1990to$7,590in2014,anaverageannualincreaseof9.0percent(WorldBank,2015).Incontrast,China’spoliticalinstitutionsremainedstable.TheCommunistPartyretainedabsolutecontroloverpolitics,exceptvillageelections;eventhere,partymembersdominatedelectedcouncils(OiandRozelle,2000;Lu,2012).Thereforethepoliticalregimewaspersistentlyauthoritarian:since1976,China’sPolityIVscorehasbeen–7onascalerunningfrom–10(fullyinstitutio-nalizedautocracy)to+10(fullyinstitutionalizeddemocracy)(http://www.system1icpeace.org/polity/polity4.htm).Thepartyretainedcontroloverthestate1Thisscoretakesintoconsiderationthecompetitivenessofpoliticalparticipation,theopennessandcompetitivenessofexecutiverecruitment,andconstraintsonthechiefexecutive. Havemanetal.71bureaucracyandthelegalsystem(Clarke,Murrell,andWhiting,2008;McGregor,2012),sotheruleoflawwasweak:lawsgoverningpropertyrights,contracts,andcompetitionwereslowtodevelopandnotenforcedimpartially(thejudiciarywasnotindependentoftheparty),andstategovernanceoftheeconomywasoftenambiguousandopaque(Clarke,Murrell,andWhiting,2008).Becauseelec-toralpoliticswerenotavailabletodisciplinestatebureaucrats,theChinesestatebecameapredatoryone:bureaucratscouldpreyonbusinessestofulfillprivategoalsaswellasofficialgoals(WittandRedding,2013).Thegradualandpiece-mealnatureofChina’seconomictransitioncreatedvariedandsometimesincom-patibleregulations,increasinguncertaintyandenhancingopportunitiesforstatepredation(Manion,2004;Naughton,2007;BianandZhang,2014).Statebureauscontrolledkeyresourcesandmanybusinessactivities.Therewerelowlevelsoftrustinthestatebureaucracy,duenotonlytothepossibilityofstatepredationbutalsotothegreatuncertaintycausedbythestatebureaucracy’sambiguity,opacity,andinternalheterogeneity(BianandZhang,2014).ThecombinationofeconomictransitionandpoliticalinertiageneratedgreatuncertaintyforbusinessinChina.Andbecausetheeconomictransitionwasgradual,changecontinued,exacerbatinguncertainty.TheChinesestate’srelationshiptotheeconomywasverycomplex—notjustpredatorybutalsoatdifferenttimesandindifferentplacesdevelopmental,entrepreneurial,andembedded(WittandRedding,2013)—duetotemporalandspatialvariationinitseconomicandpoliticalinstitutions.Newpolicieswerefre-quentlyintroducedtopromoteeconomicdevelopmentbysettingeconomicgrowthtargetsandmakingstatebureaucrats’careerprospectsdependentonmeetingthosetargets(Naughton,2007;WittandRedding,2013).Thisevenpushedlocalstatebureaustobecomeentrepreneurial,withlocalofficialslaunchingandactivelymanagingfor-profitbusinesses(Duckett,1998).Developmentaleffortsacceleratedin1992,whenthe‘‘ModernEnterpriseSystem’’turnedstate-ownedenterprisesintocorporations,someofwhichbecamethelistedfirmswestudied.Underthissystem,stateofficialsshuckedoffoldrolesasallocatorsandredistributorsofresourcesandforgednewrolesasregulatorsandbrokersofmarkettransactions;economicgrowthbecameanincreasinglyimportantdeterminantofbureaucrats’careeradvancement(Y.Lin,2001;Zhou,2010).Thedevelopmentalstatebifurcatedin1997,withthedoc-trineof‘‘graspingthelargeandlettinggoofthesmall’’[zhuadafangxiao]:cen-tralauthoritiesweretopromoteandmanageenterprisesinafewstrategicsectors,andlocalauthoritiesweretooverseeenterprisesinallothersectors.Muchpowertoregulatelanduse,issuebusinessandbuildingpermits,levytaxes,andimposefeesandfineswasdelegatedfromcentralstatebureaustolocalones(BoisotandChild,1996;Y.Lin,2001;Pei,2006;Naughton,2007;Landry,2008).Finally,theChinesestatebecameanembeddedone(Johnson,1982;Evans,1995)throughtwochangesthatforgedclosesocialtiesbetweenpoliticalandeconomicelites.Thepartyinformallyopeneditsrankstoentrepre-neursintheearly1990s,despitea1989ban;itformallylegitimatedthisprac-ticein2002(Dickson,2003,2008)andcreatedformallinkagestoentrepreneursthroughtrade,professional,andindustryassociations(Kennedy,2005).Insum,thecomplexityoftheChinesestate’srelationshiptotheecon-omyheightenedtheuncertaintycreatedbyitsgradualeconomictransition. 72AdministrativeScienceQuarterly62(2017)PoliticalConnectionsandEconomicActivitySocialscientistshavelongrecognizedthateconomicactionis‘‘embeddedandenmeshedininstitutions,botheconomicandnon-economic’’(Polanyi,1957:250).Thereforeanyanalysisofeconomicactionmustconsidertheimpactofsocialfactors,suchassocialnetworks,thatshapetheallocationandvaluationofresources(Polanyi,1944,1957;Granovetter,1985).Alleconomicexchanges‘‘arerifewithsocialconnections’’(Granovetter,1985:493)andsoareembeddedinsocialrelations,althoughsomearemoredeeplyembeddedthanothers.Bygeneratingtrustamongalters,embeddednessinsocialrelationsimprovescooperation,reducesuncertainty,andmakeseconomicexchangemorepredictable.Inturn,predictabilitylowerstheperceivedcostsofexchangeandincreasestheperceivedbenefits,makingmanyformsofeconomicexchangepossible,includingaccesstobankloans(Uzzi,1999),stockofferings(Stuart,Hoang,andHybels,1999),andtradebetweenmanufacturersandsub-contractors(Uzzi,1996).Socialrelationsareespeciallyimportantincontextsinwhichmarket-supportinginstitutionssuchasfreeinformationflowsandstrongruleoflawareabsentorweak:insuchcontexts,socialrelationsofferalterna-tivemechanismsforbuildingtrust,whichinturnfacilitatescooperationandreducesuncertainty,improvingthepredictabilityofexchange(GusevaandRo´na-Tas,2001).Inauthoritarianregimes,inwhichtheruleoflawisweak,socialrelationsareespeciallyvaluable.Relationsbetweeneconomicandpoliticalactorsareparticularlyimportant(Polanyi,1957:250;Lie,1991).Statebureaucratscreateinstitutionsthatfunda-mentallyshapeeconomicactivity(North,1990;Dobbin,1994;Fligstein,2001;AcemogluandRobinson,2012).Theycreateproperty-rightslawsthatlimitwhatcanandcannotbesoldandunderwhatcircumstances,whocanandcan-notsellit,andwhocanandcannotprofitfromsellingit.Theyalsocreatelawsconcerningcontractsandcompetition,andtheyusethoselawstoadjudicatedisputesbetweenfirms.Resource-dependencetheoryholdsthatuncertainty,suchasthatcreatedbyeconomictransition,canbereducedbydevelopingtiestothestate(e.g.,PfefferandSalancik,1978;Burt,1982)becausestateofficialsregulatebusinessand,inauthoritarianregimes,controlaccesstoresources.Especiallyinauthoritarianregimes,developingrelationswithpoliticalactorslikestatebureaucratscanaltertheopportunitiesandconstraintseconomicactorsface.Concretely,intransitioneconomieswithauthoritarianregimes,tieswithstatebureaucratscanreduceuncertaintybyclarifyingwhicharethemostprof-itableinvestmentopportunities,revealinghowtogainaccesstostate-controlledresources,andmakingiteasiertosecurefavorsfrombureaucrats.Asaresult,firmswithtiestostatebureaucratscanmorequicklyinvestinnewprojectsandmaximizereturnsfromthoseinvestments.Increasinguncertaintyalsomakesregulationsambiguousandfacilitatespredatorybehaviorbystateauthorities.Tieswithstatebureaucratscanclarifyregulations,therebyeasingfirms’decisionmaking,andimprovegovernmentoversight,therebyforestallingpredation.Becauseuncertaintyincreasesasmarketsdevelop,thebenefitsofbusiness–staterelationsalsoincreaseasmarketsdevelopinauthoritarianregimes.Notwithstandingthesebenefitsofbusiness–stateties,therearetwopossi-blecosts.First,suchtiesmayincreasepressureforfirmstoachievepolitical Havemanetal.73goals,suchasfullemploymentandpoliticalstability,andhindertheirpursuitofeconomicgoals.Second,ifstatebureaucratsarelesswelleducatedthanpeo-plewithexperienceinbusiness,andthuslesswellpreparedformakingbusi-nessdecisions,business–statetieswillmakefirmsmoredependentonbureaucratswithlower-qualityhumancapital.Asmarketsdevelop,uncertaintyincreasesandthestakesgrowlarger,sofirmshavetoattendmoretoeco-nomicgoals,anddecision-makingqualitybecomescrucial.Therefore,asmar-ketsdevelop,business–statetiesmighthaveincreasinglynegativeeffectsonfirmperformance.Thebalancebetweenthebenefitsandcostsofbusiness–statetiesdependsoncontext,specificallyontheextenttowhicheconomicgrowthisapoliticalgoalandonthestrengthofhumancapitalinstatebureau-craciesrelativetothatinbusiness.PoliticalEmbeddednessandFirmPerformanceinChinaBecausesocialrelations—guanxi—havelongbeencentraltoChinesesociety(Fei,1992;N.Lin,2001),focusingonthemisespeciallyappropriateforstudyingeconomicactivitythere.China’sreform-eraeconomicsystemhasbeendescribedas‘‘networkedcapitalism’’(BoisotandChild,1996)becausesocialrelationsarecommonlyusedtoaccomplisheconomicgoals.Onecommonbusiness–statesocialrelationishavingstatebureaucratsserveascompanyexecutivesordirectors.Doingsoliterallyembedspoliticalactorsineconomicinstitutions(ChenandDickson,2010)andcreatessocialrelationsthatarevar-2iouslybureaucratic,instrumental,andaffective(Michelson,2007:356).WithintheChinesestatebureaucracy,socialrelationshavelongbeenessen-tial,notonlyforcarryingoutone’sassignedtasksbutalsoforadvancingone’scareer(Zhou,2010;Zhou,Ai,andLian,2012).Bothlateralandverticalrelationshelpofficialsmanageuncertainty,gatherinformation,mobilizeresourcestogetthingsdone,andsolveunexpectedproblems.Intheseways,officialsusesocialtiesto‘‘soften’’theironcageofbureaucracy.Suchsocialrelationsbecameevenmoreimportantduringtheeconomictransitionasasystemofimpersonalrulesandincentiveswasputinplace,greatlyincreasingtheuncertaintyofficialsfaced(Yang,2005;Landry,2008;Zhou,Ai,andLian,2012).Becausethesocialtiesforgedbyworkinginthestatebureaucracyaresovaluable,theyarelikelytobemaintainedevenwhenofficialstakepositionsinbusinessconcerns.Inthefaceofmarketdevelopment,becomingpoliticallyembeddedbyhav-ingpeoplefromthestatebureaucracyserveascompanyexecutivesanddirec-torsreducesuncertaintyforfirmsbygivingthemaccesstoknowledgeabouthowthingsgetdoneinstateagencies—knowledgethattheexecutivesanddirectorsdevelopedfromworkingasbureaucrats.Thisknowledgehelpsfirmsnavigatestateagencies,clarifiestheregulatoryprocess,andimprovesfirms’understandingofstatebureaucrats’actions.Becomingpoliticallyembeddedalsomakesbusiness–staterelationshipslessformalandrootedmoreinsharedgoalsandintereststhaningovernmentalauthority(Wank,1999,2002;Y.Lin,2001;Yang,2005;Tsai,2007)andsoincreasestrustbetweenbusinessandstate.Incombination,theknowledgeandtrustgeneratedbybecomingpoliti-callyembeddedfacilitateexchangingfavorswithbureaucratsandpersuading2Havingstateauthoritiesownlistedfirms,evenifownershipispartial,embedseconomicactorsinpoliticalinstitutions,whichalsocreatescomplexsocialties(Yang,2005;FligsteinandZhang,2009). 74AdministrativeScienceQuarterly62(2017)bureaucratstolightentheregulatoryload,offerrelieffromtaxationandfees,andmakeiteasiertoacquireresourcesthatarecontrolledbythestate,notablybankloansandland-userights;theyalsohelpimprovelegaloversighttostaveoffpredatorybehavior(Y.Lin,2001;ChenandDickson,2010;NeeandOpper,2010,2012).ThedecentralizedstatesystemthatemergedinChinaprovidedmanyopportunitiesfornegotiationbetweenbusinesspeopleandbureaucrats;toprofitfromthesenegotiations,businesspeopleneededgoodrelationshipswithbureaucrats(ParishandMichelson,1996;Wank,1999,2002;PengandLuo,2000;Y.Lin,2001;BianandZhang,2014).Althoughbecomingpoliticallyembeddedcanbecostlyforfirmsintransitioneconomies,thiswasnotasmuchofaprobleminChinaasinothercountries.Startinginthe1990s,officialpoliticalgoalsbecamecloselylinkedtoeco-nomicdevelopment(Naughton,2007;WittandRedding,2013)andbecameimportantforbureaucrats’careeradvancement(Y.Lin,2001;Zhou,2010).Thisalignedstatebureaucrats’goalswiththoseoffirmsandreducedthelikelihoodthattheypressuredfirmstoachievepoliticalgoalsthatcontra-dictedeconomicones.Moreover,mostbureaucratswereCommunistPartymembers,andtheeducationlevelofpartymembersincreaseddramaticallyinthereformera(Zhou,2001).Overtime,thisshiftreducedanyhuman-capitaldifferencebetweenstatebureaucratsandothercandidatesforexecu-tiveanddirectorpositions.WeconcludethereforethatinChina,thebenefitsofbecomingpoliticallyembeddedincreasinglyoutweighedthecosts.Accordingly,wepredict:Hypothesis1(H1):Asmarketdevelopmentproceeded,becomingpoliticallyembeddedhadincreasinglypositiveeffectsonfirmperformance.ContingenciesIndustry.Economictransitionscancreatewidevariationacrossindustriesinthenumberandrelativesizeoffirms:someindustriesbecomehighlycon-centratedwithafewlargeincumbents,whileothersbecomemuchlesscon-centratedwithmanysimilarlysizedincumbents.Inhighlyconcentratedindustries,thesmallnumberofdominantfirmsincreasesthelikelihoodofcollu-sion,whichinturnreducescompetitionanduncertainty(Tirole,1988).Ifindus-trycompetitionanduncertaintyarepositivelycorrelated,thenduringeconomictransitionsinauthoritarianregimes,business–statetieswillbenefitfirmsmoreascompetitionincreases.Moreover,ascompetitionincreases,firmshavemoretogainfromsuchties—andmoretolosewithoutthem(BianandZhang,2014).InChina,asinmanyothercountries,theeconomictransitionhaddivergenteffectsonindustries.Differencesinthepaceandnatureofeconomicchangecreatedverydifferentdistributionsoffirmsintermsofsizeandownership(stateornon-state)andthusverydifferentlevelsofcompetition.Forinstance,thesteelindustryremaineddominatedbylarge,state-ownedenterprises,socompetitionwaslimited;incontrast,thetextilesindustrywasdominatedbysmall,privatelyownedfirms,andcompetitionwasintense(Kennedy,2005;McGregor,2012).Giventhisvariationinindustrystructureandcompetition,andthepositivecorrelationbetweencompetitionanduncertainty,wepredict Havemanetal.75thatasmarketdevelopmentproceeded,thebenefitsofbecomingpoliticallyembeddedweremorepronouncedinmore-competitiveindustries:Hypothesis2(H2):Thepositivejointeffectofmarketdevelopmentandbecomingpoliticallyembeddedonfirmperformancewasstrongerinmore-competitiveindustries.Firmsize.Thereareseveralimportantdifferencesbetweenfirmsofdiffer-entsizes(Scott,2003).Comparedwithsmallerfirms,largerfirmshavemoreresources.Largerfirmsalsohavemorepowertonegotiatebettertermsfromsuppliers,distributors,andcustomers,andtheybenefitmorefromeconomiesofscale;together,marketpowerandeconomiesofscalemakeiteasierforlargerfirmstoholddowncostsandcompeteonprice.Finally,largerfirmshavegreatervisibility,legitimacy,andstatus.Becauselargerfirmshavelargerresourcebasesandcanuseeconomiesofscaleandmarketpowertodrivedowntheircosts,theycanmoreeasilyrecoverfromthecostsofmistakes.Thereforetheyarebetterequippedthansmalleronestohandletheuncertainterrainofatransitioneconomyandmayhavelessneedtodeveloptiestostateauthoritiesandreduceuncertainty(andthusmis-takes):largerfirmscandealwithuncertaintybyrelyingontheirlargerresourcebasesandtheirlowercoststructures,whilesmalleronescannot.Butatthesametime,largerfirms’superiorvisibility,legitimacy,andstatusmakeiteasiertoattracttheattentionofstateauthoritiesandtoforgetieswiththem—eventhoughtheyhavelessneedofsuchtiestodealwithuncertainty.ThislogicapplieswelltofirmsinChina:largerChinesefirmshavemoreresources,benefitmorefromeconomiesofscale,havemorebargainingpowerwithexchangepartners,andhavegreatervisibility,legitimacy,andstatus.Butthereareotherreasons,specifictoChina,whylargerfirmsmayfinditeasiertocultivatetiestostateagents(PengandLuo,2000),eveniftheyhavelessneedtodoso.LargerfirmstendtobeinmorecentralpositionsinChina’spowerstructure:theyhavemoretiestobureaucratsandtiestomore-powerfulbureaucrats(Guthrie,1999;Y.Lin,2001).Largerfirmsdominatetheindustryassociationsthatserveasconduitstostateauthorities,andtheirscalehelpsthemattractbureaucrats’attention(Kennedy,2005).Andresearchoninterorga-nizationaltiesinChinashowsthattheyare‘‘weaponsoftheweak’’(Yang,1994:126).Smallerenterprisesmustleveragetheirsocialtiestosecurethebenefitsthataccrueeasilytolargerones:lessexposuretoregulations,taxes,fees,andfines(e.g.,Gold,1990;Tsai,2002);easieraccesstostate-controlledresourcesandlowerrisksofgovernmentexpropriationofassets(CullandXu,2005;LiandZhang,2007;Lietal.,2008);andbetteropportunitiestoacquireresourcesfromsourcesoutsidethestate,includingforeigninvestors(Walder,1994;Keister,2004).BothgeneralandChina-specificdifferencesbetweenlargerandsmallerfirmsleadustopredictthatasmarketdevelopmentpro-ceeds,thebenefitsofbecomingpoliticallyembeddedarelesspronouncedforlargerfirmsthanforsmallerones—eventhoughitmaybeeasierforlargerfirmstobecomepoliticallyembeddedinthefirstplace:Hypothesis3(H3):Thepositivejointeffectofmarketdevelopmentandbecomingpoliticallyembeddedonfirmperformanceisweakerforlargerfirms. 76AdministrativeScienceQuarterly62(2017)CausalMechanismsToprobecausalmechanisms,weconsidertwocommonflowsoffundsintoandoutoffirmsintransitioneconomies:bankloansandrelated-partytransactions.Accesstobankloans.Accesstobankloansfacilitatesexpansion,allowingfirmstoachieveeconomiesofscaleandmarketpowervis-a`-vissuppliersandcustomers,therebyenhancingtheircompetitivenessandimprovingtheirper-formance.Demandforsuchloansincreasesasmarketsdevelopandnewbusi-nessopportunitiesarise.Totheextentthatbanksareagentsofthestate,tieswithstateauthoritieswillbeimportantinsecuringsuchloans.Duringourstudyperiod,Chinesebankswerefirmlyunderstatecontrol(Shih,2008;NeeandOpper,2010;Allenetal.,2013).In2007,theendofourstudyperiod,theassetsofthe‘‘bigfive’’state-ownedbanksaccountedfor63percentofallbankassets.Other,smallerstate-ownedorstate-controlledbanksaccountedforanother24percent,andnon-state-controlledbanksjust13percent.ManyChinesefirmshadlimitedaccesstobankloansfromstate-ownedandstate-controlledbanks,relyinginsteadontheirownretainedearnings,whichwereoftencon-strained;forexample,theaveragereturnonassetsforlistedfirmsduringourstudyperiodwasjust3.1percent.Theonlyotheralternativeswereprivatelenders,whichchargedhigherinterestratesthanbanks(Allen,Qian,andQian,2005).Politicallyembeddedfirmscouldborrowmoneyfrombanksatdramaticallylowerinterestratesandwithfarlesscollateral(Allen,Qian,andQian,2005).Thuswepredict:Hypothesis4a(H4a):Asmarketdevelopmentproceeded,becomingpoliticallyembeddedhadincreasinglypositiveeffectsonfirms’accesstobankloans.Hypothesis4b(H4b):Accesstobankloansmediatedthejointeffectofmarketdevelopmentandbecomingpoliticallyembeddedonfirmperformance.Related-partytransactions.Intransitioneconomies,manyfirmsaremem-bersofbusinessgroups,collectionsoffirmscontrolledbythesameshareholderthatareconnectedthroughamixofequity,debt,personnel,andtrade(Granovetter,1993;KhannaandYafeh,2007).Inbusinessgroups,ownershipofanyfirmslistedonthestockmarketsissplitbetweentheircontrollingshareholdersandother(minority)shareholders,whileownershipofother,non-listedfirmsissolelyinthehandsofcontrollingshareholders.Inresponsetopressurefromcon-trollingshareholders,businessgroupmembersoftenexchangeinformation,funds,goods,orserviceswithothermembersinwhatarecalledrelated-partytransac-tions.Suchtransactionsallowcontrollingshareholderstopullresourcesoutofthelistedfirmsinwhichtheyhavepartialownershipandtransferthemtothenon-listedfirmsinwhichtheyhavefullownership,whichimpairslistedfirms’perfor-mancebutincreasesaggregatereturnstocontrollingshareholders.Despitetheharmsuchtransactionsdotolistedfirms,controllingshareholderscanpushforrelated-partytransactionsbecausetheyexerciseconsiderableinfluenceoverlistedfirms’strategiesandoperations.Moreover,whentheruleoflawisweak,control-lingshareholdersfacefewexternalcorporategovernancemechanisms,suchaslawsuitsbyminorityshareholders,thatwoulddetersuchtransactions(ShleiferandVishny,1997;LaPortaetal.,1998).Pressurestoengageinrelated-party Havemanetal.77transactionsincreaseasmarketsdevelopbecausetheperformanceofbusiness-groupmembersoftendeterioratesduetoincreasedcompetitionanduncertainty.Listedfirmscandefendthemselvesfrompressurestoengageinrelated-partytransactionsbyforgingrelationshipswithstatebureaucrats.Theweakruleoflawinauthoritarianregimesmeansthejudiciaryisnotindependentofthestatebureaucracy,andstatebureaucratshavediscretioninapplyingregula-tions.Relationshipswiththosebureaucratscanprovideanalternativetotheruleoflaw(GusevaandRo´na-Tas,2001):suchrelationshipsbuildtrustbetweenfirmsandstateauthoritiesandsomakeiteasiertosecureenforce-mentofanyformalchecksandbalancesregulatorshaveimposed,includingthoseconcerningcontrollingshareholders’behavior.Business–statetiescanprotectfirmsagainstpressurebycontrollingshareholderstoengageinrelated-partytransactions;asmarketsdevelopandthesepressuresmount,suchtiesincreasinglyimprovefirmperformance.InChina,mostlistedfirmsaremembersofbusinessgroups(Keister,2000;Zheng,2013)withinwhichcontrollingshareholdersexertaninfluencethatexceedstheirownershipsharebecausetheruleoflawisweak(Clarke,2003).Controllingshareholderscanpushforrelated-partytransactionsevenwhentheyhavelessthanmajorityownershipstakes(FismanandWang,2010)becausetheyfacefewexternalgovernancemechanismsthatwoulddetersuchtransactions(ZhengandKim,2011).Ouranalysisfocusesonrelated-partyloans(andloanguarantees)madebylistedfirmstootherbusiness-groupmembers.BecauseChineselistedfirmsareusuallythefinanciallystrongestintheirbusi-nessgroup(Zheng,2013),controllingshareholdersoftenpressurethemtosup-portotherbusiness-groupmembersfinancially.Listedfirms’financialstrengthandtheirtransparentfinancialstatements,whichfollowtightlyregulateddisclo-surerequirements,makeitpreferableforthemtoguaranteethird-partyloanstootherbusiness-groupmembers(FismanandWang,2010),butdoingsoputslistedfirmsatriskofrepayingtheseloansifborrowersdefault.Becauseinterestonmanyloansisnotpaidandmanyloansareneverrepaid,thesetransactionsfunnelfinancialresourcesoutoflistedfirmsandimpairtheirperformance(FismanandWang,2010;Jiang,Lee,andYue,2010;Jia,Shi,andWang,2013).BecomingpoliticallyembeddedcanhelpChineselistedfirmsdefendthem-selvesagainstpressurefromcontrollingshareholderstomakerelated-partyloans.Chineseregulatorshavebeguntoviewtheseloanswithgreatsuspicionandhaveimposedincreasinglystringentcontroloverthem.Forexample,onAugust28,2003,theChinaSecuritiesRegulatoryCommissionissuedaregula-tiondemandingplansfromlistedfirmsfordecreasingrelated-partyloansby30percentperyear.Listedfirmsthatbecomepoliticallyembeddeddevelopclosertiestoandgreatervisibilitywithstateofficials,sotheycanmoreeasilyper-suadebureaucratstoenforceregulationsthatprotectfirms’interests(Li,3Meng,andZhang,2006).Pressurestomakerelated-partyloansincreasedas3Previousresearchonbusiness–statetiesinChinademonstratedthatpoliticalembeddednessoffersprotectionacrossanarrayofsettings:lawyersreceivedfavorsfromcourtofficials(Michelson,2007);firmsweremorelikelytousethecourtsystembecausetheirpoliticaltiesgavethemadvantagesovertheiropponents(AngandJia,2014);andwhenfirmsexperiencedinfringe-mentoftheirpropertyrightsbystateofficials,theycouldappealtohigher-levelofficialsforcorrec-tiveaction(Li,Meng,andZhang,2006).Thus,althoughbusiness–statetieshavegenerallybeenlinkedtocorruption,wepredictthattheymaysometimesfendofftunneling(seealsoMichelson,2007). 78AdministrativeScienceQuarterly62(2017)marketdevelopmentproceeded,sobecomingpoliticallyembeddedwasincreasinglyimportanttoresistthosepressures.Inturn,protectionfromrelated-partyloansimprovedfirmperformance.Thereforewepredict:Hypothesis5a(H5a):Asmarketdevelopmentproceeded,becomingpoliticallyembeddedhadincreasinglynegativeeffectsonfirms’related-partyloans.Hypothesis5b(H5b):Related-partyloansmediatedthejointeffectofmarketdevelopmentandbecomingpoliticallyembeddedonfirmperformance.METHODSampleWestudyChinesefirmslistedonthedomesticstockexchanges,whichareamongthelargestfirmsinChinaandwhichdominatemostindustrialsectors.Muchmore,andmore-reliable,informationisavailableonlistedfirmsthanonnon-listedfirms,whichmakesthemastrategicsiteforresearchonChina’seconomictransition.Ouranalysisbeginsin1992,theyearaftertheChinesestockmarketsopened,andendsin2007whenChinafinallyadoptedlawstofullyproclaimtherightstoprivatepropertyinbusiness.Usingthisendpointavoidshavingtheanalysisconfoundedbytheglobalfinancialcrisisthateruptedin2008andthesubsequentdeclineintheChinesestockmarket.Thenumberoflistedfirmsincreasedovertime,from26in1992to801in1999and1,371in2007.Thefulldatasetcontains11,145firm-yearrecords.Alllistedfirmswerepoliticallyembeddedtosomeextentbecausetheyhadtoreceivestateapprovaltolistonthedomesticstockexchangesinthefirstplace(Zheng,2013).Upto1999,thestatesetstrictquotasforlisting,andfirmscompetedvigorouslyforthosequotas;startingin1999,thestatedroppedquo-tasandallowedinvestmentbankstonominatefirmstobecomelisted,butthestatestillneededtoapprovelistings.Thatthefirmswestudyhadtobepoliti-callyembeddedtogetlistedshouldreduceourabilitytofindanyimpactofthespecificformofpoliticalembeddednesswestudy,renderingouranalysescon-servativetestsofourarguments.ToconsiderwhetherourargumentsgeneralizetootherfirmsinChina,weconductedasupplementaryanalysisofnon-listedfirms,usingdatafromaseriesofsurveysdonebythePrivatelyOwnedEnterprisesResearchProjectTeam,agroupofscholarsinawidearrayofgovernmentalandnon-governmentalorganizations,includingtheChineseAcademyofSocialScienceandtheAll-ChinaFederationofIndustryandCommerce.Ineachsurveyyear(1995,1997,2000,2002,2004,and2006),theresearchteamusedstratifiedrandomsamplingtogeneratenationallyrepresentativesamplesofprivatelyownedfirmsacrossallChineseprovincesandallindustries.Theaveragesam-4plesizeeachyearwas1,385firms.Thesedatahavethreelimitations.First,theyconsistofrepeatedcross-sections,ratherthanpanels,soourabilitytopin-pointcausationislimited.Second,theymeasureindustrycoarsely,sowecouldnotassessthatcontingency.Third,theycontainnodataonbankborrowingor4Weexcludedfromanalysistherarelistedfirmsinthisdataset,whichconstituted0.2percentofthetotal. Havemanetal.79related-partyloans,sowecouldnottestthesecausalmechanisms.Still,thisanalysisallowedustotestourcentralargumentandthecontingenteffectoffirmsize.DataSourcesandMeasuresPoliticalembeddedness.Wehand-collectedthere´sume´sofeachlistedfirm’schiefexecutiveofficer(CEO),allothertopexecutives(allthosewith‘‘Chief’’intheirtitles),andalldirectorsfromthewebsiteofSina(finance.sina.com.cn),whichpublishescomprehensiveinformationonChineselistedfirms.Becauselistedfirms’executivesanddirectorschangedovertime,wecollectedthesedataforeveryyeareachfirmwaslistedonadomesticstockexchange.Wescrutinizedeachperson’sre´sume´todeterminewhetherheorshehadservedasabureaucratand,ifso,atwhatrank(chiefofficer,deputy,etc.)andatwhatlevelintheofficialhierarchy.Thishierarchyhassixmainlevels:ministry(bu),department(ju),division(chu),section(ke),staffmem-ber(keyuan),andclerk(banshiyuan).Wecodedafirm’spoliticalembedded-nessasabinaryindicatorvariablesettooneinyearswhenanexecutiveordirectorofthefocalfirmhadservedasthechiefofficerordeputychiefoffi-ceratthedivision(chu)levelorabove,andzerootherwise.ThisthresholdiscommonlyusedinstudiesofChinesebureaucrats(e.g.,Zhou,2000;Walder,2004).Lower-levelofficialsarenotfundedthroughthecentralfiscalsystem,sothedivisionlevelindicatesmembershipinthepoliticalelite,whichin1998includedroughlyhalfamillionindividuals(Walder,2004:195)outofatotalpopulationof1.25billionthatyear—justfouroutof10,000Chinesepeople.Becausethevastmajorityofhigh-levelbureaucratsareChineseCommunistPartymembers,ourmeasurecapturesasecondaspectofpoliticalembeddedness.Forrobustnesschecks,weconstructedtwoothermeasures:thenumberandtheproportionofexecutivesanddirectorswhohadservedasthechiefofficerordeputychiefofficeratthedivision(chu)levelorabove.Tocodepoliticalembeddedness,wehiredfiveresearchassistantseachattwouniversities,oneinJiangxiProvinceandoneinBeijing.Eachschool’sgroupofstudentsindependentlycodeddataaccordingtodetailedcodinginstructions.Wecomparedcodesgeneratedbythetwogroups,andfor96.3percentofmanager-yearobservations,theircodeswereinagreement.Fortheremaining3.7percentofobservations,wehadbothgroupscheckthereasonsunderlyingtheircodingdecisionsandworkedwiththemtoresolvealldisagreements.Anyrelationshipwefindbetweenbecomingpoliticallyembeddedandperfor-mancemaybeepiphenomenal:ifstateofficialsservingasexecutivesanddirectorslaterreturntothestatebureaucracy,itmaybethatthoseofficialsareplacedinlistedfirmstopolishtheirre´sume´s,andthehigher-levelofficialswhoorchestratethesemoveswillprotectpoliticallyembeddedfirms.Thisisunlikelytoexplainourfindingsintheirentiretybecauseonlyasmallfractionofofficialsservingasexecutivesanddirectorslaterreturnedtothestatebureaucracy.Amonglistedfirms,2.5percentofCEOsmadesuchmovesbetween2005and2011(Caoetal.,2014).Amonglistedfirmscontrolledbythecentralstate,such 80AdministrativeScienceQuarterly62(2017)movesconstituted11percentofallmovesoftopmanagersanddirectorsbetween2004and2007(LinandMilhaupt,2013).Suchmovesarerareinpartbecausebureaucratsusuallymoveintofirmsnearretirementage(60yearsformalecivilservantsand55yearsforfemaleones).Still,therearesomereturns,andreturnratesarelikelytobehigherforlargerfirmsandforfirmsownedby5thecentralstate.Tohandlethispossibility,wedidfourthings.First,wematchedtreated(politicallyembedded)andcontrol(unembedded)firmsalongkeyobservabledimensions,includingstateownership.Second,weincludedfirmfixedeffects,whichmeansweanalyzedchangesinpoliticaltieswithinfirmsovertime.Ownershipisconstant,andifowners’strategiesforappointingexecutivesanddirectorsremainedconstant,differencesbetweenstate-controlledandnon-state-controlledfirmsintermsofwhetherpoliticallytiedleaderswereappointedbythestatearecapturedbythefirmfixedeffects.Third,wecon-ductedancillaryanalysesthatdroppedfirmsownedbythecentralstatefromthedataset;theseresultswerehighlyconsistentwiththeresultsshownhere.Fourth,weanalyzedstate-controlledandnon-state-controlledfirmsseparatelyandfoundnostatisticallysignificantdifferenceininteractionsbetweenmarketdevelopmentandpoliticalembeddedness.Marketdevelopment.Wemeasuredtheproportionsoflaborandcapitalallocatedthroughmarketsratherthanbythestate,basedonhowlaborandcapitalweredividedbetweenstate-ownedandnon-state-ownedfirms.WefollowedtheNationalBureauofStatistics’definitionofstate-ownedfirmas(a)thoseinwhichthestateownsmorethan50percentofsharesand(b)thoseinwhichthestateownslessthan50percentofsharesbutthestateisthelargestshareholderorthecontrollingshareholder(HolzandLin,2001).Thefirstmeasure,marketdevelopment(employment),istheloggedpercentageofthelaborforceworkinginnon-state-ownedfirmseachyear.Thesecondmeasure,marketdevelopment(investment),istheloggedpercentageoffixed-assetinvestmentsinnon-state-ownedfirmseachyear.Bothvariablesaremeasuredforthenationasawhole,asmanylistedfirmsoperateinmultiplelocations.BothwerebasedondataintheChina6StatisticalYearbook(variousyears).Weloggedthembecausetheywereright-skewed.Inresultsnotshownheretosavespace,weusedanindicatorofmarketdevelopmentthatwasbasedonstatepolicyandthecreationofmarket-5XueguangZhou,personalcommunication,April29,2015.6Startingin1998,thedefinitionoftheshareoflaborandinvestmentinstate-ownedfirmswaschangedtoincludepartofthelaborandinvestmentinnon-state-controlledfirms,proportionatetostateownershipstakesinthosefirms(HolzandLin,2001).Becausetheshareoflaborandinvest-mentinstate-ownedfirmsthatcouldbeapportionedtonon-stateshareholders,basedontheirownershipstakes,wasnotdeducted,measuresofmarketdevelopmentstartingin1998wereartifi-ciallyreducedrelativetomeasuresforprioryears.Thischangeshouldmakeitmoredifficultforustofindsupportforourhypotheses. Havemanetal.81supportinginstitutions,suchasfreeinformationflowsorlawsgoverningprop-ertyrights,contracts,andcompetition,ratherthanmarketsize.Theyear2003wasamilestoneinthedevelopmentofsuchinstitutions,spurredbyChina’saccessionintotheWorldTradeOrganizationinDecember2001(e.g.,Yang,72005).Firmandindustryvariables.WeobtainedmostdataonfirmsfromtheGuoTaiAnInformationTechnologyCompany(GTA,alsocalledChinaSecuritiesMarketandAccountingResearch),aHongKongfirmthathasdevel-opeddatabasesontheChinesebankingindustry,stockmarket,andeconomyforacademicandindustryresearchers.WeusedGTA’sChinaStockMarketTradingDatabase.WealsoobtainedsomedatafromWindInformationCo.,aleadingfinancialdatafirminChina.Tomeasurefirmperformance,weusednetreturnonnetassets(ROA),ameasurethatiscomparableacrossindustriesandacrossfirmsoperatingondif-ferentscales.Forrobustnesschecks,wealsocalculatednetreturnonequity(ROE),whichrevealshowmuchprofitthecompanygenerateswiththemoneyshareholdersinvested.GiventheriseoftheshareholdervaluemodelintheUnitedStates(LazonickandO’Sullivan,2000;FligsteinandShin,2007)andthenumberofAmericaneconomicadvisorsworkinginChina,thismeasuremaybehighlysalienttoChineseexecutivesanddirectors.ButinChina,accountingforassetsislessambiguousthanaccountingforequity(PengandLuo,2000),sowefocusprimarilyonROA.Tomeasurecompetition,weaggregateddataonfirmstotheindustrylevel.Wedefinedindustryusingtwo-digitChineseStandardIndustrialClassificationcodes.WeusedtheHerfindahl–Hirschmanindexofconcentration,calculatedusingthemarketshare,basedonsales,ofalllistedfirmsineachindustryeachyear.Theindexofconcentrationisaninversemeasureofcompetition:highervaluesindicatelowerlevelsofcompetition.Mediatingvariables.Wefirstassessedeachfirm’saccesstobankloansbycalculatingitsborrowingratio,theratioofbankborrowingtototalassets.Scalingbyassetsmakesthisvariablecomparableacrossfirmsofdifferentsizes.Wethenexaminedrelated-partyloansbycalculatingforeachyearthetotalvalueofloansandloanguaranteesprovidedbythelistedfirmtothecon-trollingshareholderortootherfirmscontrolledbythatshareholder,minusthevalueofloansrepaidtothoseentitiesandloansorloanguaranteesprovidedtothelistedfirmbythoseentities.Wescaledthisvariablebythelistedfirm’stotalassetstomakeitcomparableacrossfirmsofdifferentsizesandcreatedtherelated-partyloanratio.Dataonrelated-partyloansbecameavailableonlyin1998,whenanewreportingrequirementwasputinplace,soouranalysisofthiscausalmechanismislimitedto1998onward.Fourconditionsmustbepresentforavariabletobeamediator:(1)theinitialexplanatoryvariablemustbecorrelatedwiththeoutcomevariable,(2)theinitial7Seealsomediaaccounts,forexample,‘‘Theyearof2003:AfirststeptoimproveChina’seco-nomicreforms,’’XinhuaNewsAgency,November28,2011(http://news.xinhuanet.com/newscenter/2003-11/28/content_1202985.htm)and‘‘Milestonesofreform:30yearsof3rdPlenums,’’China.org.cn,October8,2008(http://www.china.org.cn/china/reform-opening-up/2008-10/08/content_16645095_2.htm). 82AdministrativeScienceQuarterly62(2017)explanatoryvariablemustbecorrelatedwiththemediator,(3)themediatormustaffecttheoutcomevariable,and(4)theeffectoftheexplanatoryvariableontheoutcomevariableinamodelthatincludesthemediatorshouldbesignifi-cantlysmallerthaninamodelthatdoesnotincludethemediator(BaronandKenny,1986).Asweshowbelow,wetestedallfourconditions.Toassessthestatisticalsignificanceofthechangeintheeffectoftheexplanatoryvariable,weusedthemethodderivedbySobel(1982).Controlvariables.Wecontrolledforothervariablesthatmayaffectfirmperformance,accesstobankloans,andrelated-partyloans.First,wecon-structedanindicatorofstateownership,codedoneifthefirmwascontrolledbyastateagencyandzerootherwise.InChina,listedfirms’controllingshare-holdersexerciseconsiderablecontroloveroperationsthroughtheirpowertomakeappointmentstoboardsofdirectors,evenifthoseshareholdershavesig-nificantlylessthanmajoritystakes(Clarke,2003;FismanandWang,2010).Differenttypesofcontrollingshareholdersinfluencenotonlyfirmperformancebutalsowhatfirmscangainfromtheirpoliticalembeddedness(Walder,1996).State-ownedfirmscanusetheirformalpoliticalconnectionstomaneuverformoreadvantageousmarketpositions(e.g.,Calomiris,Fisman,andWang,2010).Non-state-ownedfirmscanuseinformaltiestobureaucrats,suchashavingbureaucratsserveasdirectorsorexecutives,togetprotectionfromcompetitionandaccesstofinancingandindustryinformation(e.g.,Luo,2003;LiandZhang,2007),soinformaltiesmaybeespeciallyrewardingfornon-state-ownedfirms.Second,wecontrolledforfirmsizeusingassets,loggedtonor-malizethedistribution,becauselargerfirmsaremorelikelytohavepoliticalties,althoughsmalleronesmaybemorelikelytoactivelyusethem(Guthrie,1999;Y.Lin,2001).Inrobustnesschecks,weusedanalternativemeasureofsize,basedonsales,againloggedtonormalizethedistribution.ModelSpecificationandEstimationFirmsthatbecomepoliticallyembeddedmaydifferconsiderablyfromthosethatdonot.Forexample,firmsinlesscompetitiveindustriesmaybemorelikelytobecomepoliticallyembeddedandmorelikelytoperformwellasmar-ketsdevelopindependentoftheirbecomingpoliticallyembedded.Ifso,regres-sionanalyseswillbebiasedbyselectionintothe‘‘treatmentcondition’’(becomingpoliticallyembedded)ratherthanthe‘‘controlcondition’’(notbecomingpoliticallyembedded).Toputitsimply,industrycompetitivenessmayexplainanyobservedassociationbetweenthetreatmentandthedepen-dentvariable.Propensity-scorematchinghelpsalleviatesuchconcernsbymatchingfirmsinthetreatmentandcontrolconditionsonobservableconfounds,thuseliminat-ingspuriousresultsduetothoseconfounds(RosenbaumandRubin,1983).Toimplementthistechnique,researcherspredictselectionofcasesintothetreat-mentconditionusingasetofvariablesthatarenotaffectedbythedependentvariable,calculateeachcase’spredictedprobabilityofbeingselectedintothetreatmentcondition—its‘‘propensityscore’’forexperiencingthetreatment—andmatchcasesinthetreatmentandcontrolconditionsonthebasisoftheirpropensityscores.Forourstudy,wefirstestimatedalogisticregression Havemanetal.83predictingfirmsbecomingpoliticallyembeddedusingfirmsize,borrowingratio,stateownership,industry,andyear.Wethencalculatedthepropensityscoreforeachfirmeachyear—itspredictedprobabilityofbeingpoliticallyembedded.Nextweconstructedasubsetoffirmsthatdidnotbecomepoliticallyembeddedbutthathadsufficientlyhighpropensityscores(the‘‘matchingsam-ple’’)thattheyresembledthesetoffirmsthatdidbecomepoliticallyembeddedinallobservablerespects,exceptforreceivingthetreatment.Todeterminewhichpropensityscoreswerehighenoughforinclusioninthematchingsam-ple,weusednearest-neighbormatchingwithoutreplacement.Webeganbysortingthefirmsinthetreatmentconditionrandomlyandthenmatchedeachwiththeclosestfirminthecontrolcondition.Tomatch,weusedacaliperof0.25standarddeviations.Allmatchesnotwithinthiscaliperweredropped(317firms:201thatbecamepoliticallyembeddedand116thatdidnot),sothesam-pleanalyzedincluded1,054firms.Totestthequalityofthematchingprocess,weassessedwhetherthetwogroupsoffirmswerestatisticallysimilarintermsofobservables.Althoughfirmsthatbecamepoliticallyembeddedhadslightlyhigherpropensityscoresthanthosethatdidnot,thedistributionsoverlappedgreatly.Forstateowner-ship,firmsize,andborrowingratio,thepercentagebiasbetweenthetwogroupswassmallerthanthecommonlyacceptedthresholdof5percent(RosenbaumandRubin,1983);moreover,ttestsofdifferencesbetweenthetwogroupswerenonsignificant(p>.05)fortheborrowingratio,stateowner-ship,andfirmsize.Forallindustryandyearindicatorvariables,thepercentagebiasbetweenthetwogroupswassmallerthan5percent.Differencesbetweenthetwogroupswerenotsignificantlydifferentfortheoverwhelmingmajorityofindicatorvariables:ttestsyieldedp>.05for92percentofindustryindica-torsand97percentofyearindicators.Theseresultssuggestthattheoverallmatchisgood:afterpoolingthetwogroups,becomingpoliticallyembeddedcanberegardedasexogenoustotheextentthatwehaveruledoutselectiononobservables.Ourdependentvariablesarecontinuous,soweestimatedordinaryleast-squaresregressionsonthematchedsamples.Welaggedallexplanatoryvari-ablesbyoneyear.Becausewehavemultipleobservationsoneachfirm,datapointsarenotindependent.Todealwiththis,weincludedfirmfixedeffectsinallmodelsandclusteredstandarderrorsbyfirm.Includingfirmfixedeffectsmeansthatwecomparedwithin-firm,over-timechangesintheimpactofpoliti-calembeddedness;doingsoeliminatesconcernsabouttime-invariantunobser-vabledifferencesbetweenfirmsthatmayyieldspuriousresults(Halaby,2004).Thesemodelsofferconservativetestsofourhypothesesbecausetheydonotmodelvariationbetweenfirms.Thenextstepwastoaddyearfixedeffects,whicheliminatedconcernsaboutallunobservedover-timedifferences(e.g.,unemploymentorinflation),notjustmarketdevelopment.Becausemarketdevelopmentismeasuredatthenationallevel,thesemodelsdonotincludethemaineffectofmarketdevelopment—thatcoefficientissweptawaybytheyearfixedeffects.Modelswithbothfirmandyearfixedeffectsaremoreconservativethanmodelswithjustfirmfixedeffects:theformercapturenotjustover-timevariationinmarketdevelopmentbutalsoover-timevariationinothernational-levelfactors,someofwhichmaybecausedbymarketdevelopment. 84AdministrativeScienceQuarterly62(2017)Totesthypothesis1,weinteractedmarketdevelopmentwiththepoliticalembeddednessdummy.Totesthypotheses2and3,weconductedsubsampleanalyses.Wetestedhypothesis2bycomparingregressionsacrosssubsam-plesoffirmssplitbythelevelofcompetitiontheyfacedintheirprimaryindus-try.Wecodedanindustryasmorecompetitiveifconcentrationwasbelowthemedianacrossallindustriesinthefocalyearandlesscompetitiveifconcentra-tionwasabovethemedian.Wetestedhypothesis3bycomparingregressionsacrosssubsamplessplitbasedonfirmsize.Wedefinedafirmaslargeifitstotalassetswereabovethemedianforsizeinthefocalyearandsmallifitstotalassetswerebelowthemedian.Weusedsubsampleanalysisbecauseitisflexible,asitallowsothercovariatestodifferbetweenfirmsinmoreandlesscompetitiveindustriesandbetweenlargeandsmallfirms.Bycontrast,estimat-ingasinglemodelconstrainsthecoefficientsofthesecovariatestobethesameforallfirms,whichmaynotbejustifiable.Subsampleanalysisisalsomorestraightforwardtointerpretthanthree-wayinteractions.RESULTSTrendsoverTimeandacrossSpaceFigure1plotsthenumberoflistedfirmseachyearandthenumbereachyearthathaveexecutivesordirectorswhoworkedasbureaucratsatthedivision(chu)leveloraboveinthecadrehierarchy.Itshowsastrikingpattern:overtime,listedfirmswereincreasinglylikelytohaveformerbureaucratsasexecu-tivesanddirectors.In1992,only27percentoflistedfirms(7of26)werepoliti-callyembeddedinthisway.Thenumberroseto54percentin1998(378of704),thento69percentin2002(721of1,051),beforedecliningslightlyto66percentin2007(908of1,368).Figure1.Numberoflistedfirmsandnumberofpoliticallyembeddedlistedfirms.1,400NumberofListedFirmsNumberofPoliticallyEmbeddedListedFirms1,2001,0008006004002000199119931995199719992001200320052007 Havemanetal.85Bothmeasuresofmarketdevelopmentincreasedsteadilyduringourstudyperiod.In1992,non-state-ownedfirmsaccountedfor16percentofemploy-mentand1.9percentofassetsinvested.By2007,non-state-ownedfirmsaccountedfor68percentofemploymentand41percentofassetsinvested.Ateverypointintime,marketconcentrationvariedtremendouslyacrossindus-tries.Amongthemostcompetitive(leastconcentrated)industriesweremachineryandindustrialequipment,metalandnon-metalproducts,andtextilesandclothing,withHerfindahl–Hirschmanindexesof.026,.054,and.056,respectively,onaverageacrossourstudyperiod.Socialservicesandfinanceweretheleastcompetitive(mostconcentrated)industries,withindexesof.842and.933,respectively.TestingHypothesesTable1presentsunivariatestatisticsandcorrelations.Noneofthecorrelationsishighenoughtoraiseconcernsaboutmulticollinearity,exceptbetween(a)thetwomeasuresofmarketdevelopmentand(b)thetwointeractionsbetweenpoliticalembeddednessandthesemeasuresofmarketdevelopment,butnei-therpairofvariablesisincludedinthesameregressionmodel.Table1.UnivariateStatisticsandBivariateCorrelations*NumberofVariableMeanS.D.Min.Max.observations1231.Returnonassets.031.060–.360.17910,8612.Borrowingratio.499.195.000.92010,704.3673.Related-partyloanratio.057.106.000.5149,152–.154–.3064.Politicalembeddedness.609.488.0001.00010,861.001–.077.0025.State-controlledfirm(yes=1).785.411.0001.00010,861.036.105–.092À6.Industrycompetition(concentration).462.499.0001.00010,861.174–.037–.0267.Firmsize(logassets,RMB)20.903.98114.93727.11110,861.032–.186.1338.Marketdevelopment(employment)–1.055.190–1.338–.77210,860–.166–.285.2869.Marketdevelopment(investment)–.565.168–1.113–.33110,860–.169–.283.28510.Marketdevelopment×Political–.627.523–1.338.00010,860–.027.022.051embeddedness(employment)11.Marketdevelopment×Political–.331.294–1.113.00010,860–.044–.012.081embeddedness(investment)Variable456789105.State-controlledfirm(yes=1).105À6.Industrycompetition(concentration).126.1737.Firmsize(logassets,RMB).147.141.4068.Marketdevelopment(employment).120–.174–.120.3149.Marketdevelopment(investment).119–.174–.123.313.98710.Marketdevelopment×Political–.963–.132–.152–.076.094.093embeddedness(employment)11.Marketdevelopment×Political–.905–.143–.164–.030.217.222.985embeddedness(investment)*Thistablepresentsstatisticsonthecombinedtreatmentandcontrolsamplesthatwerecreatedbypropensity-scorematching.ÀLowvaluesindicatestrongcompetition. 86AdministrativeScienceQuarterly62(2017)Table2.TheImpactofPoliticalEmbeddednessonFirmPerformance(ROA),ContingentonMarketDevelopment(Employment)*(5)(6)(1)(2)(3)(4)More-Less-(7)(8)AllAllAllAllcompetitivecompetitiveSmallLargeVariablefirmsfirmsfirmsfirmsindustriesindustriesfirmsfirms••••••••Politicalembeddedness–.001.039.002.032.038.009.047.013(.002)(.011)(.002)(.010)(.014)(.015)(.019)(.014)••••••Marketdevelopment–.037–.060(.006)(.009)•••••••Marketdevelopment×.036.027.031.013.038.014Politicalembeddedness(.010)(.009)(.014)(.014)(.016)(.013)••••••••••••••••••••••••Firmsize(log–.021–.021–.019–.019–.038–.015–.038–.019assets,RMB)(.002)(.002)(.002)(.002)(.003)(.003)(.004)(.003)••••••••••••State-controlled–.010–.011–.010–.011–.015–.007–.013–.009firm(yes=1)(.003)(.003)(.003)(.003)(.004)(.006)(.005)(.006)Chowtest(Prob>F)=.07(Prob>F)=.05Numberofobservations10,86010,86010,86010,8605,8475,0135,3645,496AdjustedR-squared.248.251.290.291.293.381.276.363•p<.05;••p<.01;•••p<.001.*Robuststandarderrors(clusteredonfirms)areshowninparentheses.Marketdevelopmentisthe(logged)percentageofthelaborforceemployedinnon-state-ownedfirmseachyear.More-competitiveindustriesarethosewithbelow-medianconcentrationindexesinthefocalyear.Smallfirmsarethosewithbelow-medianassetsinthefocalyear.Allmodelsincludefirmfixedeffects,andmodels3–8includeyearfixedeffects.Firmperformance.Table2presentsresultsonreturnonassets(ROA)usingmarketdevelopment(employment).Model1reportstheresultsofthebaselinemodelwithallexplanatoryvariablesexcepttheinteractionbetweenmarketdevelopmentandpoliticalembeddedness.Itshowsasignificantnega-tiveeffectofmarketdevelopmentandanear-zeroeffectofpoliticalembedded-ness.Italsoshowsthatlargerfirmsandstate-controlledfirmsperformworsethansmallerandnon-state-controlledones.Model2addstheinteractionbetweenpoliticalembeddednessandmarketdevelopmentandshowsasignifi-cantpositiveeffect,whichindicatesthatbeingpoliticallyembeddedincreas-inglyenhancedfirmperformance,supportinghypothesis1.Model3addsyearfixedeffectsanddropsthemaineffectofmarketdevel-opment.Model4addstheinteractionbetweenmarketdevelopmentandpoliti-calembeddednessandshowsapositiveandstatisticallysignificanteffect,whichindicatesthatbecomingpoliticallyembeddedincreasinglyenhancedfirmperformance,supportinghypothesis1.Thiseffectisquitelarge.Whenmarketdevelopmentwasatitsmean,becomingpoliticallyembeddedincreasedpre-dictedROAby.004.Whenmarketdevelopmentwasonestandarddeviationabovethemean,theincreasewas.009.Thusasmarketdevelopmentincreasedfromthemeantoonestandarddeviationabovethemean,becomingpoliticallyembeddedincreasedROAbyanadditional.005—a16-percentincreasefromthesample-averageROA.Totesthypothesis2,models5and6comparemoreandlesscompetitiveindustries,basedonamediansplitofthesample.Theinteractionbetweenpoliticalembeddednessandmarketdevelopmentispositiveandstatistically Havemanetal.87Table3.TheImpactofPoliticalEmbeddednessonFirmPerformance(ROA),ContingentonMarketDevelopment(Investment)*(5)(6)(1)(2)(3)(4)More-Less-(7)(8)AllAllAllAllcompetitivecompetitiveSmallLargeVariablefirmsfirmsfirmsfirmsindustriesindustriesfirmsfirms•••••••Politicalembeddedness–.001.017.002.020.027.004.031.006(.002)(.007)(.002)(.007)(.009)(.010)(.012)(.009)••••••Marketdevelopment–.036–.055(.007)(.010)••••••Marketdevelopment×.031.030.038.014.043.014Politicalembeddedness(.011)(.010)(.016)(.015)(.018)(.015)••••••••••••••••••••••••Firmsize(logassets,RMB)–.022–.022–.019–.019–.038–.015–.038–.019(.002)(.002)(.002)(.002)(.003)(.003)(.004)(.003)••••••••••••State-controlledfirm(yes=1)–.009–.010–.010–.011–.015–.007–.013–.009(.003)(.003)(.003)(.003)(.004)(.006)(.005)(.006)Chowtest(Prob>F)=.16(Prob>F)=.06Numberofobservations10,86010,86010,86010,8605,8475,0135,3645,496AdjustedR-squared.2460.2470.2900.2910.2930.3810.2760.363•p<.05;••p<.01;•••p<.001.*Robuststandarderrors(clusteredonfirms)areshowninparentheses.Marketdevelopmentisthe(logged)percentageofallfixed-assetinvestmentsmadebynon-state-ownedfirmseachyear.More-competitiveindustriesarethosewithbelow-medianconcentrationindexesinthefocalyear.Smallfirmsarethosewithbelow-medianassetsinthefocalyear.Allmodelsincludefirmfixedeffects,andmodels3–8includeyearfixedeffects.significantinmodel5(more-competitiveindustries)andpositivebutnonsignifi-cantinmodel6(lesscompetitiveindustries),andthemagnitudeoftheinterac-tioneffectisgreaterinmodel5thaninmodel6.ButaChowtestshowsthatthedifferencebetweentheinteractioneffectsisnotstatisticallysignificant(p=.07),whichfailstosupporthypothesis2.Totesthypothesis3,models7and8comparesmallandlargefirms,againbasedonamediansplitofthesample.Theinteractionbetweenpoliticalembeddednessandmarketdevelopmentispositiveandstatisticallysignificantinmodel7(smallfirms)andpositivebutnonsignificantinmodel8(largefirms),andthemagnitudeoftheinteractioneffectisgreaterinmodel7thaninmodel8.AChowtestshowsthatthedifferencebetweentheinteractioneffectsisstatisticallysignificant(p=.05).Thissupportshypothesis3andindicatesthatasmarketdevelopmentproceeded,becomingpoliticallyembeddedbecamemorebeneficialonlyforsmallerfirms.Table3replicatestable2,substitutingmarketdevelopment(investment),andtheresultsarelargelyconsistentwiththoseintable2.Inmodel1,theeffectofthismeasureofmarketdevelopmentisnegativeandsignificant.Inmodel2,theinteractionbetweenpoliticalembeddednessandmarketdevelop-mentisagainpositiveandsignificant,indicatingthatbeingpoliticallyembeddedbecameincreasinglybeneficial,whichfurthersupportshypothesis1.Asintable2,model3addsyearfixedeffectsanddropsthemaineffectofmarketdevelopment.Model4addstheinteractionbetweenpoliticalembeddednessandmarketdevelopment,whichisagainpositiveandsignificant,furthersup-portinghypothesis1.Again,theeffectisquitelarge.Whenmarket 88AdministrativeScienceQuarterly62(2017)developmentwasatitsmean,becomingpoliticallyembeddedincreasedROAby.003.Whenmarketdevelopmentwasonestandarddeviationabovethemean,theincreasewas.008.Thus,asmarketdevelopmentincreasedfromthemeantoonestandarddeviationabovethemean,becomingpoliticallyembeddedincreasedROAbyanadditional.005—a16-percentincreasefromthesample-averageROA.Models5and6comparemoreandlesscompetitiveindustries,basedonamediansplitofthesample,totesthypothesis2again.Asbefore,theinterac-tionbetweenpoliticalembeddednessandmarketdevelopmentispositiveandstatisticallysignificantinmodel5(more-competitiveindustries)andpositivebutnonsignificantinmodel6(lesscompetitiveindustries).Theinteractionisalsolargerinmodel5thaninmodel6.ButaChowtestshowsthattheinteractioneffectsarenotsignificantlydifferent(p=.16),whichfailstosupporthypothesis2.Models7and8comparesmallerandlargerfirms,basedonamediansplitofoursample,totesthypothesis3again.Theinteractionbetweenpoliticalembeddednessandmarketdevelopmentispositiveandstatisticallysignificantinmodel7(smallfirms)andpositivebutnonsignificantinmodel8(largefirms),andtheinteractionislargerinmodel7thaninmodel8.ButaChowtestshowsthedifferencebetweentheinteractioneffectsisnotstatisticallysignificant(p=.06),whichfailstosupporthypothesis3.Robustnesschecks.Toassessthesensitivityofourresultstomodelspeci-fication,weconductedseveralrobustnesschecks,whicharenotshowntosavespace.First,weestimatedmodelswithrandom(latent)effectsforfirminsteadoffixedeffects.Suchmodelsarelessconservativethanfixed-effectsmodelsbecausetheymakeastrongassumptionthatfirm-levelheterogeneitydoesnotpersistformorethanasingleyear(Wooldridge,2001).Second,weincludedtheprior-yearvalueofthedependentvariablesasanadditionalcontrolvariable.Third,weestimatedmodelsusingtwo-stageleast-squares.Weusedacombinationofvariablestopredictthelikelihoodoffirmsbecomingpoliticallyembedded,includingmarketdevelopmentandfirmsize,plusdummiesforthestateinwhichthefocalfirmwasheadquarteredandthefocalfirm’sindustry.Wethenusedthepredictedvalueofpoliticalembeddednessinthesecond-stageanalysis.Forallalternativemodelspecifications,theresultsshowthesamepatternsasthosediscussedabove.Wetriedmeasuringthedependentvariableasreturnonequity.Wealsoexperimentedwithexplanatoryvariables.Wereplacedthebinarymeasureforpoliticalembeddednesswithtwoalternatives:firstthelogarithmofthenumberofexecutivesanddirectorswhohadservedasthechiefofficerordeputychiefofficeratthedivision(chu)levelinthecadrehierarchyorabove,andsecondtheproportion.Wereplacedthecontinuousmeasureofmarketdevelopmentwithanindicatorfortheyearsafter2003,toreflectlegalchangesinthewakeofChina’saccessiontotheWorldTradeOrganizationinDecember2001.Finally,weusedadifferentmeasureoffirmsize,basedonsales.Analysesusingeachofthesealternativesyieldedresultsthatwerealmostidenticaltotheresultsdiscussedabove.Probingmechanisms:Accesstobankloans.Thefirstfourcolumnsintable4showtheresultsoftheanalysisofbankborrowing,whichtests Havemanetal.89Table4.TheImpactofPoliticalEmbeddednessonCausalMechanisms,ContingentonMarketDevelopment*(1)(2)(3)(4)(5)(6)(7)(8)BorrowingratioRelated-partyloanratioVariablesEmploymentInvestmentEmploymentInvestment••••Politicalembeddedness–.003.055–.003.033–.001–.064–.001–.040(.006)(.030)(.006)(.019)(.004)(.021)(.004)(.014)•••••Marketdevelopment×.053.061–.060–.070Politicalembeddedness(.027)(.032)(.020)(.022)••••••••••••Firmsize(loggedassets,RMB)–.040–.040–.040–.040.005.005.005.005(.005)(.005)(.005)(.005)(.003)(.003)(.003)(.003)••••••State-controlledfirm(yes=1).021.020.021.020–.000.001–.000.001(.008)(.008)(.008)(.008)(.006)(.006)(.006)(.006)Numberofobservations9,9839,9839,9839,9839,1369,1369,1369,136AdjustedR-squared0.6290.6300.6290.6300.3830.3850.3830.385•p<.05;••p<.01;•••p<.001.*Robuststandarderrors(clusteredonfirms)areshowninparentheses.Allmodelsincludefirmfixedeffectsandyearfixedeffects.hypothesis4a.Models1and2usemarketdevelopment(employment),whilemodels3and4usemarketdevelopment(investment).Models1and3aremain-effectsmodels;models2and4addtheinteractionbetweenpoliticalembeddednessandmarketdevelopment.Theinteractionisbothpositiveandstatisticallysignificantinmodel2andnonsignificant(p<.06)inmodel4.Takentogether,theseresultsprovidepartialsupportforhypothesis4a,whichpredictedthatasmarketdevelopmentproceeded,becomingpoliticallyembeddedincreasedbankborrowing.LimitedaccesstocreditisanobstaclethatconfrontsmanyChinesefirms.Theseresultssuggestthatonemechanismthroughwhichpoliticalembeddednesscanhelpfirmsistofacilitatetheiraccesstothisscarceresource,whichislargelystate-controlled.Butthiseffectismodest.Whenmarketdevelopment(employment)wasatitsmean,becomingpoliti-callyembeddedincreasedthepredictedborrowingratioby.001.Whenitwasonestandarddeviationabovethemean,theincreasewas.009.Thatis,asmar-ketdevelopmentincreasedfromthemeanvaluetoonestandarddeviationabovethemean,becomingpoliticallyembeddedincreasedtheborrowingratiobyanadditional.008—a2-percentincreasefromthesample-averageborrowingratio.Probingmechanisms:Related-partyloans.Thelastfourcolumnsintable4showtheanalysisofrelated-partyloanstotesthypothesis5a.Models5and6usemarketdevelopment(employment),whilemodels7and8usemarketdevelopment(investment).Models5and7aremain-effectsmodels;models6and8addtheinteractionbetweenmarketdevelopmentandpoliticalembedd-edness.Theinteractionisnegativeandstatisticallysignificant,indicatingthatasmarketdevelopmentproceeded,becomingpoliticallyembeddedincreasinglyhelpedlistedfirmsdefendthemselvesfrompressuresexertedbytheir 90AdministrativeScienceQuarterly62(2017)controllingshareholderstoengageinvalue-destroyingloansandloanguaran-teesformembersoftheirbusinessgroup.Thusbecomingpoliticallyembeddedbufferedlistedfirmsfromexploitativeshareholders,whichsupportshypothesis5a,andtheseeffectsaresubstantial.Whenmarketdevelopment(employment)wasatitsmean,becomingpoliti-callyembeddeddecreasedtherelated-partyloanratioby.001.Whenitwasonestandarddeviationabovethemean,thedecreasewas.012.Inotherwords,asmarketdevelopmentincreasedfromthemeanvaluetoonestandarddeviationabovethemean,becomingpoliticallyembeddeddecreasedtherelated-partyloanratiobyanadditional.011—a19-percentdecreasefromthesample-averagerelated-partyloanratio.Similarly,whenthemarketdevelop-ment(investment)wasatitsmean,becomingpoliticallyembeddeddecreasedtherelated-partyloanratioby.0004,andwhenitwasonestandarddeviationabovethemean,thechangewas.012.Thatis,asmarketdevelopmentincreasedfromthemeantoonestandarddeviationabovethemean,becomingpoliticallyembeddeddecreasedtherelated-partyloanratiobyanadditional.0116—a21-percentdecreasefromthesample-averagerelated-partyloanratio.Mediationtest.Totesthypotheses4band5banddeterminewhethereas-ieraccesstobankloansandgreaterabilitytofendoffrelated-partyloansarechannelsthroughwhichbecomingpoliticallyembeddedbenefittedfirmsasmarketsdeveloped,weconductedformalmediationtests.Table4showsthattheinteractionbetweenpoliticalembeddednessandmarketdevelopmentwasassociatedwithbothmediators(bankloansandrelated-partyloans),whiletables2and3showthat,withoutcontrollingforeithermediator,thisinterac-tionwasassociatedwithROA.Thefinalstepistoassesswhether,aftercon-trollingforthemediators,theassociationbetweenthisinteractionandROAiseliminatedoratleastweakened.Table5showsthisstep,reportingmodelsofROAthatincludetheborrowingratioandrelated-partyloanratio.Models1to3inthistableusemarketdevelopment(employment);models4to6usemarketdevelopment(investment).Inbothsetsofmodels,wefirstincludedonlytheborrowingratio,thenonlytherelated-partyloanratio,andfinallybothvariables.ResultsshowthattheborrowingratiomediatestheeffectonROAofbecomingpoliticallyembedded,butonlypartly.Inmodel1oftable5,theinter-actionbetweenpoliticalembeddednessandmarketdevelopment(.023)islessthaninmodel1oftable2(.027).Thechangeintheinteractionisstatisticallysignificant.Asimilarbutlesspronouncedpatternisseeninmodel4oftable5:theinteractionthereis.026,comparedwith.030inmodel1oftable3.Again,thechangeisstatisticallysignificant.Yetinbothcases,theinteractionremainsstatisticallysignificant.Table5alsoshowsthattherelated-partyloanratiomed-iatestheeffectonROAofbecomingpoliticallyembeddedbutagainonlypartly.Intable5,theinteractionsbetweenpoliticalembeddednessandmarketdevel-opment(.025inmodel2;.027inmodel4)arelessthanthoseintables2and3(.027intable2,model1;.030intable3,model1).Again,whilebothchangesarestatisticallysignificant,theinteractionsremainstatisticallysignificant.Themorefullyspecifiedmodels(3and6),whichcontainbothcausalmechanisms,showstrongermediationeffects:theinteractionsbetweenpoliticalembedded-nessandmarketdevelopmentdecreasefurther(.021inmodel3;.023inmodel Havemanetal.91Table5.MediationTest:TheImpactofPoliticalEmbeddednessonFirmPerformance(ROA),NetofCausalMechanisms,ContingentonMarketDevelopment*(1)(2)(3)(4)(5)(6)VariableEmploymentInvestment••••••••••Politicalembeddedness.026.030.025.017.019.016(.010)(.011)(.010)(.006)(.007)(.007)••••••Marketdevelopment×.023.025.021.026.027.023Politicalembeddedness(.009)(.010)(.010)(.010)(.012)(.011)••••••••••••••••State-controlledfirms(yes=1)–.014–.012–.014–.014–.012–.014(.003)(.004)(.004)(.003)(.004)(.004)••••••••••••••••••Firmsize(loggedassets,RMB)–.014–.023–.016–.014–.023–.016(.002)(.002)(.002)(.002)(.002)(.002)••••••••••••Borrowingratio.117.113.117.113(.007)(.008)(.007)(.008)••••••Related-partyloanratio–.043–.015–.043–.015(.010)(.010)(.010)(.010)••••••••••••••••••Changeinmarketdevelopment×.004.002.006.004.003.007ÀPoliticalembeddedness(.0001)(.0001)(.0001)(.0001)(.0002)(.0002)Numberofobservations10,7039,1529,13110,7039,1529,1312AdjustedR.344.307.350.344.307.350•p<.10;••p<.05;•••p<.01.*Robuststandarderrors(clusteredonfirms)areshowninparentheses.Allmodelsincludefirmfixedeffectsandyearfixedeffects.ÀFormarketdevelopment(employment),thechangeintheinteractiontermiscalculatedrelativetomodel4intable2.Formarketdevelopment(investment),thechangeintheinteractiontermiscalculatedrelativetomodel4intable3.6).Thesechangesarestatisticallysignificant,buttheinteractionsstillremainsignificant.Figures2and3showthemediationtestgraphically:figure2formarketdevelopment(employment)andfigure3formarketdevelopment(investment).Takentogether,theseresultsindicatethatbankloansandrelated-partyloanssignificantly(albeitpartially)mediatedthejointeffectonfirmperformanceofbecomingpoliticallyembeddedandmarketdevelopment.Theseresultssup-porthypotheses4band5b.Theseflowsoffundsaretwoofthemechanismsthroughwhichbecomingpoliticallyembeddedhelpedfirmsasmarketdevelop-mentproceeded.Othermechanismsbywhichbecomingpoliticallyembeddedincreasinglyimprovedfirmperformance(whichwecannottestwiththesedata),suchasimprovedaccesstobuildingandbusinesspermits,governmentcontracts,andimportorexportpermits,mayexplainwhytheinteractionbetweenpoliticalembeddednessandmarketdevelopmentremainedstatisti-callysignificant.Alternativeexplanations.Onealternativeexplanationfortheseresultsisthatcausalityisreversed:high-rankingstateofficialsmighthaveappointedlower-rankingofficialstobetter-performingfirmsasarewardfortheirservice(Caoetal.,2014)andmighthavebeenmoreinclinedtodosoasmarketdevel-opmentproceeded.Ifthisweretrue,wewouldexpecttoobservestrongerinteractionsbetweenbecomingpoliticallyembeddedandmarketdevelopment 92AdministrativeScienceQuarterly62(2017)Figure2.Mediationanalysis:Employment-basedmeasureofmarketdevelopment.*AccesstoBankLoans.053•(.027).113•••(.008)MarketDevelopment.021•(.010)Firm(Employment)Performance×PoliticalEmbeddedness–.060••(.020)–.015(.010)Related-partyLoansMarketDevelopment.027••(.009)(Employment)Firm×PerformancePoliticalEmbeddedness••••••p<.05;p<.01;p<.001.*Thisfigureisbasedontheestimatesinmodel4oftable2,models2and6oftable4,andmodel3oftable5.Estimatedstandarderrorsareshowninparentheses.Figure3.Mediationanalysis:Investment-basedmeasureofmarketdevelopment.*AccesstoBankLoans.061(.032).113•••(.008)MarketDevelopment.023•(.011)Firm(Investment)Performance×PoliticalEmbeddedness–.070••(.022)–.015(.010)Related-partyLoansMarketDevelopment.030••(.010)(Investment)Firm×PerformancePoliticalEmbeddedness••••••p<.05;p<.01;p<.001.*Thisfigureisbasedontheestimatesinmodel4oftable3,models4and8oftable4,andmodel6oftable5.Estimatedstandarderrorsareshowninparentheses. Havemanetal.93amongstate-ownedthanamongnon-state-ownedfirms,becauseofficialshavemorepowerovertheappointmentinthestate-ownedfirms.Toassessthispossibility,weconductedseparateanalysesofstate-ownedandnon-state-ownedfirms.Amongstate-ownedfirms,theinteractionispositiveandstatisti-callysignificant;amongnon-state-ownedfirms,itispositivebutsmallerandnonsignificant.Butneitherinteractionwassignificantlydifferent.Moreover,therobustnesscheckusingthemilestoneyearof2003asanalternativemeasureofmarketdevelopmentshowedstronger,notweaker,effectsfornon-state-ownedfirmsthanforstate-ownedfirms.Takentogether,theseresultsindicatethatreversecausalityisunlikely.Asecondalternativeexplanationisthattheseresultsaredrivenbysomefac-torthatisnotincludedinouranalysisandthatcausesboththeobservedpat-ternofpoliticalembeddednessandtheobservedpatternoffirmperformance,suchasacommonenvironmentalshock.Butweestimatedmodelsthatincludedyearfixedeffects,whichabsorbtheeffectsofsuchunobservedfac-tors.Moreover,aslongassuchomittedfactorsareuncorrelatedwithhowthesampleissplit,thesuspectedspuriousrelationshipshouldpersistinallsub-samples.Yetoursplit-sampleanalysesindicatethattheinteractionbetweenpoliticalembeddednessandmarketdevelopmentwasstrongerinsmallerfirms.Thusouranalysesobviateconcernsaboutspuriousnessduetoomittedvariables.Althoughwearguedthatrelated-partyloansharmfirms,previousresearchhasshowntheexistenceofacoinsurancerelationshipbetweencontrollingshareholdersandlistedfirmsatparticularpointsintime(Jia,Shi,andWang,2013).Whencontrollingshareholdersexperiencedcreditcrunches,related-partyloansoccurred,funnelingfinancialassetsoutoflistedfirms.Then,whenlistedfirms’performancedropped,non-loan-basedrelated-partytransactionsoccurred,funnelingnon-financialassetsintolistedfirms.Althoughrelated-partyloanswerenotalwaysreciprocated—onlywhenlistedfirms’performancedipped—andrelated-partyloanswerealwaysdetrimentaltolistedfirms,wedidtwothingstoalleviateconcernsthatcoinsuranceconfoundedourresults.First,becausetimingiskeytothecoinsuranceargument,wecontrolledfortimeswhencontrollingshareholdersexperiencedcreditcrunches.Whenwedidthis,ourresultsstillheld.Second,wemodelednon-loan-basedrelated-partytransactions.Whenwedidthis,theinteractionbetweenpoliticalembeddednessandmarketdevelopmentwasnonsignificant,indicatingthatasmarketsdeveloped,politicalembeddednessdidnothelpfirmsgainaccesstothesenonfinancialresources.Together,theseresultsindicatethatrelated-partyloansalwaysharmedlistedfirms’performance,andlistedfirmshadincentivestofightthemoff.Scopeconditions:Non-listedfirms.TocheckthegeneralizabilityofouranalysisoflistedfirmsinChina,weanalyzedanationallyrepresentativesurveyofnon-listedfirms.Forthesefirms,politicalembeddednessoccurredwhentheownerwasadeputyoftheNationalPeople’sCongressorthePeople’sPoliticalConsultativeConference,thecountry’smainpoliticalassemblies.Thisisamoreappropriatemeasureforsmallerfirms:lessthan2percentofnon-listedfirmshadconnectionstothestatebureaucracythroughownerswhohadworkedasbureaucratsatthedivision(chu)levelorabove,comparedwith61percentoflistedfirms.Thepaucityofformerbureaucratsinnon-listedfirmsis 94AdministrativeScienceQuarterly62(2017)likelyduetothefactthattherearemillionsofnon-listedfirmsbutonlyalimitednumberofhigh-levelbureaucrats.Incontrast,justover40percentofnon-listedfirmshadownerswhoweremembersoftheCongressorConference.Weusedpropensity-scorematchingtobalancepoliticallyembeddedandunem-beddedfirmsonobservables,thenanalyzedthebalancedsamples,regressingreturnonequity(dataonassetswerenotavailable)onpoliticalembeddedness,marketdevelopment,theirinteraction,andfirmsize,plusprovinceandindustryfixedeffects.Becausethesedataarerepeatedcross-sections,notapanel,itisdifficulttoassessthedirectionofcausaleffects;thereforeweinterpretresultsonthissampleasevidenceofstatisticalassociation,notcausation.Table6showstheseresults.Models1and2showmodelswithonlymaineffects,whilemodels3and4addinteractionsbetweenpoliticalembedded-nessandthetwomeasuresofmarketdevelopment.Bothinteractionshadpos-itiveandstatisticallysignificanteffects,indicatingthatourfindingsonlistedfirmsgeneralizetounlistedfirms.Becausewehaddataonfirmsize,wecouldtesthypothesis3,whichpredictedasmallerassociationwithlargefirmsthansmallones,withthesedata.Wesplitthedataonthissamplebysize,definingafirmassmallifitstotalassetswerebelowthemedianforsizeinthefocalyearandlargeifitstotalassetswereabovethemedian.Theseresultsareshowninmodels5to8.Interactionsbetweenpoliticalembeddednessandbothmeasuresofmarketdevelopmentwerepositiveandstatisticallysignifi-cantforbothsizegroups,buttheinteractioneffectswerelargerforsmallfirmsthanforlargeones,andbothdifferenceswerestatisticallysignificant(Chowtestp<.01).Theseresultspersistedwhenweaddedyearfixedeffects.ThisTable6.AnalysisofPrivatelyOwnedNon-listedFirms:TheImpactofPoliticalEmbeddednessonFirmPerformance(ROE),ContingentonMarketDevelopment*(1)(2)(3)(4)(5)(6)(7)(8)AllAllSmallLargeSmallLargefirmsfirmsfirmsfirmsfirmsfirmsVariableEmploymentInvestmentEmploymentInvestment••••••Politicalembeddedness.011–.051.010–.096–.046–.031–.101–.058–.009(.016)–.009(.031)(.028)(.014)(.055)(.029)••••••••••••••••••Marketdevelopment–.272–.425–.172–.301–.672.033–.495.049–.055(.065)–.048(.059)(.116)(.062)(.099)(.060)••••••••••Marketdevelopment×.297.251.403.155.323.139Politicalembeddedness(.075)(.074)(.131)(.074)(.129)(.074)••••••••••••••••••••••Firmsize(loggedequity,RMB)–.027–.027–.025–.025–.042–.004–.039–.004–.002(.002)–.001(.001)(.003)(.001)(.002)(.001)Chowtest(Prob>F)<.01(Prob>F)<.01Numberofobservations6,9726,9726,9726,9723,4583,5143,4583,514AdjustedR-squared.074.075.072.074.093.037.090.038•p<.05;••p<.01;•••p<.001.*Politicalembeddednessiscodedoneifthefocalfirm’sownerwasadeputyofthePeople’sCongressorthePeople’sConferenceandzerootherwise.Smallfirmsarethosewithbelow-medianassetsinthefocalyear.Robuststandarderrors(clusteredonfirms)areshowninparentheses.Allmodelsincludeprovinceandindustryfixedeffects. Havemanetal.95analysisbolsterstheconclusionthatourfindingsonlistedfirmsgeneralizewelltootherfirmsinChina.DISCUSSIONOuranalysesgenerallysupportourarguments.AsmarketsdevelopedinChina,havingstatebureaucratsserveasexecutivesanddirectorsinlistedfirmshadincreasinglypositiveeffectsonfirmperformance.And,aspredicted,theseeffectswerestrongerforsmallerfirms.Analysisofarandomsampleofnon-listedfirmsshowedparallelresultsforallfirmsandstrongerresultsforsmallerfirms,suggestingthatourargumentsapplywidelyinChina.Yetcontrarytopre-dictions,amonglistedfirms,interactionsbetweenpoliticalembeddednessandmarketdevelopmentdidnotdiffersignificantlybetweenmoreandlesscom-petitiveindustries.Finally,aspredicted,becomingpoliticallyembeddedincreasedlistedfirms’accesstobankloansandtheirabilitytowardoffrelated-partyloans;bothflowsoffundsaffectedoverallperformance(positivelyforbankloans,negativelyforrelated-partyloans)andpartlymediatedtheinterac-tionbetweenpoliticalembeddednessandmarketdevelopment.Theseempiricalrelationshipsweredrivenbyuncertainty,whichincreasedasmarketsdevelopedandwasworsenedbythepersistentlyauthoritarianpoliticalregime.Chinaexperiencedaseriesofeconomicchanges.Theprivatelyownedsectoroftheeconomygrewanddisplacedthestate-ownedsector;markettransactionsdisplacedstateredistribution;newformsoforganizationswerefoundedtoreplacestate-runproductionfacilitiesandministries;newkindsofgoodsandservicesweredevelopedtoreplacestate-issuedstandardones;andeconomicgrowthtargetswereimposedonstatebureaucrats,whosecareersincreasinglydependedonmeetingthosetargets.Thesechangesexpandedthearrayofopportunitiesforbusinessesbutalsogreatlyincreasedtheuncertaintyfacingbusinesses,increasingtheirlikelihoodofmakingcostlymistakes.Thesechangeswereimplementedgraduallyandatdifferenttimesindifferentregions;suchpersistentandunevenchangefurtherincreaseduncertainty.TheCommunistPartyretainedcontroloverthestatebureaucracyandtheruleoflawremainedweak,sothestateretainedcontrolofkeybusinessresources,andelectoralprocesseswerenotavailabletodisciplinebureaucrats.Thussta-sisinpoliticalinstitutionsfurtherincreaseduncertaintyandweakenedtrustinstateauthorities.Tiestostatebureaucratswereneededtoreducethisuncer-taintyandstrengthentrustinstateauthorities,andthisneedgrewasmarketdevelopmentproceededanduncertaintyincreased.OuranalysisshowedthatChina’seconomicandpoliticalsystemswereinterdependent—theeffectsofitseconomictransitionwerecontingentonitspoliticalregime—andbothhadtobeconsideredtoexplainfirmperformance.Moreover,becausethiseconomictransitionwasbyitsverynaturedynamic,analysisofitseffectsonfirmperformancehadtobedynamic—thuswefocusedontheeffectsofmarketdevelopmentovertimeinthefaceofpoliticalstasis.Ourtheoreticalapproachandempiricaldesignwereadvancesonprevi-ous,mostlystaticresearchonChina’seconomictransitionandshouldprovideatemplateforfutureresearch.Ouranalysisendsin2007,theyearbeforetheglobalfinancialcrisiserupted.Didwhatwefindherepersistafterthatevent?Therearetwodivergentpossi-bilities.First,thefalloutfromthefinancialcrisisslowedChina’seconomic 96AdministrativeScienceQuarterly62(2017)growthconsiderably,from14percentconstant-dollarGDPgrowthin2007to11percentin2010and7percentin2014(WorldBank,2015),whichintensifiedpressureonstatebureaucratstospureconomicgrowthandraiseextra-budgetaryrevenues.Bureaucratscouldaccomplishbothobjectivesbydevelop-ingevencloserrelationshipswithbusiness.Second,thedeeplyentwineddevelopmentofstateandmarketcreatedonesevereproblem:rampantcorrup-tionthroughpredatorystateaction(Manion,2004;Sun,2004;Pei,2006).Anonpoliticized,legallytransparentanti-corruptioncrackdownmightforcebureaucratsandbusinesspeopletoerectmorearm’s-lengthrelationshipsandbecomelessinterdependent.Untilsuchacrackdownbegins,weexpectthatthebenefitsderivedfrombusiness–statetiesliketheoneswestudiedhereincreasedafter2007,butthattrendmayeventuallybereversed.Theinsightsgeneratedinthispaperapplytomanyothertransitioneconomies—thoseinwhicheconomicchangehasbeenaccompaniedbylittle,ifany,politicalchangeandeliteshaveretainedpoliticalpower,suchasinHungaryandRussia(Ro´na-Tas,1994;Walder,1996,2004).Insuchcontexts,theruleoflawisweakandcannotreducetheuncertaintygeneratedbyeco-nomictransition—instead,becausethelegalprocessismurky,theuncertaintycreatedbyeconomictransitionisexacerbated.Insuchcontexts,socialnet-worksofferalternativemechanismsforbuildingtrustandpredictability,facilitat-ingeconomicexchange,andimprovingfirmperformance(GusevaandRo´na-Tas,2001).Becausepoliticalactorsinsuchcontextsretainmuchpowerovertheeconomy,tiesbetweenbusinessesandthestateareespeciallyimportant.Incontrast,theinsightsgeneratedhereapplylesswelltotransitioneconomies,suchasVietnamandPoland,inwhichthepoliticaladvantagesofstatebureau-cratsdissipateduringmarketdevelopment(Walder,2004;WalderandNguyen,2008)andthusthebenefitsoftiestostateauthoritiesarealsolikelytodimin-ish,ratherthanincrease,asmarketsdevelop.Inadditiontothenatureandtrajectoryofthepoliticalregime,thetrajectoryoftheeconomictransitionalsomatters.InChina,thetransitionwasgradualandvariedacrossregions,occurringsoonerandmorecompletelyinsomeregions(thecoast)thanothers(theinterior).Inothercountriesinwhicheliteshaveretainedpoliticalpower,theeconomictransitionhasoccurredmorequicklyanduniformly(e.g.,Russia).Uncertaintywillbeespeciallygreatintheaftermathofa‘‘bigbang’’economictransitionbutmaydeclinelater,althoughthetimepathofuncertaintyinsuchcontextsisanempiricalquestion.Ifuncer-taintydoesdeclineastimepassesincountrieslikeRussia,thepositiveassocia-tionbetweenpoliticalembeddednessandfirmperformancewillbelargestintheearlieststagesoftransitionandweakerinlaterstages,althoughifelitescontinuetodominatepolitics,theassociationwillstillbepositive.Incontrast,incountrieslikeChina,thepositiveassociationwillincreaseaseconomictransi-tionproceeds.Ouranalysissuggestsitisnotpoliticalinstitutionsperse,suchasdemo-craticvs.authoritariangovernments,thatdeterminethevalueoftiestostateauthoritiesduringeconomictransition.Instead,whatmattersiswhetherpoliti-calinstitutionscreateandsustainastrongruleoflaw—clearandimpartiallyadjudicatedlawsconcerningpropertyrights,contracts,andcompetition—andthereforewhetherpoliticalinstitutionslimitstatebureaucrats’powerintheeco-nomicsphereandrendertheiractionstransparentandremediable.Ouranalysishingesonasingleaspectofpoliticalinstitutionsthatshapedtheruleoflaw: Havemanetal.97governmentsthatremainauthoritarianduringtransitionsfromstateredistribu-tiontomarketexchange,sotheruleoflawisweakandpoliticalelitesretainpowerovereconomicactors.Futureresearchcouldfocusonotherpoliticalpro-cessesthatshapetheruleoflawanddeterminethepowerofstateauthorities.Perhapsmostfruitfulwouldbetoinvestigatethedifferentwayspublicassets—here‘‘ChinaInc.’’butinothercontextsregionalcomponentsof‘‘USSRInc.’’—havebeenconvertedtoprivateownership(Walder,2004;Walder,Isaacson,andLu,2015).Themoretheconversionprocessesadoptedbystatebureaucratsallowthemtoappropriatestateassets,themorepowerthosebureaucratsretainovertheeconomy,andthemorerelationswiththemwillaffectbusinesses’performance.InChina,assetappropriationvariedacrossregions(easierinurbanthanruralareas),acrosssectors(easierinrealestatethansteel),andovertime(generallyeasierlaterinthetransition).Assetappro-priationhasalsovariedgreatlyacrosscountries(easierinRussiaandUzbekistan;moredifficultinPolandandVietnam),whichsuggeststhatthisisafruitfulavenueforfutureresearch.AcknowledgmentsWethanktheassociateeditorandthreeanonymousreviewersforhelpfulcomments,aswellasNeilFligstein,IvanaKatic,JohnMatsusaka,BillParish,MikePeng,AndyWalder,andparticipantsattheHongKongInstituteforMonetaryResearchInternationalConferenceontheChineseEconomyandtheU.C.Berkeley/FudanUniversityconfer-ence‘‘TheFieldofGuanxiStudies.’’YongxiangWangthankstheChinaNationalScienceFoundation(GrantNumber:71472180)forfinancialsupport,andYongxiangWangandJingShithanktheAustralianResearchCouncil(DP150102339)forfinancialsupport.REFERENCESAcemoglu,D.,andJ.A.Robinson2012WhyNationsFail:OriginsofPower,Poverty,andProsperity.NewYork:RandomHouse.Allen,F.,J.Qian,andM.Qian2005‘‘Law,finance,andeconomicgrowthinChina.’’JournalofFinancialEconomics,77:57–116.Allen,F.,J.‘‘QJ’’Qian,C.Zhang,andM.Zhao2013‘‘China’sfinancialsystem:Opportunitiesandchallenges.’’InJ.P.H.FanandR.Morck(eds.),CapitalizingChina:63–145.Chicago:UniversityofChicagoPress.Ang,Y.,andN.Jia2014‘‘Perversecomplementarity:PoliticalconnectionsandtheuseofcourtsamongprivatefirmsinChina.’’JournalofPolitics,76:318–332.Assiotis,A.,andK.Sylwester2015‘‘Doesdemocracypromotetheruleoflaw?’’JournalofEconomicDevelop-ment,40:63–92.Baron,R.M.,andD.A.Kenny1986‘‘Themoderator–mediatorvariabledistinctioninsocialpsychologicalresearch:Conceptual,strategic,andstatisticalconsiderations.’’JournalofPersonalityandSocialPsychology,51:1173–1182.Bian,Y.,andL.Zhang2014‘‘CorporatesocialcapitalinChineseguanxiculture.’’InD.J.Brass,G.Labianca,A.Mehra,D.Halgin,andS.P.Borgatti(eds.),ResearchintheSociologyofOrganiza-tions,40:421–443.Bingley,UK:EmeraldBooks. 98AdministrativeScienceQuarterly62(2017)Boisot,M.,andJ.Child1996‘‘Fromfiefstoclansandnetworkcapitalism:ExplainingChina’semergingeco-nomicorder.’’AdministrativeScienceQuarterly,41:600–628.Burt,R.S.1982TowardaStructuralTheoryofAction:NetworkModelsofSocialStructure,Perception,andAction.NewYork:AcademicPress.Calomiris,C.W.,R.Fisman,andY.Wang2010‘‘Profitingfromgovernmentstakesinacommandeconomy:EvidencefromChineseassetsales.’’JournalofFinancialEconomics,96:399–412.Cao,J.,M.Lemmon,X.Pan,M.Qian,andG.Tian2014‘‘Politicalpromotion,CEOincentives,andtherelationshipbetweenpayandper-formance.’’Workingpaper,SingaporeManagementUniversity.Availableathttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=2504930;retrievedApril2015.Chen,J.,andB.J.Dickson2010AlliesoftheState:China’sPrivateEntrepreneursandDemocraticChange.Cambridge,MA:HarvardUniversityPress.Choo,J.M.2014‘‘Thechange-up:TheemergenceoftherealestatesectorinurbanChina,1978–2004.’’UnpublishedPh.D.dissertation,UniversityofCalifornia,Berkeley,DepartmentofSociology.Clarke,D.C.2003‘‘CorporategovernanceinChina:Anoverview.’’ChinaEconomicReview,14:494–507.Clarke,D.C.,P.Murrell,andS.Whiting2008‘‘TheroleoflawinChina’seconomicdevelopment.’’InT.RawskiandL.Brandt(eds.),China’sGreatEconomicTransition:375–428.Cambridge:CambridgeUniver-sityPress.Cull,R.,andL.C.Xu2005‘‘Institutions,ownership,andfinance:ThedeterminantsofprofitreinvestmentamongChinesefirms.’’JournalofFinancialEconomics,77:117–146.Dickson,B.J.2003RedCapitalistsinChina:TheParty,PrivateEntrepreneurs,andProspectsforPoliticalChange.Cambridge:CambridgeUniversityPress.Dickson,B.J.2008WealthintoPower:TheCommunistParty’sEmbraceofthePrivateSector.NewYork:CambridgeUniversityPress.Dobbin,F.1994ForgingIndustrialPolicy:TheUnitedStates,Britain,andFranceintheRailwayAge.Cambridge:CambridgeUniversityPress.Dougherty,S.,R.Herd,andP.He2007‘‘HasaprivatesectoremergedinChina’sindustry?Evidencefromaquartermillionfirms.’’ChinaEconomicReview,18:309–334.Duckett,J.1998TheEntrepreneurialStateinChina:RealEstateandCommerceDepartmentsinReformEraTianjin.NewYork:Routledge.Evans,P.B.1995EmbeddedAutonomy:StatesandIndustrialTransformation.Princeton,NJ:PrincetonUniversityPress.Fei,X.1992FromtheSoil:TheFoundationsofChineseSociety.G.HamiltonandW.Zhang,trans.(Originallypublishedin1948asXiangtuZhongguo.)Berkeley,CA:UniversityofCaliforniaPress. Havemanetal.99Fisman,R.,andY.Wang2010‘‘TradingfavorswithinChinesebusinessgroups.’’AmericanEconomicReview:PapersandProceedings,100(May):429–433.Fligstein,N.2001TheArchitectureofMarkets:AnEconomicSociologyofTwenty-First-CenturyCapitalistSocieties.Princeton,NJ:PrincetonUniversityPress.Fligstein,N.,andT.Shin2007‘‘ShareholdervalueandthetransformationoftheUSeconomy,1984–2000.’’SociologicalForum,22:399–424.Fligstein,N.,andJ.Zhang2009‘‘AnewagendaforresearchonthetrajectoryofChinesecapitalism.’’Manage-mentandOrganizationReview,7:39–62.Gold,T.1990‘‘Urbanprivatebusinessandsocialchange.’’InD.DavisandE.Vogel(eds.),ChineseSocietyontheEveofTiananmen:157–180.Cambridge,MA:HarvardUniver-sityPress.Granovetter,M.S.1985‘‘Economicactionandsocialstructure:Theproblemofembeddedness.’’Ameri-canJournalofSociology,91:481–510.Granovetter,M.S.1993‘‘Coaserevisited:Businessgroupsinthemoderneconomy.’’InN.SmelserandR.Swedberg(eds.),HandbookofEconomicSociology:453–475.Princeton,NJ,andNewYork:PrincetonUniversityPressandRussellSageFoundation.Guseva,A.,andA´.Ro´na-Tas2001‘‘Uncertainty,risk,andtrust:RussianandAmericancreditcardmarketscom-pared.’’AmericanSociologicalReview,66:623–646.Guthrie,D.1999DragoninaThree-pieceSuit:TheEmergenceofCapitalisminChina.Princeton,NJ:PrincetonUniversityPress.Halaby,C.N.2004‘‘Panelmodelsinsociologicalresearch:Theoryintopractices.’’InK.S.CookandJ.Hagan(eds.),AnnualReviewofSociology,30:507–544.PaloAlto,CA:AnnualReviews.Holz,C.A.,andY.-M.Lin2001‘‘The1997–1998breakinindustrialstatistics:Factsandappraisal.’’ChinaEco-nomicReview,12:303–316.Jia,N.,J.Shi,andY.Wang2013‘‘Coinsurancewithinbusinessgroups:Evidencefromrelatedpartytransactionsinanemergingmarket.’’ManagementScience,59:2295–2313.Jiang,G.,C.M.C.Lee,andH.Yue2010‘‘Tunnelingthroughintercorporateloans:TheChinaexperience.’’JournalofFinancialEconomics,98:1–20.Johnson,C.1982MITIandtheJapaneseMiracle:TheGrowthofIndustrialPolicy,1925–1975.Stanford,CA:StanfordUniversityPress.Keister,L.A.2000ChineseBusinessGroups:TheStructureandImpactofInterfirmRelationsduringEconomicDevelopment.NewYork:CambridgeUniversityPress.Keister,L.A.2004‘‘Capitalstructureintransition:ThetransformationoffinancialstrategiesinChina’semergingeconomy.’’OrganizationScience,15:145–158.Kennedy,S.2005TheBusinessofLobbyinginChina.Cambridge,MA:HarvardUniversityPress. 100AdministrativeScienceQuarterly62(2017)Khanna,T.,andY.Yafeh2007‘‘Businessgroupsinemergingmarkets:Paragonsorparasites?’’JournalofEco-nomicLiterature,45:331–372.Landry,P.F.2008DecentralizedAuthoritarianisminChina:TheCommunistParty’sControlofLocalElitesinthePost-MaoEra.NewYork:CambridgeUniversityPress.LaPorta,R.,F.Lopes-de-Silanes,A.Shleifer,andR.Vishny1998‘‘Lawandfinance.’’JournalofPoliticalEconomy,106:1113–1155.Lazonick,W.,andM.O’Sullivan2000‘‘Maximizingshareholdervalue:Anewideologyforcorporategovernance.’’EconomyandSociety,29:13–35.Li,H.,L.Meng,Q.Wang,andL.-A.Zhou2008‘‘Politicalconnections,financingandfirmperformance:EvidencefromChineseprivatefirms.’’JournalofDevelopmentEconomics,87:283–299.Li,H.,L.Meng,andJ.Zhang2006‘‘Whydoentrepreneursenterpolitics?EvidencefromChina.’’EconomicInquiry,44:559–578.Li,H.,andY.Zhang2007‘‘Theroleofmanagers’politicalnetworkingandfunctionalexperienceinnewventureperformance:EvidencefromChina’stransitioneconomy.’’StrategicManage-mentJournal,28:791–804.Li,X.-H.,andX.Liang2015‘‘AConfuciansocialmodelofpoliticalappointmentsamongChineseprivate-firmentrepreneurs.’’AcademyofManagementJournal,58:592–617.Lie,J.1991‘‘EmbeddingPolanyi’smarketsociety.’’SociologicalPerspectives,34:219–235.Lin,L.-W.,andC.J.Milhaupt2013‘‘Wearethe(national)champions:UnderstandingthemechanismsofstatecapitalisminChina.’’StanfordLawReview,65:697–760.Lin,N.2001‘‘Guanxi:Aconceptualanalysis.’’InA.So,N.Lin,andD.Poston(eds.),TheChi-neseTriangleofMainlandChina,Taiwan,andHongKong:153–166.Greenwich,CT:Greenwood.Lin,Y.2001BetweenPoliticsandMarkets:Firms,Competition,andInstitutionalChangeinPost-MaoChina.Cambridge:CambridgeUniversityPress.Lu,J.2012‘‘VarietiesofelectoralinstitutionsinChina’sgrassrootsdemocracy:Cross-sectionalandlongitudinalevidencefromruralChina.’’ChinaQuarterly,210:482–493.Luo,Y.2003‘‘Industrialdynamicsandmanagerialnetworkinginanemergingmarket:ThecaseofChina.’’StrategicManagementJournal,24:1315–1327.Manion,M.2004CorruptionbyDesign.Cambridge,MA:HarvardUniversityPress.McGregor,R.2012TheParty:TheSecretWorldofChina’sCommunistRulers.NewYork:HarperPerennials.Michelson,E.2007‘‘Lawyers,politicalembeddedness,andinstitutionalcontinuityinChina’stransi-tionfromsocialism.’’AmericanJournalofSociology,113:352–414.Naughton,B.2007TheChineseEconomy:TransitionsandGrowth.Cambridge,MA:MITPress.Nee,V.,andS.Opper2010‘‘Politicalcapitalinamarketeconomy.’’SocialForces,88:2105–2132. Havemanetal.101Nee,V.,andS.Opper2012CapitalismfromBelow:MarketsandInstitutionalChangeinChina.Cambridge,MA:HarvardUniversityPress.North,D.C.1990Institutions,InstitutionalChange,andEconomicPerformance.Cambridge:CambridgeUniversityPress.Oi,J.C.,andS.Rozelle2000‘‘Electionsandpower:Thelocusofdecision-makinginChinesevillages.’’ChinaQuarterly,162:513–539.Parish,W.L.,andE.Michelson1996‘‘Politicsandmarkets:Dualtransitions.’’AmericanJournalofSociology,101:1042–1059.Pei,M.2006China’sTrappedTransition:TheLimitsofDevelopmentalAutocracy.Cam-bridge,MA:HarvardUniversityPress.Peng,M.W.,andY.Luo2000‘‘Managerialtiesandfirmperformanceinatransitioneconomy:Thenatureofamicro–macrolink.’’AcademyofManagementJournal,43:486–501.Pfeffer,J.,andG.R.Salancik1978TheExternalControlofOrganizations:AResourceDependencePerspective.NewYork:HarperandRow.Polanyi,K.1944TheGreatTransformation:ThePoliticalandEconomicOriginsofOurTime.Boston:BeaconPress.Polanyi,K.1957‘‘Theeconomyasinstitutedprocess.’’InK.Polanyi,C.M.Arensberg,andH.W.Pearson(eds.),TradeandMarketintheEarlyEmpires:EconomiesinHistoryandTheory:243–270.Glencoe,IL:FreePress.Przeworski,A.,M.E.Alvarez,J.A.Cheibub,andF.Limongi2000DemocracyandDevelopment:PoliticalInstitutionsandWell-beingintheWorld,1950–1990.NewYork:CambridgeUniversityPress.Ro´na-Tas,A´.1994‘‘Thefirstshallbelast?Entrepreneurshipandcommunistcadresinthetransi-tionfromsocialism.’’AmericanJournalofSociology,100:40–69.Rosenbaum,P.R.,andD.B.Rubin1983‘‘Thecentralroleofthepropensityscoreinobservationalstudiesforcausaleffects.’’Biometrika,70:41–55.Scott,W.R.2003Organizations:Rational,Natural,andOpenSystems,6thed.UpperSaddleRiver,NJ:Prentice-Hall.Shih,V.C.2008FactionsandFinanceinChina:EliteConflictandInflation.NewYork:CambridgeUniversityPress.Shleifer,A.,andR.Vishny1997‘‘Asurveyofcorporategovernance.’’JournalofFinance,52:737–783.Simon,H.A.1957ModelsofMan.NewYork:JohnWiley&Sons.Simon,H.A.1962‘‘Thearchitectureofcomplexity.’’ProceedingsoftheAmericanPhilosophicalSociety,106:467–482.Sobel,M.E.1982‘‘Asymptoticconfidenceintervalsforindirecteffectsinstructuralequationmod-els.’’SociologicalMethodology,13:290–312. 102AdministrativeScienceQuarterly62(2017)Stark,D.,andL.Bruszt1998PostsocialistPathways:TransformingPoliticsandPropertyinEastCentralEurope.NewYork:CambridgeUniversityPress.Stuart,T.E.,H.Hoang,andR.C.Hybels1999‘‘Interorganizationalendorsementsandtheperformanceofentrepreneurialven-tures.’’AdministrativeScienceQuarterly,44:315–349.Sun,Y.2004CorruptionandMarketinContemporaryChina.Ithaca,NY:CornellUniversityPress.Tirole,J.1988TheTheoryofIndustrialOrganization.Cambridge,MA:MITPress.Tsai,K.S.2002Back-alleyBanking:PrivateEntrepreneursinChina.Ithaca,NY:CornellUniver-sityPress.Tsai,K.S.2007CapitalismwithoutDemocracy:ThePrivateSectorinContemporaryChina.Ithaca,NY:CornellUniversityPress.Uzzi,B.1996‘‘Thesourcesandconsequencesofembeddednessfortheeconomicperfor-manceoforganizations:Thenetworkeffect.’’AmericanSociologicalReview,61:674–698.Uzzi,B.1999‘‘Embeddednessinthemakingoffinancialcapital:Howsocialrelationsandnet-worksbenefitfirmsseekingcapital.’’AmericanSociologicalReview,64:481–505.Walder,A.1994‘‘Thedeclineofcommunistpower:Elementsofatheoryofinstitutionalchange.’’TheoryandSociety,23:297–323.Walder,A.G.1996‘‘Marketsandinequalityintransitionaleconomies:Towardtestabletheories.’’AmericanJournalofSociology,101:1060–1073.Walder,A.G.2004‘‘ThepartyeliteandChina’strajectoryofchange.’’China:AnInternationalJour-nal,2:189–209.Walder,A.G.,A.Isaacson,andQ.Lu2015‘‘Afterstatesocialism:Thepoliticaloriginsoftransitionalrecessions.’’AmericanSociologicalReview,80:444–468.Walder,A.G.,andG.H.Nguyen2008‘‘Ownership,organization,andincomeinequality:MarkettransitioninruralViet-nam.’’AmericanSociologicalReview,73:251–269.Wang,J.,D.Guthrie,andZ.Xiao2012‘‘TheriseofSASAC:Assetmanagement,ownershipconcentration,andfirmperformanceinChina’scapitalmarkets.’’ManagementandOrganizationReview,8:253–281.Wank,D.L.1999CommodifyingCommunism:Business,Trust,andPoliticsinaChineseCity.Cambridge:CambridgeUniversityPress.Wank,D.L.2002‘‘Business–stateclientelisminChina:Declineorevolution?’’InT.Gold,D.Guthrie,andD.Wank(eds.),SocialConnectionsinChina:Institutions,Culture,andtheChangingNatureofGuanxi:97–115.Cambridge:CambridgeUniversityPress.Whitley,R.1999DivergentCapitalisms:TheSocialStructuringandChangeofBusinessSys-tems.Oxford:OxfordUniversityPress. Havemanetal.103Witt,M.A.,andG.Redding2013‘‘Asianbusinesssystems:Institutionalcomparison,clusters,andimplicationsforvarietiesofcapitalismandbusinesssystemstheory.’’Socio-EconomicReview,11:265–300.Wooldridge,J.M.2001EconometricAnalysisofCrossSectionandPanelData.Cambridge,MA:MITPress.WorldBank2015WorldDevelopmentIndicators.Washington,DC:WorldBank.Accessedathttp://databank.worldbank.org/data/reports.aspx?source=2&country=&series=NY.GDP.MKTP.KD.ZG&period.RetrievedDecember2,2015.https://www.google.fr/publicdata/explore?ds=d5bncppjof8f9_&met_y=ny_gdp_pcap_cd&idim=country:CHN:IND&hl=en&dl=en.Yang,D.2005RemakingtheChineseLeviathan:MarketTransitionandthePoliticsofGovern-anceinChina.Stanford,CA:StanfordUniversityPress.Yang,M.M.-H.1994Gifts,Favors,andBanquets:TheArtofSocialRelationshipsinChina.Ithaca,NY:CornellUniversityPress.Zheng,L.2013‘‘Politicalembeddednessasadouble-edgedsword:FirmsinChina’sstockmar-ket.’’Workingpaper,DepartmentofSociology,TsinghuaUniversity.Zheng,L.,andB.-S.Kim2011‘‘Spin-offsandcorporategovernance.’’InJ.C.Oi(ed.),GoingPrivateinChina:ThePoliticsofCorporateRestructuringandSystemReforminthePRC:241–269.PaloAlto,CA:Asian-PacificResearchCenterandtheBrookingsInstitution.Zhou,X.2000‘‘EconomictransformationandincomeinequalityinurbanChina:Evidencefrompaneldata.’’AmericanJournalofSociology,105:1135–1174.Zhou,X.2001‘‘PoliticaldynamicsandbureaucraticcareerpatternsinthePeople’sRepublicofChina,1949–1994.’’ComparativePoliticalStudies,34:1036–1062.Zhou,X.2010‘‘TheinstructionallogicofcollusionamonglocalgovernmentsinChina.’’Mod-ernChina,31:47–78.Zhou,X.,Y.Ai,andH.Lian2012‘‘Thelimitofbureaucraticpowerinorganizations:ThecaseoftheChinesebureaucracy.’’InD.Courpasson,D.Golsorkhi,andJ.J.Sallaz(eds.),ResearchintheSociologyofOrganizations,34:81–111.Bingley,UK:EmeraldBooks.Authors’BiographiesHeatherA.HavemanisaprofessorofsociologyandbusinessattheUniversityofCalifornia,Berkeley,DepartmentofSociology,410BarrowsHall,Berkeley,CA94720-1980(e-mail:haveman@berkeley.edu).Shestudieshoworganizations,industries,andemployees’careersevolve;inadditiontoChinesefirms,sheiscurrentlyanalyzingAmericanmagazinesandstate-legalmarijuanamarkets.ShereceivedaPh.D.inorgani-zationalbehaviorandindustrialrelationsfromUCBerkeley.NanJiaisanassistantprofessorofstrategicmanagementattheMarshallSchoolofBusiness,UniversityofSouthernCalifornia,BRI306,3670TrousdaleParkway,LosAngeles,CA90089-0808(e-mail:nan.jia@marshall.usc.edu).Herresearchinterestsincludecorporatepoliticalstrategy,business-governancerelationships,andcorporate 104AdministrativeScienceQuarterly62(2017)governanceininternationalbusiness.ShereceivedherPh.D.instrategicmanagementfromtheRotmanSchoolofManagement,UniversityofToronto.JingShiisaprofessoroffinanceatSchoolofEconomics,FinanceandMarketing,RMITUniversity,445SwanstonStreet,Melbourne,VIC3000,Australia,andInternationalInstituteforFinancialStudies,JiangxiUniversityofFinanceandEconomics(e-mail:Jing.Shi@rmit.edu.au).Hisresearchfocusesonempiricalcorporatefinance,politicaleconomy,andemergingmarkets.HereceivedhisPh.D.infinancefromtheResearchSchoolofFinance,ActuarialStudiesandStatisticsattheAustralianNationalUniversity.YongxiangWangisanassistantprofessoroffinanceattheMarshallSchoolofBusiness,UniversityofSouthernCalifornia,BRI308,3670TrousdaleParkway,LosAngeles,CA90089-0804(e-mail:yongxiaw@marshall.usc.edu).Hisresearchfocusesonempiricalcorporatefinance,politicaleconomy,developmenteconomics,andtheinterac-tionofthesefields.HeobtainedhisPh.D.infinancefromtheColumbiaBusinessSchool.
此文档下载收益归作者所有
