您的当前位置：首页正文

Political constraints and performance measurement in China

来源：榕意旅游网

Accounting,OrganizationsandSociety31(2006)157–177

www.elsevier.com/locate/aos

Politicalconstraints,organizationdesignandperformance

measurementinChinaÕsstate-ownedenterprises

NealeG.O’Connor

aa,*,JohnnyDengb,YadongLuo

cDepartmentofAccountancy,CityUniversityofHongKong,83TatCheeAvenue,KowloonTong,Kowloon,HongKong

bRoutonElectricCoLtd.,MiaoshanDistrict,Wuhan,PRChina

cDepartmentofManagement,SchoolofBusinessAdministration,UniversityofMiami,FL33124,USA

Abstract

Thisstudydevelopsatheoreticalmodeltotesthowpoliticalconstraintsonlabordecisionsmediatetheeﬀectsofeconomicliberalizationforcesonaspectsoforganizationaldesignsuchasdelegation,performancemeasurement,andincentivesinChinesestate-ownedenterprises.Hypothesestestsusingalargesurveyofdivisionalmanagersgener-allyconﬁrmthemodel:thattheinﬂuenceofthreeliberalizationforces(industrylevelgrowthandforeignﬁrmcompe-tition,jointventureexperienceandstockmarketlisting)onorganizationdesignismediatedbypoliticalconstraints.Ó2005ElsevierLtd.Allrightsreserved.

Introduction

Since1997,moststate-ownedenterprises(SOEs)inChinahavebecomemarkedlymoreindependentasaresultofamassiveliberalizationprogram,whichhasnotonlycreatedgreaterautonomyforbusinessexpansionbuthasalsointroducedcompetitivepressureunderwhichSOEsstrivetogroworstruggletosurvivethroughsigniﬁcantimprovementsofproductivity,

eﬃciency,innovation,andservices.Toachievetheseends,thedevelopmentofamore‘‘Western’’micro-levelorganizationdesignisnowcommonlyviewedasthefundamentalforcethatdeterminesthesuccessfulrestructuringofSOEs(Jeﬀerson,Rawski,&Zheng,1996;Qian,1996;Shirley&Xu,2001).1Forinstance,recentresearchindicatesthattheuseofincentiveschemesinChinaÕsSOEsenhancesproductivity(Qian,2001;Xu,2000;

ThetermÔWesternÕisusedtoreﬂecttheeconomicprescrip-tions(eﬃciencyrationale)thatunderliethechoiceoftheorganizationaldesigncomponents(forexample,delegation,performancemeasurement,andincentivecompensation)intheaccountingandeconomicsliterature(Milgrom&Roberts,1995).

1Correspondingauthor.Tel.:+85227888964;fax:+85227887944.

E-mailaddresses:acno@cityu.edu.hk(N.G.O’Connor),dengf@routon.com(J.Deng),yadong@miami.edu(Y.Luo).

*0361-3682/$-seefrontmatterÓ2005ElsevierLtd.Allrightsreserved.doi:10.1016/j.aos.2005.01.005

158N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Zhuang&Xu,1996).Lee(2001)reportsincreaseddelegationduringtheprocessoftheﬁnancialrestructuringofalargeSOE.Otherstudiesﬁndthatliberalizationforcessuchasmarketcompeti-tion(Firth,1996),foreignjointventureexperience(Firth,1996;OÕConnor,Chow,&Wu,2004),andstockmarketlisting(OÕConnoretal.,2004)areassociatedwiththelevelofadoptionofvariousWesternmanagementaccountingtechniques,includingperformancemeasurementsystems.

Todate,however,therehasbeenlittlesystem-aticexaminationofhowtheseorganizationalde-signcomponentsareconstrainedbypoliticalconditions,whichtendtobeanimportantfeatureofeconomiesintransition,especiallyChina(Qian,1996).Recentstudieshighlighttheprevalenceofpoliticalconstraints.Forinstance,Li(2000)pro-videsevidencethattightergovernmentalcontrolresultsinmoreunproﬁtableproductionandsur-plusemployment.Xu,Zhu,andLin(2002)ﬁndthatpoliticalinterferencetendstodominatelabordecisions,whereasotherdecisionsaredominatedbyagencycosts.AccordingtoQian(1996,p.429),‘‘understandingthisinteractionbetweentheeﬀectivecontrolbymanagersoversomedecisionsandtheultimatecontrolbythePartyandthegov-ernmentoverotherdecisionsisthekeytounder-standingtheproblemswiththepastreformandtheissuestobeaddressedinthefuture.’’Theques-tionofhowtobreakuptheoldvestedinterestsinstate-ownedindustryhasbeendescribedasa‘‘for-biddenarea’’ofreform,becauseitaﬀectsthegov-ernmentÕsabilitytoregulate,monitor,orcontrolemploymentandotherresourceallocationissues.‘‘ThecontradictionsinthesituationareobvioustomanyChineseenterprisemanagersandaca-demicanalysts,butthereisonlylimitedresearchonhowtoresolvethembecauseoftheissueÕsgreatpoliticalsensitivity.’’(Hassard,Sheehan,&Mor-ris,1999,p.76).

ThispaperextendstheliteratureonChineseSOEreformbyexaminingthemediatinginﬂuenceofpoliticalconstraintsonorganizationaldesigninChinaÕsSOEs.PoliticalconstraintsaredeﬁnedinthispaperasthedegreetowhichgovernmentalauthoritiesandCommunistPartyrepresentativesintervene,regulate,orcontrolanSOEÕslabordeci-sions(hiring,ﬁring,andpromotion).Asthese

politicalconstraintsareinplayalongwithmarketliberalizationforces(Li,2000),itislikelythattheeﬀectofliberalizationforcesontheorganizationaldesignordecisionsofSOEsismediatedbypoliti-calconstraints.Forexample,marketcompetitionhasapositivedirecteﬀectontheadoptionofWes-ternmanagementcontrols(e.g.,Firth,1996),butalsoanindirecteﬀectbecauseitisassociatedwithhighergrowthindustries,whichweretheﬁrsttoexperiencethereducedpoliticalconstraints(Chen,2000)thatcanslowtheadoptionofWesternman-agementcontrols.Thedirecteﬀectpotentiallypro-videsamisleadingimpressionoftheinﬂuenceofcompetition.Wesuggestthatthispossiblemediat-ingeﬀectisimportantbecause,inasocialistmar-keteconomysuchasChina,economicreformstakeplaceundervariouspoliticalconstraints,andaccordinglySOEsneedtocopewithnotonlyeconomictransformationandmarketliberaliza-tionbutalsopoliticalconstraintsandregulatoryconditions.

Wedevelopatheoreticalmodeltotestthemedi-atinginﬂuenceofpoliticalconstraintsontheasso-ciationsbetweenliberalizationforces,andtheuseofthreeorganizationaldesigncomponents(dele-gation,performancemeasurement,andincentives)withinSOEs.ConsistentwiththerecentChinare-formliterature,weuseagencytheorytodeveloptheseparatelinksinthemodel.2Weusetheﬁrmastheunitofanalysisandfocusontheorganiza-tionaldesigncomponentsatthedivisionalman-agerlevelsuchasproﬁt-centermanagers(invariousdivisions,branches,orunits)andcost-cen-

2Researcherssuggestthatcollectiveculturalvalueschallengetheself-interestassumptionthatunderliesagencytheoryandthedesignofmanyaccountingcontrolsystems(e.g.,Howell&Sakurai,1992).However,thereisgrowingevidenceoftheprevalenceofself-interestedbehaviorinChina.Forexample,Chen(1995)ﬁndsthatmainlandChineseemployeesinreform-orientedcompaniesactuallyfavormoremerit-basedpay.InanotherstudyinvolvingUS,Japanese,andChinesesubjects,Bailey,Chen,andDou(1997)ﬁndthattheChinese‘‘consis-tentlydepartedfromprediction,’’hypothesizingtheeﬀectofcollectivismontheirpreferenceforperformancefeedback.TheirresultsledthemtoconcludethattheChinesemay‘‘departfromcollectivistvaluesinwaysthatresembleAmericansmorethanJapanese’’(pp.605,620).

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177159

termanagers(invariousdepartments),asthedeci-sionsandactionsofmanagersatthislevelarelikelytohaveafargreatereﬀectontheenterprisethanthoseoflowerlevelmanagers(OÕConnoretal.,2004).WeuseananalyticalframeworkthatwasdevelopedbyJensenandMeckling(1992),MilgromandRoberts(1992,1995),andBrickley,Smith,andZimmerman(1995,2001),inwhichthedelegationofdecision-makingauthority(dele-gation),objectiveperformancemeasures,andincentivecompensationwithinanorganizationÕshierarchyjointlyconstitutetheorganizationÕsde-sign.Theconsiderationoftheinterdependenciesamongthesecomponentsaddsanotherdimensiontoourunderstandingoforganizationaldesignandisconsistentwithrecentstudiesinthemanagement(Mendelson,2000),economics(Delmastro,2002)andaccounting(Nagar,2002)literatures.Forexample,Nagar(2002)ﬁndsthatunderstandingtheinﬂuenceofﬁrmgrowthonthestrengthofincentivesisenhancedwhentheinterdependencyofdelegationandincentivesistakenintoconsideration.

Theremainderofthepaperisorganizedasfol-lows.Thenextsectiondevelopsthethreehypothe-sesthatcomprisethemodel.Theresearchmethodsandmeasurementofthevariablesarethenex-plained,followedbytheresults.Thepapercon-cludeswithadiscussionoftheﬁndings.

TheoryandhypothesesTheoreticalmodel

Liberalizationandpoliticalconstraints

OneofthemainthrustsofrecentSOEreformshasbeentoencourageSOEsingrowthindustriesincertainlocationstoseekalternativesourcesofcapitaltodecreasetheirrelianceonthestate.ThemajorliberalizationforcesthatpertaintoChi-neseSOEsincludemarketcompetition(industrygrowth,foreigncompetition),exportmarketsales,jointventureexperience,andstockexchangelist-ing(Huang&Duncan,1997;Lin,2000;Lin,Cai,&Li,1998;Xu,2000).Inthetransitionfromcen-tralplanningtoamarketeconomy,however,SOEoperationsarecomplicatedbecauseliberalization

forcesandstateinﬂuencescoexistandjointlycon-strainSOEmanagementandorganizationalchange(Child,1994).Inparticular,opaque,uncer-tain,andunpredictableregulatoryframeworksthatareformedbybothcentralandlocalgovern-mentsheightenthecomplexityoforganizingpro-ductionandmarketing,andoftennullifythestrategicplanningofSOEs(Linetal.,1998).

OneoftheprimechannelsinwhichtheStatecontrolsSOEsisinthearealabordecisionmaking.PoliticalconstraintsonlabordecisionmakinginSOEstakeplacethroughvariouspolicyregula-tionsandgovernmentrepresentatives(i.e.,partysecretaries,boardmembers,orstate-asset-manage-mentrepresentatives).SuchcontrolisorchestratedviathepowerofCommunistPartyrepresentativestointerveneinenterprisedecisionmaking,includ-ingtheappointing,ﬁring,andpromotionofdivi-sionalmanagers(Hassardetal.,1999).AlthoughpreviousSOEreformsadoptedvariousrestructur-ingpolicies,thefundamentalprincipleoftheso-called‘‘Partycontrolspersonnel’’policyremainedunchallenged.

Whenpersonnelmanagementsystems,includ-ingauthorityregimesandperformanceevaluationstructures,areinfusedwithpoliticalconsiderationsthroughgovernmentinterference(e.g.,imposing‘‘caps’’toreduceincomedisparity,orrestrictingthedismissalofredundantorunproductiveemployees),thebeneﬁtstobegainedbyadoptingthesecomponentsbecomerestricted,andagencyhazardswithintheSOEhierarchybecomemoreserious.SuchrestrictionsorhazardsarewhatShle-iferandVishney(1994)callthepoliticalcoststhatareassociatedwithprivatizationandliberaliza-tion.Thepoliticalcostsmodelstatesthattheeﬀec-tiverestructuringormodernizationofpublicenterprisesislargelydependentontheextenttowhichemploymentcontrolrightsaretransferredtomanagement(fromgovernment)intheprocessofcorporatization(Shleifer&Vishney,1994,p.1015).Thisisexpectedtodirectlyaﬀectthefunc-tionalityoftheorganizationaldesignandtheeﬀec-tivenessofitscomponents(Pannier,1996).Organizationaldesigncomponents

Topmanagementplaysamajorroleinformu-latingorganizationaldesign,includingdecisions

160N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

abouthowmuchdecisionauthoritytodelegatetodivisionalmanagers,andhowtostructuretheper-formanceevaluationandrewardsystemsothatqualiﬁedemployeesareattracted,retained,andmotivated(Brickleyetal.,2001).3Whendelegatedtomanagersinvariousproﬁtorcostcenters,deci-sionrights—whicharetherightstodecideonandtakeaction—canboostorganizationaladaptationandmarketresponsiveness(Bushman,Indjejikian,&Penno,2000).KnowledgetransfercostswithinanorganizationÕshierarchyarelowerwhendeci-sion-makingrightsaremovedtoindividualswhooperateatorganizationaledges(Christie,Joye,&Watts,2003).4Firmsalsoneedtomeasurethebehaviorandeﬃciencyofdivisionalmanagerswithamixofobjectiveandsubjectiveperformancemeasures,andtorewardthemonthisbasis(Haveman,1992;Jensen&Meckling,1995).Thegreateruseofobjectiveperformancemeasures,whichisde-ﬁnedinthisstudyastherelativeweightingplacedonobjectivemeasuresintheobjective/subjectiveperformancemeasurementmix,meansthatthemeasurementisincreasingly‘‘freefrompersonalbias’’(Merchant,1989,p.26),whichinturnre-ducesthepotentialgainsfrominﬂuencingactivi-tiesasperceivedbytheagent(Prendergast,1999).5Objectiveperformancemeasuresalsopro-videanimportantnormofexpectedperformance,ashighuncertaintyinperformanceevaluationmayotherwiseinhibitthedevelopmentofentrepreneur-

3Thisisnotanexhaustiveviewofcontrols.Controlsalsocommunicatestrategyanddevelopcapabilitiesintheorgani-zation(Simons,1995).4Jensen(1998)deﬁnesknowledgetransfercostsalongaspeciﬁc(general)knowledgecontinuumthatmeasureshigh(low)transfercosts.Werefertoknowledgetransfercostsintermsofthiscontinuum.SeeChristieetal.(2003)foranextensivereviewofdelegationandknowledgetransfercosts.5Theaccountingliteratureidentiﬁesseveraleconomicattri-butesofperformancemeasures(informativeness,sensitivity,noise,andobjectivity:Moers,2005).TheobjectivityattributeofperformancemeasurementisparticularlycriticalintheChinesetransitionaleconomy(andthereforethisstudy)becauseper-formancemeasurementsystemsatthedivisionalandlowermanagementlevelshavetraditionallybeenhighlysubjective(dependentonsuperiorpersonalassessment)(Byrd&Tidrick,1991).

ialattitudesandbehavioramongmanagers,thusenlargingagencycosts(Baker,Gibbons,&Mur-phy,1994).Subjectivityalsoplaysanimportantroleineﬃcientperformancemonitoringbecauseitcanreducegamingactivities(Gibbs,Merchant,VanderStede,&Vargus,2004;Ittner,Larcker,&Meyer,2003).However,intheChinareformcontextsubjectivityinperformanceevaluationhastraditionallybeenusedtopreservethestatusquointermsofcentralizedpowerintheSOEratherthanasaneﬃciencyenhancingmechanism(Byrd&Tidrick,1991).

Finally,agencytheoryprescribestheuseofincentivecompensationthatisbasedontheper-formanceofagents,becausetheresultantinforma-tionasymmetriesmaketheirbehaviorcostlyordiﬃculttoobserve(Eisenhardt,1989;Gerhart&Milkovich,1990).Theuseofincentivesisdeﬁnedinthisstudyasthesizeofmonetaryrewardsthatdistinguishbetweenlevelsofmanagerperfor-mance.Thesethreecomponentsareinterrelatedandinseparableinsuchamannerthatperfor-mancemonitoringisafoundationonwhichre-wardandauthorityallocationsarebuiltandappraised,whiledecisionrightsandincentivesaccentuateeachotherinacompetitiveyetvolatileenvironment(Nagar,2002;Tosi,Katz,&Gomez-Mejia,1997).ConsistentwiththeorganizationaldesignframeworkdevelopedbyJensenandMec-kling(1992)andothers(Brickleyetal.,2001,Chapter11;Milgrom&Roberts,1992,Chapters4&12)wedevelophypothesesforthejointchoiceofcomponentsthatcomprisetheorganizationaldesign.6Eachofthelinksinthemodelisdevel-opedinturn(seeFig.1).

6Thereissomedebateoverwhetherthesecomponentsareactuallychosensimultaneouslyinorganizations,withregardtoeachother,orarechoseninsomekindofsimplifyingunidirectionalorder(cf.Luft&Shields,2003;Nagar,2002).Thatquestionisnotexaminedinthisstudy.Inaddition,bytakinganorganizationallevelofanalysis,wefocusonvariationsintheuseoforganizationaldesigncomponentsacrossSOEsthatfacediﬀerentlevelsofexogenousinﬂuences.Oursurveydesignrestrictedusfromeﬀectivelyexaminingtheinﬂuencesofseveralpotentialendogenousvariablessuchasstrategy,taskuncertainty,andhumanassetspeciﬁcity(Fisher&Govindarajan,1992).

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177161

Liberalization PoliticalSOE forces H2constraints H3Organizational Industry growthon laborDesign& Foreign decision - Delegation competition making- Objective Export sales performance Joint venture H1measuresexperience - IncentivesStock exchangelisting Fig.1.Theoreticalmodel.UseoforganizationaldesigncomponentsinChineseenterprises

Liberalizationforcesandorganizationaldesigncomponents

Previousstudieshavedevelopedgeneralexpec-tationsfortheSOEadoptionofmanagementaccounting/controlsandhaveexaminedtheinﬂu-enceofliberalizationforcesononlysomeoftheorganizationaldesigncomponentsinthisstudy(e.g.,Firth,1996;OÕConnoretal.,2004).7Toexaminethemediatingeﬀectofpoliticalcon-straints,weﬁrstneedtoproposethemaineﬀectofliberalizationforcesontheorganizationalde-signcomponents.Wedevelophypothesesforindustrylevel,marketopenness,andmarketfunc-tion-relatedliberalizationforces.

Whenanindustryisfreedfromgovernmentcontrolasaresultofliberalization,rapidmarketgrowthusuallyensues.However,thisgrowthmaynotlastverylongbecauseliberalizationalsoencourageslocalandforeigncompetition(Perkins,1994).Followingagencylogic,SOEsinfastergrowingindustriesthathavehigherlevelsoffor-eigncompetition,arelikelytodependmoreontheknowledgeandexperienceofdivisionalman-agersandthusfacehigherknowledgetransfercosts(Jensen,1998).Thedelegationofdecisionauthoritytodivisionalsalesmanagersenablesthemtoeﬀectivelyusetheirinformationadvantage

7Forexample,OÕConnoretal.Õs(2004)constructcomprisesﬁvedimensions,onlytwoofwhich(approvalproceduresandperformancetargets)matchthedimensionsinthisstudy(delegationandobjectiveperformancemeasures).Theyﬁndthatthesedimensionsaresigniﬁcantlyassociatedwithonlyonetypeofliberalizationforce(jointventureexperience).

toquicklyrespondtotheenvironment(Bakeretal.,1994).Thereisalsogreaterpressureonseniormanagementtomakeaccuratejudgmentsaboutmanagerperformancewithrespecttoquality,deliveryandbudgettargets.Li(1997,p.1101)notesthatmarketcompetitionhas‘‘gener-atedconsiderablepressuretoimprovebothcostandquality’’,whichinturnincreasestheneedformoreextensivecontrolssuchasperformancemeasurementsystems.

Pressuretoadoptmoreobjectiveperformancemeasurementandperformance-basedincentivesystemstoretainandattractqualitymanagersismorelikelyinfastergrowingindustriesthathavehigherconcentrationsofforeigninvestedenter-prises(i.e.jointventureorwholly-ownedsubsidi-aries).Insuchindustries,SOEscouldsuﬀerfromthelossofmanagerstohigherpayingcompetitors;asaresult,‘‘someexpertshaveproposedraisingmanagerÕssalariesbasedonannualperformancereviews.’’(Zhongguo,1997,p.1).ForeigninvestedenterprisestypicallyattractthebestChinesegraduatesandhavedevelopedtheleadingexam-plesofmodernmanagersinChina(Branine,1996,p.37).

H1.1:Therewillbeapositiverelationshipbetweenthestrengthofindustry-levelindustrialgrowthandforeigncompetitionandtheuseofdelegation(H1.1a),therelativeuseofobjectiveperformancemeasures(H1.1b),andthelevelofincentives(H1.1c).

Attheﬁrmlevelmarketopenness-relatedliber-alizationforces(jointventurepartneringandre-latedexportopportunities)provideSOEswithﬁnancialincentives(accesstoforeigncapital)andoperationalprivileges(e.g.,priorityinaccessinginnovativetechnologyfromtheforeignpartner,includingWesternmanagementtechniques)thatarelikelytospurtheadoptionofamoreWesternorganizationaldesign.Asjointventuresalsopro-videaccesstoexportmarkets,SOEsareabletogainaccesstofurthercapitalintermsofrefundsofvalueaddedtaxandtheretentionofforeignex-changeearnings.TheseinstitutionalincentivespropelSOEstoallocateimportantresourcestocontrolthedeliveryandqualityofproducts,whichinturnincreasesthescopefortheuseof

162N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

delegation,performancemeasures,andincentivestoplanandcontrolsuchpractices.Forexample,greaterdelegationcanhelpinmoretimelyre-sponsestothequalityandon-timedeliveryde-mandsofexportcustomers.Similarly,motivatingdivisionalmanagers,especiallythoseinproductdivisions,toexportmoreisaprerequisitemeasure,whichmaymakeitnecessaryforSOEexecutivestolinkincentiveswithdivisionalmanagersÕexportperformance,toensurethatthecorporatepriorityinexportisadequatelymaterialized.

Inadditiontoprovidingaccesstoexportmar-kets,jointventuresprovideoperationalprivilegesthroughthesecondmentofstaﬀandexpatriatetraining,whichenablesSOEstolearnperceivedbestpracticestotakeadvantageofsuchexportopportunities(Branine,1996;Firth,1996).JointventuresprovideaccesstoWesternmanagementcontrolsystemsthatwouldnormallybeexpensivetointroduce.Inaddition,theyprovidethetrainingandsupportthatensuresthesuccessfulimplemen-tationofsuchsystems.This,inturn,increasesthelikelihoodthatjointventurepartneringSOEshaveagreatercapabilitytodelegatedecisions,andthattheyhaveagreaterrangeofperformancemeasuresattheirdisposaltouseinevaluatingandrewardingemployees.

GoodallandWarner(1999,pp.25–26)notethatforeign-ChineseJVshavethepotentialtopro-videtheseedsofglobalbargaininginthreepoten-tiallykeyareasofhumanresourcemanagement:employmentcontracts,rewardsystems,andwel-fare/socialinsurance.Firth(1996)andOÕConnoretal.(2004)ﬁndthatSOEswithgreaterforeignjointventureexperiencetendtohavehigherlevelsofuseofWesternmanagementcontrols(e.g.bud-getingandperformancetargets)thandotheirnon-jointventurecounterparts.Firth(1996)alsoﬁndsthatSOEswithgreaterexportsalestendtohavehigherlevelsofuseofWesternmanagementcon-trolsthantheirnon-exportcounterparts.Thesecontrolsincludetheuseofbudgetsetting,which,inthetransitionaleconomysetting,islikelytoin-creasethedegreeofobjectivityintheperformancemeasurement.

H1.2:Therewillbeapositiverelationshipbetweenexportmarketsalesandjointventureexperience

andtheuseoforganizationaldesigncomponentsintermsofdelegation(H1.2a),therelativeuseofobjectiveperformancemeasures(H1.2b),andthelevelofincentives(H1.2c).

Animportantmarketfunction-relatedliberal-izationforceisallowingSOEstolistonstockex-changes(Li,2000).Thesharesofover1000ﬁrmsarelistedandexchangedinChinaÕsstockmarket,withamarketcapitalizationofover2000billionYuan(ChinaSecuritiesRegulatoryCommission,2001).Inaddition,thecentralgovernmentnowal-lowsprivatecompaniestoacquireagreaternum-berofpoorlymanagedlistedSOEs,whichhasloweredtheaveragegovernmentshareholdingto30%(Cheng,2001).StockmarketlistingspressureSOEstoadoptmoreadvancedmanagementsys-tems,suchasperformancemeasurementsystems,toenhanceorganizationaltransparency,eﬃciency,andproductivity(Cheng,2001;Megginson&Net-ter,2001;Pannier,1996).OÕConnoretal.(2004)foundapositiverelationbetweentheuseofman-agementaccounting/controlsandstockexchangelisting.BecausemostSOEsthatarelistedonChinaÕsexchangeshavedecentralizedorganiza-tionalstructures,divisionalmanagersinvariousproﬁtorcostcentersarerealcontributorswhoareresponsiblefortheentirecompanyÕsaccountingperformance(e.g.,returnoninvest-ment,earningspershare,andreturnonassets)andmarketingperformance(e.g.,salesgrowth,marketshare,andassetturnover).Thesecondi-tionsincreasetheopportunitytousemoreobjec-tivemeasuresandtiemanagerrewardstoperformanceoutcomes(Keating,1997;Lambert&Larcker,1987).Inlightoftheabovediscussion,wehypothesize:

H1.3:Therewillbeapositiverelationshipbetweenstockmarketlistingandtheuseoforganiza-tionaldesigncomponentsintermsofdelegation(H1.3a),therelativeuseofobjectiveperformancemeasures(H1.3b),andthelevelofincentives(H1.3c).

Liberalizationforcesandpoliticalconstraints

Thelevelofpoliticalconstraintsvariesaccord-ingtoseveralfactors,suchassize,industrycompe-

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177163

titionandgrowth,export,foreignpartneringopportunitiesandstockexchangelisting.Apartfromﬁrmsize,thevigorofpoliticalconstraintsoverdiﬀerentSOEsisalsolikelytovaryaccordingtocompetition,growth,andtheextentofexportandforeignpartneringopportunities(Linetal.,1998).8Firmsindiﬀerentindustriesandregionsaresubjecttoidiosyncratictreatmentbygovern-mentalpoliciesonhumanresourcemanagement.Forinstance,Chen(2000)arguesthatinsomeindustrieswithlessgrowth,export,andemploy-mentopportunities,thegovernmenthasagreaterinﬂuenceoverhumanresources,therebyconstrain-ingtheadoptionofmoreeﬃcientformsofenter-prise.SOEsincoastalcitieshavegreaterexportandforeignpartneringopportunities,therebyreducingSOEﬁnancialrelianceonthestate.Finally,bywayofdilutionofownership,thegovernmenthaslessinﬂuenceonpubliclylistedSOEs.Forexample,inastudyofChinaÕspubliclylistedﬁrms,Wang(2003)reportsasigniﬁcantpositiverelationshipbetweenthelevelofstateownershipandthedegreeofgovernmentinter-vention.

Theseinﬂuences,however,arenotindependentofeachother.Forexample,duringthe1990sper-iodoftransition,thefastergrowingindustrieswereﬁrstopeneduptoforeigndirectinvestment,andwerethussubjecttohighermarketcompeti-tionthantheirlowergrowthcounterparts(Chen,2000).Hence,bothindustrygrowth(directly)andmarketcompetition(indirectly)inﬂuencethelevelofpoliticalconstraints.Basedonthesefac-tors,wehypothesize:

H2:Therewillbeanegativerelationshipbetweenthestrengthofliberalizationforces(industrialgrowth,foreigncompetition,exportmarketsales,jointventureexperience,andstockmarketlisting)andthelevelofpoliticalconstraints.

8Weincludedsize(numberofemployeesintheSOE)asacontrolvariableintheanalysisbecauseitispositivelyassoci-atedwithbothagencyandpoliticalcostsandhasbeenanimportantcriterionintheSOEreforms(Goodall&Warner,1999;OÕConnoretal.,2004).

Politicalconstraintsandorganizationaldesigncomponents

Directinterferenceinthegreateruseoforgani-zationaldesigncomponentsbyresidentPartysec-retariesisgenerallyprofoundbecausetheoverridingjobofthesesecretariesisthemanage-mentofSOEpersonnel.First,whenPartyrepre-sentativeshavegreatercontroloverpersonnel,theselectionofdivisionalmanagersmaybemadeonpoliticalgroundsandnotproﬁtgrounds.Thiseﬀectivelyreducesthebeneﬁtsof(andhencetheuseof)delegationtomanagerswhoaresupposedtakegreaterresponsibilityforfront-linedecisionsinthecontextofincreasingliberalization.Theresultingcentralizationofdecisions,althoughcon-sideredtobeaseﬃcientinsomecontexts,ismorelikelytocomeatthecostoforganizationaleﬃ-ciency.AccordingtoQian(2001),withthePartysecretaryactingasa‘‘superowner’’forthemain-tenanceofsocialstability,corporategovernancethatpromotesorganizationaleﬃciencyishardtoestablish.

Second,politicalconstraintsdecreasetheinten-tionofSOEexecutivestosharpenmanagerialdis-cretionandimproveorganizationaleﬃciencythroughtheuseofobjectiveperformancemeasuresandincentivesystems(Branine,1996;Peng&Heath,1996).9Partybureaucratsgenerallylacktheabilityandtheincentivestomakedecisionsaboutmanagerialselectionandcompensationaccordingtobusinesscriteriabecausetheyaremainlypoliticallymotivated(Huang&Duncan,1997;Jeﬀerson&Rawski,1994).Therefore,penal-tiesforpoorlyperformingmanagersmaybere-strictedonthebasisofmaintainingastableworkplaceundertheguiseofgradualreform.This,inturn,reducesthevisibilityandcredibilityofobjectiveperformancemeasuresandincentives,thuscurtailingtheiradoptionorlimitingtheireﬀectiveness(Shirley&Xu,1998).

9Politicalconstraintsalsohavethepotentialtodecreaseagencycoststhroughthemonitoringthatisprovidedbythegovernmentrepresentative.Weargue,however,thatthissubstitutioneﬀect(governmentmonitoringoforganizationaldesigncomponents)isalesseﬃcientoptionduetothedualroles(politicalandeconomic)ofthegovernment.Thatquestionwasnotdirectlytestedinthisstudy.

164N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Theseexpectationsarepartlysupportedbyexistingevidence.Forexample,whileLi(1997)ﬁndsthattheincreaseduseofincentivesisrelatedtoincreasesinmarginalandtotalfactorproductiv-ity,ShirleyandXu(1998)donot,andattributethefailureofcontractstopoliticalconstraints(e.g.theinabilityofthegovernmenttofollowthroughonpromisedactionsandtheineﬃcientmonitoringofcontracts).Xuetal.(2002)ﬁndthatpoliticalconstraintstendtodominatelabordecisions,whileotherdecisionsaredominatedbyagencycosts.Thus,althoughpoliticalconstraintsappeartohaveadirectinﬂuenceondelegation,followingagencylogiconecanexpectthatthesupportingcontrolcomponents(objectiveperformancemea-suresandincentives)willbeaﬀected,albeitindi-rectly.Westatethehypothesesintermsofthegeneralexpectationabouttheassociationbetweenpoliticalconstraintsandorganizationaldesigncomponents.

H3:Therewillbeanegativerelationshipbetweenthelevelofpoliticalconstraintsandtheuseofdel-egation(H3a),therelativeuseofobjectiveperfor-mancemeasures(H3b),andthelevelofincentives(H3c).

ResearchmethodsSampleanddatacollection

Wecollectedsurveydatafrom502divisionallevelmanagersin502SOEs.Therespondentsrep-resentedarangeoffunctionssuchasaccountancy(identiﬁedn=68),administration(92),humanre-sourcesmanagement(30),production(88),salesandmarketing(92),andresearchanddevelopment(36).AccesstothemanagerswasobtainedthroughalistofexecutiveMBAalumniwhoworkedinSOEsthatoperatedintwoChineseprovinces(HubeiandGuangdong).Thesurveywasdistrib-utedin1999tofourseparateMBAgroupsacrosstwoMainlandChineseuniversities,andcompletedinclass.WhileitwaspossiblethatmorethanonemanagercouldcomefromoneSOE,thiscaseishighlyunlikelyforthefollowingreasons.First,therespondentsindicatedover20speciﬁc

industries,andthatnotwoSOEsfromthesameindustrywereindicatedtohavethesamesizeorthesameage.Second,basedondiscussionswiththeMBAprogrammeleaders,theselectionprocessiscompetitiveandindividualandtheman-agersarenotsponsoredbytheSOEs.Moreover,thereareover10,000SOEsinthevicinityofthetwouniversitieswheretheMBAprogrammesaretaught.

WeconsideredissuesthatwereassociatedwiththereliabilityandvalidityofresponsesfromMBAstudentswhendesigningthestudy.First,MBAstudentsmightonlyrepresentlesstradi-tionalSOEs,suchasthosefurtherdownthepriv-atizationpath,andnotaverageSOEs.Second,MBAstudentsmightthinkdiﬀerently,andarethuslikelytobeaconcernforstudiesthatseektoexamineindividuallevelphenomenasuchascognitivecharacteristics.Forexample,PriemandRosenstein(2000,p.517)contrasttheÔcausemapsÕofMBAstudentswiththoseofCEOswithoutsucheducationandﬁndthatMBArespondentsexhibitmuchstrongercontingentthinking.Webalancedtheseissuesagainstthediﬃcultyofobtainingpri-marydataonChineseSOEsandthebeneﬁtsofexploringarelevantresearchquestion,giventhelimitedlarge-scalesurveysonSOEmanagementpracticesintheChinareformliterature.

Wealsoconsideredthelevelofanalysisthatwasrequiredinourstudy,andtookstepstotestthevalidityoftheresponses.First,wefocusedtheanalysisattheorganizationlevelratherthantheindividuallevel.Forexample,weaskedtherespon-dentstoassessthelevelofparticularcontrolsinplaceintheirﬁrm,ratherthanaskingfortheirpref-erencesortheirlevelofpersonaljobsatisfaction.Second,wetooktwostepstotestthevalidityofresponsesbycomparisonwithindustrycharacteris-tics(seetheresultssection).Therespondentswerepaidanincentive(100Yuan)tocompletethesurvey.Thiswasnecessarygiventhesurveylength(6pages)andthegreatersenseofmanagertimeurgency.

Measuringinstrument

AsummaryofthesurveymeasuresispresentedinTable1.Thesurveyinstrumentcomprisedsev-

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Table1

Conﬁrmatoryfactoranalysis—variablesusedinthemodel(n=502)Factor1.Size

Abouthowmanyfulltimeemployeesdoesyourﬁrmhave?

2.Liberalizationforces—industrylevelGWHIndustry—averagesalesgrowth(1996–1998)FCMPIndustry—percentageofforeignﬁrms

3.Liberalizationforces—ﬁrmlevelEXPORTWhatpercentageofyourﬁrmÕsoutputisexportedoutofthecountry?JVDoesyourﬁrmhaveajointventurewithaforeignenterprise?Yes(1)No(0)STOKIsyourﬁrmlistedonstockexchange?Yes(1)No(0)

0.7740.929

0.9150.4860.5280.688

165

Factorscore

4.Politicalconstraints

Towhatextentdoeseachofthefollowingparties(people)aﬀectthehumanresourcesdecisions

(hiring,ﬁring,andpromotion)ofcostorproﬁtcentermanagersinyourﬁrm?Theresponsescalerangedfrom1(notatall)to7(averyhighextent)PC1Governmentagency/ministryresponsibleforyourindustry0.648PC2CommunistPartyrepresentative0.847

WhatistheextenttowhichauthorityisgiventotheCommunistPartyrepresentativefromthegovernmentministrytomakethefollowingdecisionsforyourﬁrm?Theresponsescalerangedfrom1(notatall)to7(averyhighextent)PC3Thepromotionofcostorproﬁtcentermanagers0.931PC4Thehiringofcostorproﬁtcentermanagers0.939PC5Theﬁringofcostorproﬁtcentermanagers0.9155.Delegation

Towhatextentisauthoritydelegatedtothecost/proﬁtcentermanagersfromthegeneralmanager(orseniormangers)tomakethefollowingdecisionsfortheﬁrm?A5itemresponsescalerangedfrom0(notatall)to5(averylargeextent)DEC1Developmentofnewproductsandprojects0.463DEC2Thehiringandﬁringofpersonnel0.607DEC3Sourcingofinputs(materialsandparts,etc.)0.692DEC4Settingthebudgetforeachfunctionordivision0.783DEC5Spendingitemsinthebudgetforeachfunctionordivision0.802DEC6Spendingitemsoutsidethebudgetforeachfunctionordivision0.6356.Objectiveperformancemeasures

Whatistherelativeweightgivento(alistoffour)objectivecriteriaor(alistoffour)subjectivecriteriainevaluatingtheperformanceofcostorproﬁtcentermanagers?A5itemresponsescalerangedfrom‘‘100%Obj’’(Objectivemeasuresarethemostimportant)to‘‘100%Subj’’(Subjectivemeasuresarethemostimportant)in25%increments.Thecodingforanalysisrangedfrom0%to100%objectivemeasuresOBJ1Increasingtheirresponsibilities0.674OBJ2Increasingtheirvariablebonussalary—basedonindividualperformance0.804OBJ3Increasingtheirvariablebonussalary—basedoncost/proﬁtcenterperformance0.783

0.713OBJ4Increasingtheirvariablebonussalary—basedonoverallﬁrmperformance

OBJ5Determiningthenewcontractterms(ornon-renewal)attheendofthecurrentcontract0.6767.Incentives

Merit-basedmonetaryincentives

A5itemresponsescalerangedfrom0(notatall)to5(veryhighextent)REW1Rewardsaretiedtoanaccountingearningmeasures(e.g.salesrevenue,

returnonequity,operatingearnings,netincomebeforetax)

REW2Rewardsaretiedtoquantitativemeasures(e.g.productionoutput,

productioncosts,productionqualitytargets,salesgrowthtargets)

0.831

0.838

(continuedonnextpage)

166

Table1(continued)FactorREW3REW4

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Factorscore

TheextenttowhichcompensationcontractsclearlyspecifyhowcompensationisrelatedtomanagerperformancerelativetotheirdepartmentÕsbudgetWhatisthelevelofearningsofthe(a)highest10%and,(b)lowest10%performingmanagers?

A6itemresponsescaleforeachquestionrangedfrom:

$Yuan<=1000,1001–2500,2501–4000,4001–6000,6001–10000,>10000.Thegapbetweentheselevelsofearningswasdividedbythelowestperformingmanagerstoobtaina%score

0.7250.833

eralcategoricalanddescriptivequestionsinaddi-tiontoninepartswithnon-categoricalresponsesthatwereanchoredon5and7pointLikertscales.Severalissuesrelatingtothesurveydesignwereconsidered.First,themeasurementofthecon-structsinthesurveywasbaseduponthoseofsev-eralarticlesinthemanagement(Killing,1983)andmanagementaccountingliterature(Baiman,Larc-ker,&Rajan,1995;Baker,Jensen,&Murphy,1988,1994;Shields&Young,1993).BasedonaprevioussurveyofChineseSOEs(Firth,1996)andactualsitevisits,weexpandedthecontentofthesurveytoincludepoliticalconstraintsandlib-eralizationforces.10Second,itwasimportanttodesigntheques-tionstoreﬂecttheChinacontext.Initially,meet-ingswereconductedwithamainlandChineseprofessorandaVicePresidentofFinancefromalargeprivatecompanyinShenzhen(1998),afterwhichnumerousphonecallsande-mailswereex-changed.Thiscommunicationenabledustoagreeonseveralconstructsthatwereconsideredrelevanttothepaceofhumanresourcemanagementre-forminChineseSOEs,suchasitemsthatcom-prisedpoliticalconstraints.Wewerealsoabletoagreeontheinﬂuenceofseveralliberalizationforcesthathadalreadybeendiscussedintheliter-ature(seeFirth,1996).Anotherissuewasensuringthatthesalaryrangewhichweusedtoobtainre-sponsesaboutthelevelofbaseandbonuspayequatedwiththerangeofSOEsalariesinthetwolocations.

Aspartofanearlierproject,oneoftheauthorsconductedcasestudiesoftheuseofmanagementcontrolmechanismsinfourSOEsinShanghai.Thecasestudiesinvolvedhalf-dayvisitstoeachSOEinDecember1997.

10PriorexperienceincollectingdatainChinaalsoaidedthedesignofthescales.Forexample,basedonpastsurveyresearchinChina,theresearchersfoundthatitwasimportanttoexpresslylabeleachanchoroftheitemsthatrequiredaresponseonaLikertscale.ManyversionsofthesurveywereproducedbasedonthreemeetingsbetweentheinvestigatorsinApril,May,andJune1999.ThisreﬁnementprocessincludedseveraltranslationsfromEnglishintoChineseandthesubsequentbacktranslationofseveralvariablesanditems.11Toensureinternalreliability,abilingualChineseresearchprofessortranslatedtheﬁnalversionofthesurveyintoMandarin.Back-translation,changes,andcorrectionsweremadetothistrans-lationbyanotherChineseprofessor.

Finally,weconsideredcommonresponsebiasissues.Aswithanyattempttocollectandanalyzedatausingasurveyinstrumentthequestionofassessmentandself-reportingarose(Young,2000).Weaddressedthisissueinseveralways.Forexample,weuseddiﬀerentresponsescalesforthediﬀerentsetsofitemsthatpertainedtoeachconstruct.Wealsoreversedtheitemsthatpertainedtoobjectiveperformancemea-sures.Wealsousedarchivaldatatomeasureindustryliberalizationforces(asexplainedinthenextpart).

Liberalizationforcesandpoliticalconstraints

Followingtherecentreformliteratureweiden-tiﬁedfourmainvariablesascomprisingliberaliza-tionforces.Attheindustrylevelthelatentfactor

TheEnglishandChineseversionsofthequestionnaireareavailablefromtheauthorsuponrequest.

11N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Table2

DescriptivestatisticsofvariablesurveyitemsaAlpha

ExogenousfactorsLiberalizationforces

Industry—averagesalesgrowth(1996–1998)bIndustry—percentageofforeignﬁrmsExportsales—percentageofsalesexportJointventureexperienceStockexchangelistingcPoliticalconstraints

OrganizationaldesignDelegation

ObjectiveperformancemeasuresIncentives

Incentives—%Gapb/wtopandbottom10%Controlvariable

Size—fulltimeemployees

ab167

MeanSDMaxMin

0.9210.7400.7990.7820.100d15.43%14.14%17.82%15.94%22.51%12.3317.68268.94%8.7652.91%3020

11.79%9.77%38.32%36.64%41.81%7.504.8584.99%2.7868.20%9395

55%33%100%100%100%25.0030.00500%15.00350.00%80000

À20%2%0%0%0%0.000.000.00%0.000.00%400

Allwithin-organizationmeasuresrelatetothedivisionalcost/proﬁtcentermanagerlevel;n=502.

Over20industrieswererepresentedinthesample.Thelargestrepresentationswere:ElectronicsandCommunicationEquipment(88),Non-metalMineralProducts(77),TransportationEquipmentManufacturing(55),Construction(38),ChemicalRawMaterialandChemicalProducts(31),SpecialEquipmentManufacturing(14),Textiles(9),Electricity,Steam,HotWaterProductionandSupplies(8),Trading(7),MedicineManufacturing(6),ManufacturingofFoods(4),ElectricalMachineryandEquipment(2),TobaccoProcessing,Instruments,Meters,EducationalandOﬃceEquipment,EducationalandSportsProductsManufacturing.Averagesalesgrowthforthethreeyears(1996–1998),andthenumberofforeignﬁrmsineachindustrywastakenfromtheChinaMarketsYearbook(2000).72.87%ofthesamplewaslocatedinGuangdong(consideredasaspecialeconomiczone),while27.13%waslocatedinHubei(consideredasaninlandprovince).cIndicatesthatadichotomousresponse(forexample,yes(1)orno(0))wasobtainedforthisvariable.dPearsoncorrelationbetweensalarygapbetweentopandbottom10%performingmanagersandincentivesscale(REW1ÀREW3)(p=0.026).

ofindustryliberalizationforceswasmeasuredwithtwoobservablemeasures:industrysalesgrowthandforeignﬁrmconcentration.IndustrysalesgrowthandtheextentofforeignﬁrmconcentrationweretakenfromtheChinaMarketsYearbook(2000),whichconsistsof550industrialcodes(atthe2digitlevel).Twentydiﬀerentindustrycodeswererepresentedinourdata(seethefootnoteinTable2).Foreachindustry,themarketgrowthforthethreeyearsto1998wasaveraged.Forfor-eignﬁrmconcentration,thepercentageofforeignﬁrmsineachindustrywascomputed.Weusedthismeasuretoproxyfortheextentofforeigncompe-titioninaparticularindustry.

Wealsomeasuredthreemainliberalizationforcesattheﬁrmlevel:percentageofsalesex-ported,foreignjointventureexperienceandstockexchangelisting.Thepercentageofsalesexportedwasmeasuredbyaskingrespondentswhatper-

centageoftheirﬁrmÕsoutputwasexportedoutofthecountry.Foreignjointventureexperiencewasmeasuredbyaskingrespondentswhethertheirﬁrmwasinajointventurewithaforeignﬁrm.Adummyvariableof1wasallocatedtothoseSOEsinajointventureand0tothosethatwerenot.12Asforeignjointventureexperienceandsalesexportopportunitiesarepartlyafunctionoflocationandthesampleusedinthisstudywasgainedfromtwodiﬀerentlocations(HubeiandGuangzhou)weincludedlocationasacontrolvariableinthisstudy.Adummyvariableof1wasallocatedtothoseSOEsinGuangdong,whichisacoastalprovinceandspecialeconomiczone,andazerowasallocatedtothoseSOEslocatedinHubei,whichisaninlandprovince.Whenlocationwasenteredintothemodelasanindependentvariableofthefourliberalizationforcesandpoliticalconstraints,thesignandsigniﬁcanceoftheparameterestimatesforallofthehypoth-esizedpathsremainedunchanged.

12168N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

StockmarketlistingcomprisedlistingontheShanghaiorShenzhenStockExchanges,andweusedadummyvariablewiththevalueof1iftheSOEwaslisted,and0ifnot.

Politicalconstraintsweremeasuredwithﬁveob-servablemeasuresthatrelatedtohumanresourcedecision-makingauthority.Therewereﬁveitemsinresponsetotwoquestions.Theﬁrstquestionaskedabouttheextenttowhichthegovernmentagency/ministrythatwasresponsiblefortheindus-try(item1)andtheCommunistPartyrepresenta-tive(item2)aﬀectedthehumanresourcedecisions(hiring,ﬁring,andpromotion)ofcostorproﬁtcentermanagersintheﬁrm.Thesecondquestionaskedabouttheextentofauthoritythatwasgiventotherepresentativefromthegovernmentminis-trytomakethefollowingdecisionsfortheﬁrm:thehiring(item3),promotion(item4),andﬁring(item5)ofcostorproﬁtcentermanagers.There-sponseswereanchoredonasevenpointLikertscalethatrangedfrom1(notatall)to7(averyhighextent).TheCronbach(1951)alphafortheitemswas0.92.

Size(numberofemployeesintheSOE)wasin-cludedasacontrolvariableintheanalysissinceitispositivelyassociatedwithbothagencyandpolit-icalcostsandhasbeenanimportantcriterionintheSOEreforms,asevidencedbygovernmentinterventioninlargerSOEs,whichcarryaheavierburdenintheformofsocial-welfarecosts(Goo-dall&Warner,1999;Lee,2001;Linetal.,1998).Thenaturallogarithmofthenumberofemployeeswasusedintheanalysis.

Organizationaldesigncomponents

Threeorganizationaldesigncomponentswereexaminedinthisstudy(delegation,objectiveper-formancemeasures,andincentives).WeadaptedtheKilling(1983)delegationmeasure,andaskedtherespondentsabouttheextenttowhichauthor-itywasdelegatedbytheseniormanagement(e.g.,thegeneralmanagerortheboardofdirectors)tocostorproﬁtcentermanagersindecision-makingareasthatwererelatedtothedevelopmentofnewproductsandprojects,thesourcingofinputs,thehiringandﬁringofpersonnel,budgetsetting,andbudgetandnon-budgetspending.Thesixitemswereextractedfromnineitemsinthesurveyonthebasisoffacevalidity(budgetresponsibilityorientation)andconﬁrmatoryfactoranalysis(seethenextsection).A5itemresponsescalerangedfrom0(notatall)to5(averylargeextent).TheCronbach(1951)alphafortheitemswas0.74.Asacheckofdiscriminantvalidityitwasimportanttodistinguishbetweenpoliticalcon-straintsanddelegationvariablesbecauseitemsthatrelatetohiringandﬁringappearinbothvari-ables.However,thecontextofeachisdiﬀerentintermsof:(i)whohasthepower—forpoliticalcon-straintsitistheextenttowhichthePartymemberaﬀectshiring,promotion,andﬁringdecisions,andfordelegationitistheextenttowhichauthorityisdelegatedtoproﬁt/costcentremanagers;and(ii)thescopeofeachvariable—delegationhasabroaderscopeinthatitseeksthelevelofdelega-tiononagreaterrangeoffactorsthanhumanresourcesmanagement.Thelownegativecorrela-tionbetweenthesetwomeasures(À0.219,seeTable3)alsoprovidessupportforthediscriminantvalidity.Aswillbeshownlater,allofthecon-structsloadedontoseparatefactors.

Wemeasuredtheextenttowhichobjectiveper-formancemeasureswereusedbyaskingabouttherelativeweight(between0%and100%)thatwasgiventoobjectiveorsubjectivecriteriainevaluat-ingtheperformanceofcostorproﬁtcenterman-agersforthepurposesof:(a)increasingtheirresponsibilities;(b)increasingtheirvariablebonussalarybasedonindividualperformance;(c)increasingtheirvariablebonussalarybasedoncost/proﬁtcenterperformance;(d)increasingtheirvariablebonussalarybasedonoverallﬁrmperfor-mance;and(e)determiningnewcontractterms(ornon-renewal)attheendofthecurrentcontract.Tomakethiscleartotherespondents,wegavethemalistofobjectivecriteria(fourbudgetandaccount-ingitemstakenfromtheHopwood(1972)budgetemphasismeasure)andsubjectivecriteria(fourqualitativeitemstakenfromthesamework).TheresponsewassoughtonaﬁvepointLikertscalethatrangedin25%incrementsfrom‘‘100%Obj’’(objectivemeasuresarethemostimportant)to‘‘100%Subj’’(subjectivemeasuresarethemostimportant).Transformationsofthescaleweremadetoindicatethelevelofobjectivemeasures,thusproducingarangefrom0%to100%objective

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Table3

Constructanderrormatrixes

Pearsoncorrelations

PoliticalIndustryExportconstraintsaliberalizationsales

forces

Industry

liberalizationforcesaExportsalesJointventureexperience

StockexchangelistingDelegationaObjective

performancemeasuresaIncentivesaSize

À0.097**À0.008À0.261**À0.138**À0.229**À0.203**À0.0300.201**

0.137**À0.0240.104**0.0610.0720.158**À0.067*

169

JointventureStockDelegationObjectiveIncentivesexperienceexchangeperformance

listingmeasures

0.0490.231**

0.254**À0.012

0.181**0.135**0.172**

0.119*0.0740.0370.122*

0.256**0.582**0.053

0.499**0.046

0.236**À0.0160.0880.083

À0.204*

*p<0.05;**p<0.01.

aVariablesrepresentfactorconstructscoresfromthesevenfactormodel;n=502.

measuresforeachoftheﬁveitems.TheCronbach(1951)alphafortheﬁveitemswas0.80.13Theincentivesmeasurecomprisedtwocompo-nents:theShieldsandYoung(1993)incentivesinstrumentandtheaveragetotalmonthlyincomerangeforallcostorproﬁtcentermanagers.TheShieldsandYoung(1993)instrumentcomprisedthreeitems:theextenttowhichcompensationwasrelatedtomanagerialperformance;theextenttowhichmanagersinthetop25%ofperformersweregivenlargerrewardsthanthoseinthebottom25%;andtheextenttowhichﬁnancialrewardsin-creasedasactualperformanceexceededbudgetedperformance.Aﬁveitemresponsescalerangedfrom0(notatall)to5(verylargeextent).TheCronbach(1951)alphaforthethreeitemswas0.78.Fordeterminingtheaveragetotalmonthlyincomerangeforallcostorproﬁtcentermanag-

ers,weaskedtherespondentstoindicatethein-comelevelfor(a)thehighestperforming10%ofmanagersand(b)thelowestperforming10%ofmanagers.A6-itemresponsescalerangedfromYuan61000,1001–2500,2501–4000,4001–6000,and6001–10,000to>10,000.Thediﬀerenceinthesetworesponses(2setsof6-itemscales)wasdi-videdbythelowestperforming10%levelofmonthlyincometoobtainapercentagescore.Thedescriptivestatisticsofthevariablesarein-cludedinTable2.

Samplevalidityandreliability

Therespondentshadanaveragetermofemploymentof7years.ThissuggeststhattheyhadadequateSOEknowledgewithwhichtoan-swerthesurveyquestions.BecauseitispossiblethatmanagersÕknowledgeconcerningpoliticalconstraintsandpartsoforganizationaldesignmayvaryaccordingtothedivisionthattheyareinorthesizeoftheSOE,wemeasuredtheirdegreeofconﬁdenceintheiranswerstoitemsthatcom-prisefourvariables(politicalconstraints,delega-tion,objectiveperformancemeasuresandincentives)usinga5pointLikertscale(responsescale:1=verysmallextentto5=verylargeextent).Theaverageresponseacrossthefour

Asafurtherindicationofthereliabilityofthemeasure,wecitedtheactualrangeinresponseacrosstheﬁveitems.Withthetheoreticalrangebeing100%(i.e.,anindicationofsubjective(i.e.,0%)foroneitemandanindicationofobjective(i.e.,100%)foranotheritem)therewouldbeaconcernifasigniﬁcant(5%)portionofcaseshadarangethatwasmorethan50%.Outof502cases,in21(4.2%)therewasadiﬀerenceof75%,andinone(0.2%)therewasadiﬀerenceof100%.

13170N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

variablesrangedfrom3.233to3.750.WhilewefoundnosigniﬁcantdiﬀerenceintheaveragelevelofrespondentconﬁdenceacrossvarioussamplesthatweresplitaccordingtotheﬁvemaindivisionsandbetweenlargeandsmallSOEs,theaveragelevelofrespondentconﬁdencewasthelowestforthemeasureofincentives,thereforerelatedﬁnd-ingsmustbeinterpretedwithsomecaution.

TheSOEsamplehadanaveragesizeof3020employees,exportedanaverage18%ofsales,andexhibitedanaverage15%salesgrowth.About16%oftheSOEswereapartnerinaforeign-Chi-nesejointventureand23%oftheSOEswerelistedonthestockexchange.Theaveragemonthlyman-agersalaryrangedfrom250to11,000Yuan,withameanof3395Yuan.Thisiscomparabletothelowermonthlysalaryof1500YuanthatGoodallandWarner(1999)reportfortheworkersin38SOEs.ComparisonswithindustrystatisticswereundertakentoprovidefurtherinsightintothevalidityofsomeSOEcharacteristics.Forexample,wecorrelatedthesalesgrowthoftheSOEwiththesalesgrowthcharacteristicsoftheSOEÕsindustryandobtainedalowbutsigniﬁcantPearsoncorrela-tionof0.12(p<0.01),whichisconsistentwiththeexpectationthatSOEsinhighergrowthindustrieswillexhibit,onaverage,higherlevelsofsalesgrowth.

Wealsomeasuredthepercentagelevelofbonus-basedsalary(bonuspay/basepay*100)byaskingrespondentstoindicatetheaveragelevelofbonususingthesame7itemscaleasbasepay(rangingfromlessthan500Yuantomorethan10,000Yuan).ThePearsoncorrelationbetweenthismeasureandthemeasureoftheaveragetotalmonthlyincomerangepercentagewas0.138(p<0.05).Substitutingthismeasureforthebonusmeasureinthemodelresultedinnodiﬀerenceinthesigniﬁcanceoftheparameterestimates.Structuralequationmodel

Astructuralequationmodeling(SEM)ap-proachwasusedtoanalyzethesurveyandarchi-valdatausingtheSPSSxAMOS4.0statisticalpackage.Thefollowingequationsweresimulta-neouslytestedtoestimatetheparametersfortheinitial(unconstrainedpath)model.

DEL¼B0þB1OBJþB2INCþB3COMP

þB4EXPþB5JVþB6STOKþB7PCþB8SIZEþe

ð1Þ

OBJ¼B0þB1DELþB2INCþB3COMP

þb4EXPþB5JVþB6STOKþB7PCþB8SIZEþe

ð2ÞINC¼B0þB1DELþB2OBJþB3COMP

þb4EXPþB5JVþB6STOKþB7PCþB8SIZEþe

ð3ÞPC¼B0þB1COMPþB2EXPþB3JV

þB4STOKþB5SIZEþeð4ÞwhereDELdelegation

OBJobjectiveperformancemeasuresINCincentivecompensation

COMPindustrygrowthandcompetitionEXPexportsales

JVjointventureexperience(yes(1),no(0))STOKlistedonstockexchange(yes(1),no(0))PCpoliticalconstraints

SIZE

ﬁrmsize(logofnumberofemployees,usedasacontrolvariable)

TheSEMspeciﬁestherelationshipsbetweentheunobservedconstructs.Similartoasetofregres-sionequations,themodelisusedtodescribetheassociationsbetweenconstructs.However,incon-trasttoregressionmodels,thestructuralequationmodellinksunobserved(latent),hypotheticalcon-structsratherthanconcrete,empiricalindicators.Themodelinthisstudycomprisedﬁvelatentvari-ables(DEL,OBJ,INC,COMP,PC)andfourob-servedvariables(EXP,JV,STOK,SIZE)(seeFig.2).TheprincipaladvantageoftheSEMapproachisthatitgoesbeyondconventionallinearmodelsandaccountsformeasurementerror,allowsforsimultaneousestimatesofmeasurementandstruc-turalparameters,andhenceprovidesdiagnosticstatistics/informationforthemodelasawhole(Joreskog&Sorbom,1979).AnotheradvantageofSEMwastheabilitytomodelbi-directional

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177171

Industry growth and Foreign competitionH2a--.17 (2.47)*DelegationExport salesH1.2a+.23 (3.31)**H3a--.30 (2.72)**.59 (4.03)**Political constraintsH3c-.30 (2.15)*IncentivesJoint venture experienceH2c--.36 (6.40)**.14 (2.05)*H3b--.29 (3.26)**Objective performance measures.42(3.73)**Stock exchange listing H2d--.14 (2.45)**Fig.2.Parameterestimatesofthemodiﬁedmodel.*p<0.05;**p<0.01.Theunconstrainedpathmodelincludesallofthepaths(dottedandsolid).Thesolidlinesformthemodiﬁedmodelofbestﬁt.Standardizedcoeﬃcients(b)andplevelsaredisplayed(two-tailtest)forthesigniﬁcantpaths.Theparameterestimatesofthesizecontrolvariable(notshown)weresigniﬁcantforpoliticalconstraints(b=0.23**)andincentives(b=À0.38**).v2=395.894(261d.f.),AGFI=0.918,CFI=0.878andIFI=0.885.relationshipsbetweentwoormoreconstructs(seeByrne,2001,p.120),whichisanassumptionoftheorganizationaldesignframework(Milgrom&Roberts,1995).Assomeofourdataweremea-suredusingordinalscales(e.g.,delegationandobjectiveperformancemeasures)weusedtheGLSestimationmethod.AccordingtoGolob(2001,p.3),GLSestimationmethodsareusefulfordealingwithdiscretechoicevariablesandordi-nalattitudescales(suchastheLikertscale).14Conﬁrmatoryfactoranalysis(CFA)

Inbuildingtheproposedsystemofrelation-ships,webegantheanalysisbyexaminingonlythemeasurementpropertiesoftheinitialmodel.WeconductedCFAtotestthegoodness-of-ﬁtofthemodelthatcomprisedsize,liberalizationforcesattheindustry(growthandforeignmarketcompe-Thereisdebateintheliteratureabouttheappropriatenessofusingordinalmeasures(includingtheuseofdummyvariablestorepresentdiﬀerentcategoriesinaparticularorder)(Golob,2001).ForgeneraldiscussionsofGLS,seeBrowne

´n(1984).(1984)andMuthe

14tition)andﬁrm(exportsales,jointventure,stock

exchange)level,politicalconstraints,delegation,objectiveperformancemeasures,andincentives(Bollen,1989).Theoverallmodelﬁtcouldbeas-sessedbytheChi-squaretest,andheuristicallybyanumberofgoodness-of-ﬁtindices:theadjustedgoodness-of-ﬁtindex(AGFI),therootmeansquareresidual(RMSR),theincrementalﬁtindex(IFI)andcomparativeﬁtindex(CFI)(Anderson&Gerbing,1988;Bollen,1989).Theﬁtindicesindi-catetheextenttowhichtheassociationsbetweentheunobservedendogenousandexogenouslatentvariablesandtheirobservedindicatorsinthehypothesized(measurement)modelaresigniﬁ-cantlydiﬀerentfromarandomassociationsmodel.Valuesofgreaterthan0.80provideanacceptablelevelofincrementalﬁt(Carmines&Mclver,1981).Weperformedseveralcheckstogaugethepo-tentialthreatofmulticollinearityinthedata.First,wereviewedthecorrelationsamongthevariablesshowninTable3.Thecorrelationbetweentheindependent(dependent)variableswiththegreat-estmagnitudewas.261(.582),whichisbelowthecriterionof.80forthelevelofcorrelationthatindicatesaseriousmulticollinearityproblem

172N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

(Kennedy,1979).Asasecondcheckweranthefactoranalysiswithavarimaxrotationandsevenfactorsemergedthatcorrespondedtothevariablesinthemodel.This,alongwiththesigniﬁcantﬁtin-dexesfortheconﬁrmatoryfactoranalysismodelindicatedthatmulticollinearitywasnotaproblem.Table1showsthefactorloadingsforeachofthesevenconstructs.Althoughtheinitialfactoranalysisresultedinsevenfactors(thethreeliberal-izationforcesattheﬁrm-levelloadedontothesamefactor),weexpandedthemodeltoconsistofninefactors(ﬁvelatentandfourdirectlyob-servedvariables).Thisenabledustotestthehypothesesbyexaminingtheseparatepathsper-tainingtoeachoftheﬁrm-levelliberalizationforces.Theninefactor/variablemodelshowedgoodlevelsofﬁtagainstthebaselinemodelsandagainstthebenchmarkssetintheliterature(Bago-zzi&Yi,1988).15Thegoodness-of-ﬁtindexeswerewithintolerableranges:(e.g.AGFI=0.917,RMSR=0.033,IFI=0.890,CFI=0.883).Allofthefactorloadingsofthemeasurementinstrumentexceededthe0.4levelthatiscommonlyconsideredmeaningfulinfactoranalysis(Ford,MacCallum,&Tait,1986).Thebi-variatecorrelationsofthefactors/variablescoresprovideaninitialindicationoftherelationshipsbetweenthemeasuredandla-tentvariables(seeTable3).

Results

Totestthehypotheses,weﬁrstassessedthemodelusingaseriesofnestedmodelsbeginningwiththeleastconstrainedmodel—onethatin-cludedallofthepathsthatwereshowninFig.2.16Todothis,someofthepathsinthemodelwereconstrained(i.e.,setequalto0),whichpre-15WemadeseveraladjustmentstotheparametersinﬁttingtheninefactorCFAmodel.Weﬁxedthecovarianceinerrortermsforseveralitemsthathadsimilarfacevalidityandthesameresponsescale.Weﬁxedtheerrortermforallitemsforpoliticalconstraintsanddelegation.Weappliedthesameconstrainttotwoobjectiveperformancemeasureitems(OBJ2andOBJ3)andthreeincentivesitems(REW1,REW2andREW3).16Tocontrolforsize,thepathsbetweensizeandpoliticalconstraintsandeachoftheorganizationdesignvariableswereincludedinallofthemodels.

ventedthemfromsubsequentlybeingestimated.Thesequenceofnestedmodelswasdeterminedbyeliminatingtheleastsigniﬁcantparameter(set-tingthepathequalto0).FollowingtheprocedurethatAndersonandGerbing(1988)suggested,wecomputedtheChi-squarediﬀerencesbetweeneachnestedmodelandtestedthemforsigniﬁcancebytakingintoaccountthediﬀerenceinthedegreesoffreedom.Asarule,ifchangeintheChi-squareisnotsigniﬁcant,themodelwiththeconstrainedpathisabetterﬁt(Anderson&Gerbing,1988).AsigniﬁcantchangeintheChi-squareindicatesthattheconstrainedpathshouldnotberemoved.Wecontinuedthisprocessuntilnofurtherimprovementscouldbemadeandthemodelwasthemostparsimonious.Thisanalysisproducedthefollowingmodiﬁedhypothesizedmodel(shownwiththesolidlinesinFig.2).Thegoodness-of-ﬁtindexes(e.g.AGFI=0.918,RMSR=0.032,IFI=0.885,CFI=0.878)indicatedthatthemod-iﬁedmodelﬁttedthedataverywell.

Themain-eﬀecthypotheses(H1)proposedapo-sitiveanddirectassociationbetweenliberalizationforcesandeachoftheorganizationaldesigncom-ponents.Onlyone(exportsalesanddelegation)outof12possibleassociationswassigniﬁcantandpositive(H1.2a,b=0.23,p=0.001).Theremaininghypotheses(H2andH3)predictedthatthemaininﬂuenceofliberalizationforcesonorga-nizationaldesignwouldbeindirectviapoliticalconstraints.Fortheseindirectrelationships,thesignsandsigniﬁcancelevelsoftheparameteresti-matesprovidedsupportforﬁveoutofthesixsub-hypotheses.Threeofthefourcausalpathsbe-tweenliberalizationforcesandpoliticalconstraintswerenegativeandsigniﬁcant.Thesewereindustrygrowthandcompetition(H2a,b=À0.17,p=0.013),jointventureexperience(H2c,b=À0.36,p=0.000)andstockexchangelisting(H2d,b=À0.14,p=0.014).Finally,hypothesesthreepredictedthatpoliticalconstraintswouldbenegativelyrelatedtoeachoftheorganizationaldesigncomponents.InresultsnotpresentedinFig.2,thepathbetweensizeandpoliticalconstraintswaspositiveandsigniﬁcant(b=0.23,p=0.000),whilethepathbetweensizeandincentiveswasnegativeandsigniﬁcant(b=À0.38,p=0.001).Andwhensizewascontrolled,thereremainedsig-

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177173

niﬁcanteﬀectsofpoliticalconstraintsonorganiza-tiondesigncomponents.Whilethenegativepathsthatlinkedpoliticalconstraintsanddelegation(H3a,b=À0.30,p=0.007)andobjectiveperfor-mancemeasures(H3b,b=À0.29,p=0.001)wereconsistentwithexpectations,thepositivepathbetweenpoliticalconstraintsandincentives(H3c,b=0.30,p=0.032)wasoppositetoexpecta-tions.17Theimportanceofpoliticalconstraintsasamediatingvariableisfurtherillustratedbyexamin-ingthebi-variateassociationsbetweenliberaliza-tionforcesandorganizationaldesigncomponents(seeTable3),inwhichthereisasig-niﬁcantdirectrelationinsixoutoftwelvecases.Infourofthesesigniﬁcantrelations(excludingex-portsales)mostoftheobservedcorrelationisex-plainedduetotheindirecteﬀectviapoliticalconstraints.AspredictedinH2,politicalcon-straintsformanimportantmediatingfactorintheserelationships.Forthesixnon-signiﬁcantbi-variateassociations,ourﬁndingsindicatethattheincreaseduseofdelegationandobjectiveper-formancemeasuresarisessolelyduetothelowerpoliticalconstraintsinresponsetoliberalizationforces(industry,jointventureexperience,andstockexchangelisting),aftercontrollingforsize.Thestrengthofthepoliticalconstraints-organiza-tionaldesignlinksrelativetotheliberalizationforceslinksfordelegationandobjectiveperfor-mancemeasuresisconsistentwithourhypothesis

17WhiletheSEMapproachwaschosentoanalyzethedata,wealsotestedthesamerelationshipswithaseriesofregressionmodels,oneforpoliticalconstraintsandforeachoftheorganizationaldesigncomponents.ThesignsandsigniﬁcanceoftheexpectedassociationswereconsistentwiththeSEMmodelinmostofthecases.ConsistentwiththeSEM,theregressionmodelforpoliticalconstraintswassigniﬁcant(AdjR-square=0.19)withtheestimatedcoeﬃcientforsize(p<0.01)andthreeoutofthefourliberalizationforces(JV,STOKandCOMP)beingsigniﬁcant(p<0.05).Similarly,intheregressionsforbothDECandOBJ,theestimatedcoeﬃcientforPCwasnegativeandsigniﬁcant(p<0.05).ForDEC,SIZEandEXPwerealsopositiveandsigniﬁcant(p<0.05),whileforREW,thecoeﬃcientforOBJwaspositiveandsigniﬁcant.ContrarytotheSEM,theestimatedcoeﬃcientsforJVandREWintheregressionmodelforOBJweresigniﬁcant.Finally,intheregressionofREW,thecoeﬃcientforEXP(butnotDEC)waspositiveandsigniﬁcant(p<0.05).

thatpoliticalconstraintsandnotliberalizationforcesaredominantinorganizationaldesigninChineseSOEs.

Infocusingontheendogenousinﬂuencesbe-tweenthecomponents,thepartialresults—threeofthesixpaths(OBJtoDEL,OBJtoINC,andDELtoINC)weresigniﬁcantly(p<0.05)andpositivelyrelated—indicatethatwhilethechoiceofdelegationisafunctionofobjectiveperfor-mancemeasures,thechoiceofincentivesisafunc-tionofbothdelegationandobjectiveperformancemeasures.Thisevidencesuggeststheexistenceofsomeheterogeneityamongtheorganizationalde-signcomponents.Thatis,noteachofthesecom-ponentsisassociatedwithpoliticalconstraintsandliberalizationforcesinthesameorequalman-ner.Indeed,theabsenceofasigniﬁcantrelation-shippertainingtotheinﬂuenceofincentivesonboththelevelofdelegationorobjectiveperfor-mancemeasuresintheindirectmodelmightreﬂectthetransitionaleconomycontext,wherethelevelofriskthatisbornebytheagentmightnotbeaprimaryorganizationaldesigncriterion.18Overall,thestrengthofthepoliticalconstraints-organizationaldesignlinksrelativetotheliberal-izationforces-organizationaldesignlinksisconsis-tentwithourhypothesisthatitispoliticalconstraintsandnoteﬃciencythatisthedominantimperativedrivingorganizationaldesignintheChinareformsetting.

Conclusion

Themodelthatwedevelopinthispaperpro-videsageneraltheoreticalframeworktoexplainthedeterminationandevolutionofChineseSOEorganizationaldesign.Themodelrelatestheuseofthreeorganizationaldesigncomponentstoﬁve

18InatestoftheclusteringeﬀectofthethreeorganizationaldesigncomponentswecomparedtheﬁtoftheﬁnalmodelinFig.2withtheﬁtofamodelinwhichthejointpathsbetweenthethreecomponentswereallowedtobefreelyestimated.ThenewmodelhadabetterﬁtwithaChiSquareof386.651,whichwasadiﬀerenceofninefortheextrathreedegreesoffreedomgiven.Whilethisindicatesapossibleclusteringeﬀect,ourdataislimitedinverifyingtheassumptionthatthethreecomponentsarejointlydetermined.

174N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

determinants:liberalizationforces(industry,ex-portsales,jointventureexperience,stockexchangelisting)andpoliticalconstraintsoverlabordeci-sion-makingintheSOE.Thismodelishelpfulinunderstandingvariationsintheorganizationalde-signamongChineseSOEs.Variousmarketliberal-izationforceshaveprovedtobeadeterminantfactorinthestructuringofSOEs.LiberalizationforcesstimulateSOEstosearchformoreeﬃcientmanagementcontrolforms,whichincludethedel-egationofdecisionauthoritytodivision-levelmanagers,andthemoreextensiveuseofobjectiveperformancemeasuressuchasbudgetsand(indi-rectly)theprovisionofincentives.Incontrast,politicalconstraintsslowdownthetransitionofSOEstowardamoreeﬃciententerprise.AsChen(2000,p.47)pointsout,‘‘thegovernmenthasstrongincentivestodeviatefromtheproﬁtmaxi-mizationobjectivetopursueitsowngoals,suchasemploymentprovision,materialbalanceininput-output,tradepromotion,politicalcon-straintsandstabilization,etc.,atthecostofﬁrmeﬃciency.’’

Threekeyﬁndingsandcontributionshaveemerged.First,wefoundstrongersupportfortheindirectpathassociationsinwhichpoliticalconstraintsmediatetherelationshipbetweenliber-alizationforcesandtheuseofthreeorganizationaldesigncomponents.Withinthismodelourﬁndingthatpoliticalconstraintsareadirectnegativedeterminantofdelegationandobjectiveperfor-mancemeasuressupportsourargumentthatsuchcomponentshavethepotentialtodilutepoliticalconstraints.Thedirectpositiveinﬂuenceofpoliti-calconstraintsonincentivesmightbebecauseincentiveshavethepotentialtosupportthestatusquoifthebasisforincentivedeterminationiscon-sistentwiththeobjectivesofthepoliticalconstraint.

Second,theresultsextendtheﬁndingsofFirth(1996)andOÕConnoretal.(2004)ontheinﬂuenceofliberalizationforcesontheSOEadoptionofWesternmanagementcontrols.Whilebothprevi-ousstudiesfoundlinearrelationshipsbetweenaspectsofmarketliberalization(marketcompeti-tionandjointventureexperience,Firth,1996;stockexchangelistingandjointventureexperi-ence,OÕConnoretal.,2004)andtheadoptionofmanagementaccountingmechanisms,thisstudyfoundthatthelevelofpoliticalconstraintsmedi-atestheinﬂuenceofthesethreefactors.Thenega-tivemediatinginﬂuenceofpoliticalconstraintsappearstobeconsistentwithourtheorythatsuchinterferenceimposesadditionalcostsonseniormanagementtodelegatemoreandtoimplementmoreobjectiveperformancemeasurement.Suchinterferencealsoimposescostsondivisionalman-agersduetothelimitedtransparencyandscopeforactionrequiredtomeetcertainperformancetargets.

Third,theresultscontributetotheChinaman-agementliterature(Goodall&Warner,1999;Groves,Hong,McMillan,&Naughton,1994;Li,2000;Shirley&Xu,1998;Xuetal.,2002).Inpar-ticular,ourresultsareconsistentwiththeﬁndingsofShirleyandXu(1998)andXuetal.(2002),whosuggestthatthegovernmentdelegationofhumanresourcedecisionsisimportanttothecontinuedpaceofChinesestate-ownedenterprisereform.ForpolicymakersandotherswithaninterestinChina,andperhapsothercommandeconomies,theseﬁndingssuggestthatindustrygrowth,for-eigncompetition,jointventureexperienceandstockexchangelistingcanbepowerfulforcesintheadoptionofWesternorganizationaldesigncomponents.However,theyalsosuggestthatthechangeprocesscanbehamperedbyinstitutionalfactorssuchasgovernmentinvolvementinman-agement.Thisﬁndingsigniﬁcantlyaddstoourunderstandingofthetransitionaleconomycon-text,wherebythestateretainssigniﬁcantinﬂuenceinenterprisesthataredeemedtohavebeenpriv-atized.TheprocessoftransitionofSOEscontrastswiththeprivatizationﬁrststrategyfollowedinothertransitionaleconomiessuchasEasternEur-opeandtheformerrepublicsoftheSovietUnion.AccordingtoMcMillanandNaughton(1992),oneoftheuniquefeaturesofChinaÕsreformstrategyliesinitsslowdevelopmentofinstitutions(includ-ingmarkets)thatleadtogreatercompetitivepressure.

WhilethecurrentstudyhascontributedinsightsintothemoveofChineseSOEstowardtheuseofWesternorganizationaldesigncomponents,sev-erallimitationsareacknowledged.First,thestudyofthedesignofcontrolsystemsasadependent

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177175

variableisjustiﬁedonthebasisthattheeconomyisinequilibriumandallformsareperformingoptimally.IntheChinareformcontextwhereﬁrmsareinaconstantstateofﬂux,thisassump-tionislessrealistic.

Second,theﬁndingsaremainlybasedonman-agersÕperceptionsasreﬂectedintheirresponsestoasurvey,thusposingcommonmethodbias.Thenatureofthisbiasalsoprecludedusfromeﬀectivelyexaminingtheinﬂuencesofotherpoten-tialendogenousvariablessuchasstrategyandtaskuncertainty(Fisher&Govindarajan,1992);andpersonalfactors(e.g.,skillsandaspirations)orevenculturalfactors(e.g.,traditionalbeliefsaboutmanagement).Thereisaneedtotriangulatewithotherdatacollectiontechniques,suchasexamin-ingtheenterprisesÕprocedureandpolicymanualsandotherinternaldocuments.Third,theuseofaconveniencesamplejeopardizestheextenttowhichthesampleisrandom,andtheresultinginferencesthatcanbemadeaboutthegeneraliza-tionoftheresults.

Finally,althoughourstructuralequationesti-mateshaveconﬁrmedthevalidityofourmodelinwhichtheorganizationaldesigncomponentsaresimultaneouslyincludedasthedependentvari-ables,ourdata(basedoncross-sectionapproach)islimitedinverifyingthecausalorderwithrespecttothesequentialpathlinkingpoliticalorindustrialforcesandeachorganizationaldesigncomponent.Futureresearchmaymorethoroughlyverifythissimultaneityorcausalorderbycollectinganduti-lizingtimeseriesdatathatreﬂectslongitudinalpatternoforganizationaldesigncomponentsaswellaspoliticalandindustrialchanges.Suchreﬁnementscanshedlightontherelativeimpor-tanceofdiﬀerentorganizationaldesigncompo-nentsbywhichtheChinesegovernmentimposespoliticalconstraints,andhowthismaydiﬀeracrossenterpriseswithdiﬀerentenvironmentsandcharacteristics.

Acknowledgements

EarlierversionsofthispaperwerepresentedattheUNSWManagementaccountingsymposium,Sydney,2003,theAFAANZconference,Brisbane,

2003,theAAAAnnualMeeting,Atlanta,2001,theAcademyofInternationalBusiness,Sydney,2001,andinseminarsattheUniversityofMel-bourneandMonashUniversity,July2002.Theauthorsareindebtedtotheparticipants,CheeChow,SueHaka,JoanLuft,GeorgeMilkovich,MichaelMorris,GordanRichardson,andMichaelShieldsfortheirsuggestionsandassistance.Theworkdescribedinthispaperwaspartiallysup-portedbyStrategicResearchGrant#7001016fromtheCityUniversityofHongKong.Thepaperhasbeneﬁtedsigniﬁcantlyfromthecon-structivecommentsandadviceprovidedbytherefereesandtheeditor.

References

Anderson,J.C.,&Gerbing,D.W.(1988).Structuralequationmodellinginpractice:Areviewandrecommendedtwo-stepapproach.PsychologicalBulletin,103(May),411–423.

Bagozzi,R.,&Yi,Y.(1988).Ontheevaluationofstructuralequationmodels.JournaloftheAcademyofMarketingScience,16,74–94.

Bailey,J.,Chen,C.,&Dou,S.(1997).Conceptionsofselfandperformance-relatedfeedbackintheUS,Japan,andChina.JournalofInternationalBusinessStudies,28,605–625.

Baiman,S.,Larcker,D.F.,&Rajan,M.V.(1995).Organi-zationaldesignforbusinessunits.JournalofAccountingResearch(Autumn),205–229.

Baker,G.,Gibbons,R.,&Murphy,K.J.(1994).Subjectiveperformancemeasuresonoptimalincentivecontracts.QuarterlyJournalofEconomics(November),1125–1156.Baker,G.P.,Jensen,M.C.,&Murphy,K.J.(1988).Compensationandincentivespracticevs.theory.JournalofFinance,3,593–616.

Bollen,K.(1989).Structuralequationswithlatentvariables.NewYork:JohnWiley&Sons.

Branine,M.(1996).ObservationsontrainingandmanagementdevelopmentinthePeopleÕsRepublicofChina.PersonnelReview,25(1),25–39.

Brickley,J.A.,Smith,C.W.,&Zimmerman,J.L.(1995).Theeconomicsoforganizationalarchitecture.JournalofAppliedCorporateFinance,8(2),19–31.

Brickley,J.A.,Smith,C.W.,&Zimmerman,J.L.(2001).Managerialeconomicsandorganizationalarchitecture.USA:McGraw-HillIrwin.

Browne,M.W.(1984).Asymptoticdistributionfreemethodsinanalysisofcovariancestructures.BritishJournalofMath-ematicalandStatisticalPsychology,37,62–83.

Bushman,R.,Indjejikian,R.,&Penno,M.(2000).Privatepredecisioninformation,performancemeasurecongruity,andthevalueofdelegation.ContemporaryAccountingResearch,17(4),561–587.

176N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

Byrd,W.,&Tidrick,G.(1991).Themarketmechanismand

economicreformsinChina.Armonk:M.E.Sharpe.

Byrne,B.M.(2001).StructuralequationmodelingwithAMOS:

basicconcepts,applications,andprogramming.Mahwah,NewJersey:LawrenceErlbaumAssociatesInc.

Carmines,E.,&Mclver,J.(1981).Analysingmodelswith

unobservedvariables:analysisofcovariancestructures.InG.Bogrnstedt&E.Borgatta(Eds.),Socialmeasurement:currentissues.Sage:BeverlyHills.

Chen,C.C.(1995).Newtrendsinallocationpreferences:A

Sino-UScomparison.AcademyofManagementJournal,38,408–428.

Chen,H.(2000).TheinstitutionaltransitionofChina’stownship

andvillageenterprises.Aldershot:Ashgate.

Cheng,S.(2001).Studiesoneconomicreformsanddevelopment

inChina.NewYork:Oxford.

Child,J.(1994).ManagementinChinaduringtheageofreform.

Cambridge,England:CambridgeUniversityPress.

ChinaMarketsYearbook.,(2000).Shanghai:AllChinaMar-ketingResearchCo.,Ltd.

ChinaSecuritiesRegulatoryCommission.(2001).Introduction

toChinaSecuritiesMarket.Retrieved7November,2001,fromhttp://www.csrc.gov.cn/CSRCSite/eng/esmintr.htm.Christie,A.A,Joye,M.P.,&Watts,R.L.(2003).Decentral-izationoftheﬁrm:theoryandevidence.JournalofCorporateFinance,9,3–26.

Cronbach,L.J.(1951).Coeﬃcientalphaandinternalstructure

oftests.Psychometrika(September),297–334.

Delmastro,M.(2002).Thedeterminantsofthemanagement

hierarchy:evidencefromItalianplants.InternationalJour-nalofIndustrialOrganization,20(1),119–137.

Eisenhardt,K.M.(1989).Agencytheory:anassessmentand

review.AcademyofManagementReview,14(1),57–74.Firth,M.(1996).Thediﬀusionofmanagerialaccounting

proceduresinthePeopleÕsRepublicofChinaandtheinﬂuenceofforeignpartneredjointventures.Accounting,OrganizationsandSociety,21(7/8),629–654.

Fisher,J.,&Govindarajan,V.(1992).ProﬁtCenterManager

Compensation:Anexaminationofmarket,politicalandhumancapitalfactors.StrategicManagementJournal,13,205–217.

Ford,J.K.,MacCallum,R.C.,&Tait,M.(1986).The

applicationofexploratoryfactoranalysisinappliedpsy-chology:acriticalreviewandanalysis.PersonnelPsychol-ogy,39,291–314.

Gerhart,B.,&Milkovich,G.T.(1990).Organizational

diﬀerencesinmanagerialcompensationandﬁnancialper-formance.AcademyofManagementJournal,33(4),663–691.Gibbs,M.,Merchant,K.A.,VanderStede,W.A.,&Vargus,

M.E.(2004).Determinantsandeﬀectsofsubjectivityinincentives.Workingpaper,UniversityofChicago,Univer-sityofSouthernCaliforniaandUniversityofTexasatDallas.

Golob,T.F.(2001).TravelBehavior.Com:Activityapproaches

tomodelingtheeﬀectsofinformationtechnologyonpersonaltravelbehavior.InD.A.Hensher(Ed.),TravelBehaviourResearch:TheLeadingEdge.Oxford:Pergamon.

Goodall,K.,&Warner,M.(1999).Enterprisereform,labor-managementrelations,andhumanresourcemanagementinmultinationalcontext.EmpiricalevidencefromSino-foreignjointventures.InternationalStudiesofManagementandOrganization,29(3),21–36.

Groves,T.,Hong,Y.,McMillan,J.,&Naughton,B.(1994).

AutonomyandincentivesinChinesestateenterprises.QuarterlyJournalofEconomics,109,181–209.

Hassard,J.,Sheehan,J.,&Morris,J.(1999).Enterprisereform

inpost-DengChina.InternationalStudiesofManagement&Organization,29(3),54–83.

Haveman,H.A.(1992).Betweenarockandahardplace:

organizationalchangeandperformanceunderconditionsoffundamentalenvironmentaltransformation.AdministrativeScienceQuarterly,37(1),48–75.

Hopwood,A.G.(1972).Anempiricalstudyoftheroleof

accountingdatainperformanceevaluation.JournalofAccountingResearch(Suppl.),156–182.

Howell,R.,&Sakurai,M.(1992).Managementaccounting(and

other)LessonsfromtheJapanese.Boston,MA:HarvardBusinessSchoolPress.

Huang,Y.,&Duncan,R.(1997).HowsuccessfulwereChinaÕs

statesectorreforms?JournalofComparativeEconomics,24(1),65–78.

Ittner,C.D.,Larcker,D.F,&Meyer,M.W.(2003).

Subjectivityandtheweightingofperformancemeasure:Evidencefromabalancedscorecard.AccountingReview,78(3),725–758.

Jeﬀerson,G.H.,&Rawski,T.G.(1994).Enterprisereformin

Chineseindustry.JournalofEconomicPerspectives,8,47–70.

Jeﬀerson,G.H.,Rawski,T.G.,&Zheng,Y.(1996).Chinese

industrialproductivity:trends,measurementissuesandrecentdevelopments.JournalofComparativeEconomics,23,146–180.

Jensen,M.C.(1998).Foundationsoforganizationalstrategy.

Cambridge,Massachusetts:HarvardUniversityPress.

Jensen,M.C.,&Meckling,W.H.(1992).Speciﬁcandgeneral

knowledge,andorganizationstructure.InLarsWerin&HansWijkander(Eds.),Contracteconomics.Oxford:BasilBlackwell.

Jensen,M.C.,&Meckling,W.H.(1995).Speciﬁcandgeneral

knowledge,andorganizationalstructure.JournalofAppliedCorporateFinance,8(2),4–18.

Joreskog,D.G.,&Sorbom,D.(1979).Advancesinfactoranal-ysisandstructuralequationmodels.Cambridge,MA:ABT.Keating,A.S.(1997).Determinantsofdivisionalperformance

evaluationpractices.JournalofAccountingandEconomics,24(3),243–274.

Kennedy,P.(1979).Aguidetoeconometrics.Cambridge,MA:

MITPress.

Killing,J.P.(1983).Strategiesforjointventuresuccess.New

York:Praeger.

Lambert,R.A.,&Larcker,D.F.(1987).Ananalysisoftheuse

ofaccountingandmarketmeasuresofperformanceinexecutivecompensationcontracts.JournalofAccountingResearch,25,S85–S125.

N.G.O’Connoretal./Accounting,OrganizationsandSociety31(2006)157–177

177

Lee,C.W.(2001).Financialrestructuringofstateowned

enterprisesinChina:thecaseofShanghaiSunvePharma-ceuticalCorporation.Accounting,OrganizationsandSoci-ety,26,673–689.

Li,D.D.(2000).Insidercontrolvsgovernmentcontrol:astudy

ofChinaÕsstateenterprisesreform.WorkingPaper,HongKong:HongKongUniversityofScienceandTechnology.Li,W.(1997).Theimpactofeconomicreformonthe

performanceofChinesestateenterprises:1980–1989.Jour-nalofPoliticalEconomy,105,1080–1106.

Lin,J.Y.(2000).WTOaccessionandChinaÕsAgriculture.

ChinaEconomicReview,11(4),405–408.

Lin,J.Y.,Cai,F.,&Li,Z.(1998).Competition,policy

burdens,andstate-ownedenterprisereform.AmericanEconomicReview,88(2),422–427.

Luft,J.,&Shields,M.D.(2003).Mappingmanagement

accounting:graphicsandGuidelinesfortheory-consistentempiricalresearch.Accounting,OrganizationsandSociety,28,169–249.

McMillan,J.,&Naughton,B.(1992).Howtoreformaplanned

economy:lessonsfromChina.OxfordReviewofEconomicPolicy,8(1),130–143.

Megginson,W.L.,&Netter,J.M.(2001).Fromstateto

market:asurveyofempiricalstudiesonprivatization.JournalofEconomicLiterature(June),1–89.

Mendelson,H.(2000).Organizationalarchitectureandsuccess

intheinformationtechnologyindustry.ManagementSci-ence,46,513–529.

Merchant,K.A.(1989).Rewardingresults:motivatingproﬁt

centermanagers.Boston:HarvardBusinessSchoolPress.Milgrom,P.,&Roberts,J.(1992).Economics,organization,and

management.EnglewoodCliﬀs:Prentice-Hall.

Milgrom,P.,&Roberts,J.(1995).Complementaritiesandﬁt:

strategy,structure,andorganizationalchangeinmanufac-turing.JournalofAccountingandEconomics,19(2–3),179–209.

Moers,F.(2005).Discretionandbiasinperformanceevalua-tion:theimpactofdiversityandsubjectivity.Accounting,OrganizationsandSociety,30(1),1–98.

Muthe

´n,B.(1984).Ageneralstructuralequationmodelwithdichotomous,orderedcategoricalandcontinuouslatentvariableindicators.Psychometrika,49,115–132.

Nagar,V.(2002).Delegationandincentivecompensation.

AccountingReview,77(2),379–395.

OÕConnor,N.G.,Chow,C.W.,&Wu,A.(2004).Theadoption

of‘‘Western’’managementaccounting/controlsinChinaÕsstate-ownedenterprisesduringeconomictransition.Accounting,OrganizationsandSociety,29(3–4),349–375.Pannier,D.(1996).Corporategovernanceofpublicenterprises

intransitionaleconomies.WorldBankTechnicalPaper323.Peng,M.W.,&Heath,P.S.(1996).Thegrowthoftheﬁrmin

plannedeconomiesintransition:Institutions,organizations,

andstrategicchoice.AcademyofManagementReview,21,492–528.

Perkins,D.(1994).CompletingChinaÕsmovetothemarket.JournalofEconomicPerspective,8(2),23–46.

Prendergast,C.(1999).Theprovisionofincentivesinﬁrms.JournalofEconomicLiterature(March),7–63.

Priem,R.L.,&Rosenstein,J.(2000).Isorganizationtheoryobvioustopractitioners?Atestofoneestablishedtheory.OrganizationScience,11(5),509–524.

Qian,Y.(1996).EnterprisereforminChina:Agencyproblemsandpoliticalcontrol.EconomiesofTransition,4(2),427–447.

Qian,Y.(2001).HowreformworkedinChina.WorkingPaper.DepartmentofEconomics,UniversityofCalifornia,Berke-ley,CA.

Shields,M.,&Young,S.M.(1993).Antecedentsandconse-quencesofparticipativebudgeting:evidenceontheeﬀectsofasymmetricalinformation.JournalofManagementAccount-ingResearch,5,265–280.

Shirley,M.,&Xu,L.C.(2001).Empiricaleﬀectsofperfor-mancecontracts:evidencefromChina.JournalofLaw,Economics,&Organization,17(1),168–200.

Shirley,M.,&Xu,L.C.(1998).Information,incentivesandcommitment:anempiricalanalysisofcontractsbetweengovernmentandstateenterprises.JournalofLaw,Econom-ics,andOrganization,14(2),358–378.

Shleifer,A.,&Vishney,R.W.(1994).Politiciansandﬁrms.QuarterlyJournalofEconomics,109,995–1025.

Simons,R.(1995).Leversofcontrol:Howmanagersuseinnovativecontrolsystemtodrivestrategicrenewal.Boston:HarvardBusinessSchoolPress.

Tosi,H.L.,Katz,J.P.,&Gomez-Mejia,L.R.(1997).Disaggregatingtheagencycontract:theeﬀectsofmonitor-ing,incentivealignment,andterminoﬃceonagentdecisionmaking.AcademyofManagementJournal,40,584–602.Wang,J.(2003).Governanceroleofdiﬀerenttypesofstateshareholders:evidencefromChinaÕslistedcompanies.WorkingPaperpresentedattheAsia-PaciﬁcJournalofAccounting&EconomicsConference,Shanghai,China.Xu,L.C.(2000).Control,incentivesandcompetition.Eco-nomicsofTransition,8(1),151–172.

Xu,L.C.,Zhu,T.,&Lin,Y.M.,(2002).Politicalcontrol,agencyproblemsandownershipreform:evidencefromChina.WorkingPaper.Washington,D.C.WorldBank.Young,A.M.(2000).Theroleofindividualdiﬀerencesinthereferentselectionprocess.JournalofBehavioralandAppliedManagement,1(1),83–91.

Zhongguo,G.B.(1997).China-culture:performancereviewsdiscussed.BeijingReview,1.

Zhuang,J.,&Xu,C.(1996).Proﬁt-sharingandﬁnancialperformanceintheChinesestateenterprises:evidencefrompaneldata.EconomicsofPlanning,29,205–222.

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文