2024年3月7日发(作者：奔腾b70老款)

Available online at ific-BasinFinanceJournal16(2008)591–/locate/pacfinCorporategovernanceandtunneling:EmpiricalevidencefromChina☆LeiGaoa,1,GerhardKlingb,?aNanjingUniversityofInformationScience&Technology,SchoolofFinance,NanjingPukou114,210044,PRChinabBristolBusinessSchool,UniversityoftheWestofEngland,ColdharbourLane,BristolBS161QY,UKReceived3January2006;accepted28September2007Availableonline17November2007AbstractWeanalyzeassetappropriationbyprincipalshareholdersinChinaanduncoverthefollowingrelationships:(1)outsidersintheboardofdirectors,auditwithoutnon-cleanopinion,anddispersedownershippreventoperationaltunneling;(2)belongingutionalownershipdoesnotpreventtheembezzlementofassetsandisendogenous,sgovernancemechanisms,stockcharacteristicsmatterinthatlargerfirmsexhibitlesstunneling,adeclineoftunnelingin2001,whichmightbeduetoeconomicreforms.?ssification:G34;G38Keywords:Corporategovernance;China;uctionInChina,aseriesofscandalsrelatedtotheembezzlemeeincidentof‘Qiongminyuan’in1997tothescandalofthe‘SanjiuGroup’in2005,plentyofanecdotalevidencedemonstratestheextentwitriouscorporategovernanceproblemThispaperwasfinanciallysupportedbytheKeyProjectoftheNationalNaturalScienceFoundationofChina(ProjectNumber:70532001,ChefScientist:WeianLi)andtheProjectoftheNationalSocialScienceFundofChina(ProjectNumber:07CJY001).?.:+441173283418.E-mailaddresses:drgaolei@(),@().1Tel.:+867542902383;fax:+867542903442.0927-538X/$-seefrontmatter?:10.1016/.2007.09.001☆

,/Pacific-BasinFinanceJournal16(2008)591–605couldhindereconomicdevelopment(seeLietal.,2004),whichraisesthequestion:howtopreventtunnelingandprotecttheinterestsofsmallinvestors?Toimprovecorporategovernance,weneedtoidentifywhichcorporategoveescholarshavealreadyconductedvaluableresearchontheextentofoperationaltunneling(seeLietal.,2004;Tangetal.,2004).Yet,theinterrelationofcorporscorporategovernancemeasures,weincorporatefirmcharacteristics(,financialleverage,industryspecificeffects)andaccountforregionaldevelopment(larea,regionalGDPgrowth).OurstudyreferstoapaneldatasetofalllisnpanelOLSfixed-effectsestimationandmoreelaboratetechniques,namelycensored-normalregressions,orderedlogit,andquantileregressions,wecanassesstheimpactofgovernance,ownershipstructures,firmcharacteristics,industryeffects,er,weaddresstheinherentendogeneityissue,forcorporategovernancemeasuresmightaffectfirmbehavior(ing)—ceptofGrangercausalityandpanelvectntyears,theexploitationofminorityshareholdersbylargeshareholdershasattractedscholars\'erandVishny(1986)detectedthatwhenlargestockholderscontrolfirms,themainproblemisnolongertheconflictofinterestsbetweenmanagementandshareholders,butprnetal.(2000b)coinedtheterm‘tunneling’todescribeassetappropriationbylargeshareholders,ingnotonlyhurtstheinterestsofsmallshareholders,butalsoseriouslyhindersstockmarkets\'development(seeJohnsonetal.,2000b;Wurgler,2000;Bertrandetal.,2002).Johnsonetal.(2000a)arguedthatunrestrainedtufinancialcrises,manyfactsprovedthatncethedevelopmentofemergingfinancialmarkets,onehastoidenertriestodetectmechanismsandprovidespolicyrecommendationstostrengthencorporategovernancenotonlyinChina—erisorganizedasfollows:theliteraturerengonoperationaltunneling,weanalyzeinterrelationsbetweencorpornpanelOLS,censored-normalregressions,orderedlogit,andquantileregressions,inginChinaTunnelingisalsothefocusofcurrentChinesecorporategovernanceresearch,whichprovidesevidencethatprincipalshareholderstunnelassets;however,thesestudiesonlydescribetheproblemwithoutuncoveringtheunderlyingcau.(2004)collectedevidenceontunnelingofbigshareholdersfocusingontheembezzlementoffundsandassettransfersrelatedtomergersandacquisition;thus,operatundthatconcentrate

,/Pacific-BasinFinanceJournal16(2008)591–605593particular,companiescontrolledbythegodWong(2003)reportedthatrelatedlendingtransactions,commonlyreferredtoasfinancialtunneling,stheirscope,thesestudiessufferfrominherentendogeneityissues,icular,financialtunnelingreferstofreeze-outsanddilution—bot-outs(seeGilsonandGordon,2003;Batesetal.,2006;Subramanian,2004)describeasituationofdelistingduetoatenderofferthatisrelativelylowcomparedtotheactualmarketvalueofthecompany,a,aneffectivedelisting,freeze-outsarenotessentialinChina—butso-calleduncompensatedtransactionsinthecaseofmergersandacquisitionsareimportant(seeGaoandKling,inpress).However,uncompensatedtransactionsaremainlycommonwhenthestaterestructuresstate-ownedenterprises(SOEs)uently,on(seeBlackandKraakman,1996),whichdescribesaconsiderableincreaseinequitycapitalislesslikelyinChina,asthestatetriestokeepcontrolofpubliclylistedcompaniesandissuingnewsrast,operationaltunnelingthroughtransferpricing(seeJohnsonetal.,2000b)andotherunjustifiedtransfers,dpartytransactionsofintangibles(seeCheungetal.,2006)seemtobemorerelevantinChina(seeLietal.,2004;Tangetal.,2004).Accordingly,ourpaperfocusesonoperationaltunnelingandtriestomeasuretheextentoticalhypothesesTheresearchfocusofcorporategovernancehasshiftedfromtheproblembetweenshare-holdersandmanagementtothatbetweenmajorshareholdersandminorityshareholders(seeShleiferandVishny,1997;Johnsonetal.,2000b;DenisandMcConnell,2003).Tosolvetheproblemthatblock-holdingshareholdershurttheinterestsofsmallshareholders,oneshouldimprovethefollowingtwokindsofcorporategovernancemechanisms(seeDenisandMcConnell,2003):(1)internalmechanismsincludethestructureoftheboardofdirectors,incentivesystemsofseniormanagement,ownershipstructure,typeofblock-holdingshare-holders,institutionalinvestors\'ownership,andcorporatetransparency.(2)Externalmechanismsincludemarketcompetitionforcorporatecontrol,legalenvironment,protectionofminorityshareholders,marketdevelopment,ssthestructureoftheboardofdirectors,weusethepercentageofoutsiders,alboardmembersarefamiliarwiththefirm\'soperations(seeYermack,1996).Thir,thehigherthepercentageofinternalboardmembers,theeasierthefirmcanbecontrolledbythemanagement,ercentageofoutsidersintheboardoeoftheboardofdirectormatters,asJensenandMurphy(1990)arguedthatreasonablyincreasingthenumberofboardmemberscouldimprovetheboard\'uently,itismoredifficultfortheCEOorsenior

,/Pacific-BasinFinanceJournal16(2008)591–,theboardcanbetterfunctiontocor,whentheboardistoobig,communicationwouldbecomedifficultandtheefficiencyoftheboarddecreases(seeYermack,1996).sonablerange,quencyofboardmeetingsindicatestheimportanceoftheboardinfirm\',moremeetingswouldindica,from1998to2002,rmsconveneasfewasonceperyear,whilesomefirmsholdasmanyas37meetings.2Nevertheless,onecouldarguethatdistressedcompaniessufferingfromtunnelingholdmoremeetingsduetotheirstateofemergency;hence,berofboardmeetingsindicatestheimportanceinthedecisionprocess,gstocksmakessenioasmanagersareminorityshareholders,whotheyusuallyare,gersownstocks,trationofownershipcouldfacilitateassetappropriation(seeGomesandNovaes,2001),asmajorshareholdersnotonlydominateshareholdergatheringsandtheboardofdirectors,butalsodeterminefirm\'sdailyoperationbyappointingtheir‘owncandidate\'Shitu(2004)foundthatin2001,block-holdingshareholdersselected82.9%ofallCEOs.3Furthermore,theyfoundthat52.4%ofboarddirectorsorCEOshadpositionsatblock-holdingshareholders\'hisdependency,swithseveralmajorshareholders,majorshareholdershavetonegotiate;thus,somedecisionsthatdamagesmallshareholderscouldpossiblybeavoided(seeZwiebel,1995).BlochandHege(2001)arguedthatifthereareseveralshareholdersstrugglingforcontrol,ithasingleblock-holdingshareholderexhibitmoincipalshareholdersandlistedfirmscanconductrelatedpartytransactions,tunnelingismorelikely(seeKhannaandPalepu,2000a,b;JianandWong,2003).Conductingtransactionsiseasierwhentheblock-holdingshareholderisorganizedaicular,businessgroups,inwhichprincipalshareholderscontrolYetabnormalconditionsinthefirm\'8.78%sly,nalysis,weaccountedforabnormallyhighnumberofmeetings,butwedidnotfindanysignificantimpactonassetappropriation.3Actually,thestatedirectlyorindirectlyappoints69%ofalldirectorsandCEOsbasedonfiguresfor2001;hence,becomingaCEOismainlypoliticallydetermined(seeQuiang,2003).4HighlydispersedownershipstructureisratherrareinChina—butitcanserveasreferencegroup.2

,/Pacific-BasinFinanceJournal16(2008)591–605595themembercompaniesbyapyramidownershipstructure,ernalcapitalandfactormarketsofbusinessgroupsbecomethebestvenueforbigshareholderstotunnel(seeKhannaandPalepu,2000a).Therearestrongincentivesandmultiplemeasurestoappropriatemembercompanies\'resources;therefore,tunnelinginbusinessgroupsismoresevere(seeJianandWong,2003).ithblock-holdaseofprivatizations,well-performingunitsofcompaniesareseparatedandlisted,uently,whentheparentcompanyencountersdifficulties,itcol.(2004)arguediththeslandPoulsen(1987),ShleiferandVishny(1986),andMcConnellandServaes(1990)andBrickley,LeaseandSmith(1988)uncoveredthatinstitutionalinvestorstendtoopposefirm\'rasttofindingsforothercountries,Tang,Luo,andWang(2004)foundthatChineselistedcompanieswhosesecondlargestshareholderisinstitutionalinvestorbadlysufferfromtunneling.5YetXiaoandWang(2004)showedthatChineseinst,onecanarguethatinstitutionalinvestorsselectcompanieswithgoodcorporategovernance,whichindicatesanallegedendogeneitybias(seesectionfive).ercentageofinstitutionalinvestorsenhancesgoodcorporategovernancepractices;thus,reportsarethemostimportantfinancialinformationdisclosedbylistedfirms,andirmscanissueanon-cleanopinion,orth,auditswithoutnon-cleanopinioninwithoutnon-cleally,largerauditfirmsandfirmswithinternaingtosignalinggames,whenlistedfirmshireauditfirmsthatchargehighfeesandofferhigh-qualityauditopinions,thsly,hiringoneofthetopfiveauditfikthatthereasonforthisfindingisthattheyincludedprivatefundsasinstitutionalinvestors;however,privatefundsinChinacannotberegardedasinstitutionalinvestors.5

,/Pacific-BasinFinanceJournal16(2008)591–eadingfiveinternationalauditfirmsproveannualreports,alsyaetal.(1998)uncoveredthatincommonlawcountries,thelevelofcorporategovernanceishigh,rast,incivillawcountries,a,thestockmarketissegmentedintoAandB-shares,andsomecompaniescanissuestockslistedonforeignexchanges(suchasH-shareinHongKong,andADRintheUSA).ssuingB-ore,orporategovernanceevaluationsystemofStandard&Poor\'s,thecountry\'smarketdevelopmentisanimportantindicatorregardingcountry\'egionaldisparitiesconcerningeconomicdevelopmentinChina,theeastuldimplythatcorporategovernansusingthissimpleregionaldivision,weincorporateGDedtofirmsinwesternandcentralChina,firmslocatedintheeasterncoastalregionexhibitbettercorporategovernancestructures,gerswasteresources,competitionintheprodua,someindustriesareprotected,andfirmsoperatingintheseindustriesfacelesscompetition;thus,rstudyusesdataprovidedbytheChinaCenterforEconomicResearch(CCER)databaseonoperationaltunneling,companycharacteristics(salesandfinancialleverage),andcorporategovernancemechanismsfrewlylistedcompanies,yzetheextentoftunneling,wedeterminethedifferencebetweenaccstedcompanieshavetransactionswithoneoftheirblock-holdingshareholders,weregardthesetransactionsasrelatedpartytransactions,ral,relatedpartytransactionsarethegatewayforoperationaltunneling(seeJohnsonetal.,2000b).Albeitnotallrelatedpartytransactionsleadtoassetappropriation,therisknfivediscussesthemeasurementproblemsandrelatedissuesindetail.6WefollowthedefinitionofindustriesprotectedbythestateinChenetal.(2005).

,/Pacific-BasinFinanceJournal16(2008)591–605Table1BasicdescriptivestatisticsYearsTunnelingCasesoftunneling/observationsSingleMultiBoardsizeMeetingsOutsiderSharesAuditBigfiveHBsharesStateGroupFundshareProtectedLn(Sales)LeverageObservations19980.0570.6720.6030.1119.7574.2120.0030.0010.8530.0300.1170.8600.7960.0040.1098.5950.4.0560.6980.5990.1289.6344.8630.0050.0010.8260.0310.1110.8680.8130.0090.1058.6410.42278220000.0580.6760.5780.1399.4815.3930.0110.0010.8520.0360.0970.8500.8210.0070.1128.6870.43293620010.0380.4920.5670.1399.4506.2560.0610.0010.8860.0570.0990.8430.8140.0050.1118.7180.455103820020.0400.4500.5600.1489.9228.4980.2370.0030.9040.0900.0970.8140.7990.0070.1148.7590.4981108All5970.0490.5730.5780.1359.6506.0730.0750.0010.8680.0520.1030.8450.8090.0060.1118.6890.4484559Erminetheconcentrationofownership,namelysingleorseveralmajorshareholders,wefollowthedefinitionsbyZhaoandYang(2003).Thepercentageofstocksheldbyseniormanatageofstocksheldbyinstitutionalinvestorsinaproxyfortheprestigeofauditfirms,weuseadummythatfiveconsistsofAnderson,KPMGLLP,Ernst&Young,De-loitte&Touche,2002,Andersonwithdrewfromtheauditbusiness;hereafter,iningtheextentofstatecontrolisdifficult,asthestatecontrolsonly8.5%r,dueto‘pyramidshareholdingschemes’,thestatecanultimatelycontrol84%ofequityin2001(seeQuiang,2003).Toidentifythetypeofblock-holdingshareholders,wefollowthedefinitionofTangetal.(2004)inthatnon-corporationgroupsrefertomanagementagenciesofstateassets,researchinstitutions,institutionsofhighereducation,socialgroups,banks,insurancecompanies,owthedefinitionofindustriesprotectedbythestateinChen,Chen,andWan(2005).Thisincludesthepetroleumandchemicalindustry,untforregionaldisparitiesregardingmarketdevelopment,g,Tianjin,Shanghai,Jiangsu,Zhejiang,Fujian,ShandongandGuangdongbelongtothedevelopedeasterncoastalregion,sthisstraightforwardgeographicaldivision,stratetheextentoftunnelingandobtainarelativemeasure,wedividedt5.68%oftotalassetswerelostinrelatedpartytransactionsin1998comparedto4.04%4yearslater;hence,sthedeclineintheextentoftunneling,thefrequencyofassetappropriationdeclinedfrom67.19%to45.04%.Inspiteofthisconsiderableimprovement,2summapinassetappropriationin2001and2002isaccompaniedbyanincreaseinthenumberofboard

,/Pacific-BasinFinanceJournal16(2008)591–605Table2DefinitionandcomputationofvariablesNameDefinitionandcomputationTypeContinuousDummyTunnelingProxy:tspayable/receivablerefertorelatedpartytransactions,whicharedisclosedinannualreportsSingleDummyisequaltoone,ifoneshareholdercontrolsmorethan50%wnershipisbetween40%and50%andhigherthatthesumofthesharesheldbythesecondtofifthlargestshareholder,weregardthelargestshareholderstillassinglemajorshareholderMultiDummyisequaltooneiflargestshareholderholdsbetween10%and50%,secondlargestatleast10%,andthepercentageofthelargestissmallerthanthesummedpercentageofthesecondtothefifthlargestshareholderBoardNumberofboardmemberssizeMeetingsNumberofboardmeetingsperyearOutsiderPercentageofoutsidersintheboardofdirectorsDummyContinuousContinuousContinuousbetween[0,1]SharesPercentageofstocksheldbyseniormanagers(boardmembersandtopmanagement)Continuousbetween[0,1]AuditDummyvariablethattakesthevalueoneifanauditwithoutnon-cleanopinionisissuedDummyBigfiveDummythatisequaltooneifthefiveleadingauditfirmsareinchargeDummyHBDummythatisequaltooneifthefirmissuesBorHsharesorothersharesnottradedontheDummysharesShanghaiandShenzhenStockExchangeSlisbasedonDummythepyramidownershipstructureGroupDummythatisequaltooneifblock-holdingshareholdersarebusinessgroupsDummyFundPercentageofstocksheldbyinstitutionalinvestorsContinuoussharebetween[0,1]ProtectedDummythatisequaltooneifthefirmoperatesinaprotectedindustryDummyCoastDummythatisequaltooneifthefirmislocatedintheeasterncoastalregionDummyLn(Sales)Thenaturallogoffirm\'snetsalesisaproxyforfirmsizeContinuousLeverageToassessthecapitalstructure,wedividelong-termdebtsbytotalassetsContinuousmeetings,thepercentageofoutsidersasboardmembers,thepercentageofstockheldbymanagersandinstitutionalinvestors,hangesincnourtheoreticalhypotheses,ing?ai?b1?Singleit?b2?Multiit?b3?BoardSizeit?b4?Meetingsit?b5?Outsiderit?b6?Sharesit?b7?Auditit?b8?BigFiveit?b9?HBShareit?b10?Stateit?b11?Groupit?b12?FundShareit?b13?Protectedit?b14?Coastit?b15?GDPit?b16?Growthit?b17?ln?Sales?it?b18?Leverageit?eit?1?Beforecarryingoutanyregressions,relationcoefficientsbetweenexplanatoryvariablesarerathersmall;hence,multicollinearity

,/Pacific-BasinFinanceJournal16(2008)591–605599doesnotaffectourresults.7Aswedealwithpaneldata,weembedfirmspecificeffectsintheconstantterm(fixed-effectsmodel).Toimproveourmeasureofregionaldevelopment,weusedthelevelofGDPpercapita,andGDPgroinguishbetweenhighlydevelopedandlessdevelopedprovinces,wecalculatedthedeviationofGDPpercapitaandGDPgrowthratefromthenationaltrend.8However,theresultsinTable3donotindicatethatregionaleffectsmatter,tion,neitherGDPpercapitanorGinreferenceestimatesofEq.(1),weuseapanelOLSestimationwithoutconsideringtimeorfirmspecificeffects(seemodels1and2).Model3incorporatestimeeffectsusingafixed-effectapproach,whereawithoutnon-cleanopinionhaveinallmoalvariables,namelythedummyforcoastalprovinces,relativeGDPpercapita,andrelapecifications,firmsizeasmeasuredbysalesreducedassetappropriation,hipstructuremattersinthatahigherextentofconcentration(singlemajorshareholder)eoutsidersarereprestingly,controllingforfirmspecificeffects(seemodel4)showsthatinstitutionalownershipandbeinginaprotectedindustryarenotrelevantdriversforassetappropriationassuggestedbymodels1–r,therearetwomainshortcomingsofourOLSestimatesthatshouldbeclarifiedbeforeclaimingthatourresultsarerobust:(1)ourproxyfortunnelingiscensored,asonlyabout50%untforatruncateddistributionofourdependentvariable,weuseacensored-normalregressionmodel(seeTobin,1958).(2)Besidesobservingmanycaseswithouttunneling,someobservationsareextremevalues,forinstancesomecompanieslost48.04%oftotalassetsinrelatedpartytransactions.9Theseextremecasesoftunnelingdeviatestronglyfromthemedianofourmeasure,whichreachesonly0.44%.Toexplainsevereformsofassetappropriation,weapplyquantileregressionsforthe90and95percentile(seeKoenkerandHallock,2001).(3)Analternativeapproachisconstructingranksfortunneling,,thecontinuousmeasureyzearankeddependentvariable,weapplyanorderedlogitapproach(seeZavoinaandMcElvey,1975).Table4providestheresultofthecensored-normalregression,quantileregressions,ficientsofthepanelOLSandthemoreadvancedmethodsarenoteasilycomparable,weillustratetheaveragecontributionofeverycorporategovernancevariableontheextentoftunnelinginTable5.10Thestepsareasfollows:(1)wedeterminethepredictedvaluesoftunnelingbasedontherespectiveestimationtechnique.11(2)Then,wecomputethepartialThelargestcorrelationcoefficientof–0.4625canbedetectedbetweenthedummiathighlydispersedownershipstructureisthereferencegroupforourdiscussionconcerningconcentrationofownership.8ThisapproachalsoavoidtrendcorrelationandstationarityproblemsofGDPtimeseries,onalandnationalGDPtimeseriesexhibitacomovement,calculatingarelativemeasureyieldsastationarytimeseriesthatindicatesrelativedevelopmentofprovinces.9Thisvalueisbasedonthe99%percentile.10Weuseaveragevaluesofexplanatorvethepartialimpact,wecomputemarginaleffects(elasticity).11Notethatallmodelssuggestalowerleveloftunnelingifacompanyreachesthemeanofallexplanatoryvariables.7

,/Pacific-BasinFinanceJournal16(2008)591–605Table3PanelOLSwithandwithoutfixed-effectsYearsModel(1)PanelOLSSingleMultiBoardsizeMeetingsOutsiderSharesAuditBigfiveHBsharesStateGroupFundshareProtectedCoastGDPgrowthGDPpercapitaLn(Sales)LeverageYear1999Year2000Year2001Year2002ConstantObservationsAdjustedR20.008???0.0070.0000.000?0.059????0.053?0.052????0.0110.017???0.007?0.014????0.363???0.018???0.000?0.025???0.051???Model(2)PanelOLS0.008???0.0070.0000.001?0.059????0.055?0.052????0.0100.017???0.007?0.014????0.365???0.017???0.0000.000?0.024???0.051???Model(3)Yeareffects0.007???0.007?0.0000.001?0.064????0.055?0.051????0.0100.017???0.007?0.014????0.386???0.018???0.000?0.024???0.051????0.0010.004?0.013??0.0010.264???45660.11Model(4)Firmeffects0.002?0.016???0.0010.001?0.032???0.094?0.014???0.0050.0260.0140.012??0.053?0.011?0.007?0.044???0.046???0.264???45660.110.234???45660.1145660.60?Significanceatthe0.10level.??Significanceatthe0.05level.???butionofthodhelpstoidentifynotjuststatisticallybutalso‘economically`significantexternalandinternalgovernancemechanisms.12Inallmodels,auditswithoutnon-cleanopinionhaveanegativeandstatisticallysignificantimpactontunneling;however,theeconomicimpactisratherlimited,asitreducestunnelingonlyby2.83%to14.69%.Atafirstglance,themosteffectivegovernancetoolisthepercentageofinstitutionalinvestors,whichcanreduceassetappropriationbyatleast43.91%(lowestestimate).Especially,forseverecasesofembezzlementexplainedbythequantileregressions,institutionalownershipseemstoprevent(reductionof85.37%and87.19%)eoftheallegedimpactofinstitutionalownership,erkey-elementforimprovingcorporategovernanceissharesheldbymanagers,whichleadstoadeclineinpredictedassetappropriationbyatleast45.21%.However,managers\'ownershipislessrelevantforextremecases,asthepartialcontributionreachesonly6.62%to6.85%yspecificfactors,namelyfirmsize(measuredbylognetsales)andfinancialleverage,areinallmodelshighlysignificant—icular,largercompaniesBy‘economically’significant,wemeanthatthemagnitudeofimpactoftherespectiveexplanatoryvariableontunnelingishigh.12

,/Pacific-BasinFinanceJournal16(2008)591–605Table4Censored-normalregression,orderedlogit,andquantileregressionsYearsSingleMultiBoardsizeMeetingsOutsiderSharesAuditBigfiveHBsharesStateGroupFundshareProtectedCoastLn(Sales)LeverageYear1999Year2000Year2001Year2002ConstantObservationsCensoredregression?0.000?0.025????0.0000.002???0.153????0.895???0.066????0.027??0.031???0.011?0.025????0.823???0.024???0.001?0.030???0.064???0.0030.004?0.037????0.0140.274???4559Orderedlogit(10percentile)?0.070?0.379???0.0030.021???1.855????10.295???0.717????0.357??0.381???0.146?0.306????10.377???0.278????0.049?0.247???0.441????0.039?0.075?0.671????0.446???4559Orderedlogit(20percentile)?0.074?0.391????0.0020.022???1.811????10.622???0.640????0.366???0.395???0.1270.283????9.944???0.263????0.059?0.208???0.381????0.017?0.079?0.674????0.450???4559Quantileregression(95%)0.011?0.035??0.0020.002?0.094?0.098?0.178????0.044?0.0310.0060.056????1.294???0.052???0.003?0.073???0.174???0.0190.001?0.020?0.0090.898???4559601Quantileregression(90%)0.016??0.021??0.0000.000?0.049?0.065?0.139????0.0290.024??0.0020.035????0.809??0.040????0.001?0.046???0.099????0.014?0.015?0.052????0.035?0.616???4559?Significanceatthe0.10level.??Significanceatthe0.05level.???sufferlessfromassetappropriation,stheimpactofgovernancemechanismsandfirmcharacteristics,weconfirmageneraldeclineoftunnelingin2001indicatedbysignificantdummyvariables(seeTable4).Thismightbeduetoeconomicreformsin2001,forChinahastriedtosellstate-ownedstocksandtoimprovecorporategovernance(seeQuiang,2003).neityissuesandrobustnesschecksTocheckwhetherourresultsarerobust,icular,theberofboardmeetingsmightincreaseduetoahighnumberofrelatedpartytransactionsthatneedapproval;hence,heless,ahighnumberofboardmeetingscanbeinterpretedasasignargumentsarevalidfromatheoreticalpointofview,wehavetotesttheendogenousrelationshipbetweencotion,institutionalinvestorsmaychoosetoinvestinfirmswithgoodcorporategovernanceasshownbyXiaoandWang(2004);therefore,thereductionoftunnelingmaybeduetothegoodcorporategovernanceandnotdrivenbyinstitutionalinvestors\'rmore,therecouldbeaself-selectionbiasinthecaseofthevariable‘bigfiveauditors’,forcompaniesmighthiretheseleadingauessthisallegedself-selectionbiasandendogeneityissues,weapplyatwo-stageleastsquaresmodel.

,/Pacific-BasinFinanceJournal16(2008)591–605Table5IllustrationofresultsforarepresentativefirmYearsSingleMultiBoardsizeMeetingsOutsiderSharesAuditBigfiveHBsharesStateGroupFundshareProtectedCoastLn(Sales)LeverageCensoredregression?0.02?1.35?0.010.09?8.22?48.04?3.56?1.431.650.611.32?44.181.260.03?1.593.43Orderedlogit(10percentile)?0.31?1.670.010.09?8.15?45.21?3.15?1.571.670.641.34?45.571.22?0.22?1.081.94Orderedlogit(20percentile)?0.33?1.73?0.010.10?8.00?46.91?2.83?1.621.750.561.25?43.911.16?0.26?0.921.68Quantileregression(95%)0.74?2.35?0.140.10?6.32?6.62?11.99?2.982.100.403.79?87.193.480.18?4.9311.73Quantileregression(90%)1.67?2.26?0.020.00?5.17?6.85?14.69?3.042.53?0.203.74?85.374.18?0.11?4.8510.43POLS1.41?9.51?0.610.49?19.82?57.67?8.473.0115.718.536.69?30.74?6.75?3.87?26.8128.40Thistableshowstheontributionsareexpressedinpercentagepoints,whichprelspredictlowervaluesoftunnelingfortheaveragefirm;hence,-stageleastsquaresapproach,however,requiresfindi,weusedallpotentialsetsofinstrumentsbasedonourexplanatoryvariables(seeEq.(1)).Regardlesswhichspecificationisused,Wu-HausmanF-testsandDurbin-Wu-HausmanChi-squaredtestscannotrejectthenullhypothesisthatthevariablesboardmeetings,institutionalownershipand‘bigfiveauditors’dinherentspecificationproblemsofatwo-stageleastsquaresapproach,orth,w6showstheresultsandhighlightsthattheextentofassetappropriationisnotGrangercausedbythenumberofboardmeetings,institutionalownershiporchoosingoneofthe“bigfive”rast,soundcorporategovernancewithlowlevesendogeneityproblems,themeasurementofassetappropriationisnotstraightforward,fordirectmeasurescannotbeobservedfrompublicsources(reports).BasedonJohnsonetal.(2000b),esinannualreportsindicaterelatedpartytransactions;hence,wefocusontheaccountspayableandreceivableduetotransactionsbetweentherespectivinarelativemeasurethataccountsforfirmsize,wedividethedifrvesasaproxyforassetappropriation,asrelatedpartytransactionwithahighimbalancebetweenaccountspayableandreceivateroffact,,themeasureisnoisyinthatassetappropriationcannotbedirectlyobserved,andrelateorth,itisdifficulttodistinguishbetweennormalrel,theaccrualsmaybearesultofpriorearningmanagementanddonotindicate

,/Pacific-BasinFinanceJournal16(2008)591–605Table6GrangercausalitytestsNullhypothesesBoardmeetings,institutionalownershipandthe“bigfive”auditfirmsdonotGrangercauseassetappropriationAssetappropriationdoesnotGrangercausesthenumberofboardmeetingsAssetappropriationdoesnotGrangercauseschoosingoneofthe“bigfive”auditfirmsAssetap,anincreaseinaccountsreceivablecanalsoimplythatthelargeshareholderisembezzlingassetsfromthelistedcompany,forthelistedcsaccountingmeasures,empiricalmeasuresbasedonfinancialandstockmarketperformancehavebeenused(Baeetal.,2002;Cheungetal.,2006).However,themainproblemwiththesemeasuresisthattheyrequireanefficientstockmarketwithhighliquiditysottcrucialshortcomingisthatbytestingtheimpactoftunnelingonstockperformanceadditionalfactors(ecificeffects,institutionalownership)areusuallynotconsidered,ingly,wesuggestusinganaccountingbasedmeasure;however,\'sQdefinedasthemarketvalueoffirmassetsdividedbytherepurchasevalueoffirmassetsindicateswhetheracompany\'gthisdifferently,acompanywithaTobin\'sQexceedingonecreatesvaluebycombiningitsresources,ingly,companieswithahighTobin\'sQseemtohavede,assetappropriationshouldreduceTobin\'sQinfuture,asessentialresourcesdisappear,roxyforassetappropriationisanadequateindicatorforassetappropriation,weshouldobservethefollowingrelationships:(1)companieswithhighTobin\'sQshouldbeapreferredtargetforassetappropriation;thus,theaccountingproxyshouldhavehighvalues;(2)whenhighlyvaluedassetsdisappearindicatedbyahighvalueofourproxy,companiesshouldexhibitalowerTobin\'-effectsmodelsshowthatahighTobin\'sQcausesmoreassetappropriation(p-value:0.000),andassetappropriationreducesTobin\'sQinthefollowingyear(p-value:0.025).Consequently,sionOurstudyuncoveredahighextentofoperationaltunnelinginChina,asabout4%situationhasimprovedbecausetheextentoftunnelingdeclinedfrom1998to2002from5.68%to4.04%,andonly45.04%ofthefirmssufferfromassetappropriationin2002comparedto67.19%iricalanalysisidentifiedfourkey-mechanismsofcorporategovernancethatcouldhelptosignificantlyimproveiwithoutnon-cleanopinionsignalfinancialhealthandsoundgovernanceloweringtheextentofassetappropriationby2.83%to14.69%.Consequently,minorityshareholdersshouldbealarmedwhenauditfirmsissuenon-cleanopinions,tion,sorth,enhancing

,/Pacific-BasinFinanceJournal16(2008)591–605stockownership(koptionplans)rast,institutionalownershipalonedoesnorcausalitytestsrevealtha,theargumentthatinstitutionalownershipshouldbheless,wedisagreewithTangetal.(2004),rthy,thestateorSOEsasprincipalshareholdersdonotsignificantlyaffectassetappropriation;hence,tunnelinheless,severalblock-holdingshareholdersthatcompeteforcontrollingacompanymakeembezzlementlesslikely—ingly,ourresultsdonotconfirmthepessimisticviewofChenetal.(2005)verthatintheyear2001,inwhicheconomicreformsoccurred(seeQuiang,2003),,economicreformsliketheattempttoimprovecorporategovernanceandtolimittheinfluenceofthestateinpubliclylistedcompaniesarerecommendedancesBae,K.H.,Kang,J.K.,Kim,J.M.,ingorvalueaddition?lofFinance57,2695–,C.E.,Liu,Q.,Lu,J.,Song,F.M.,Zhang,J.X.,lofComparativeEconomics32,599–,T.,Lemmon,M.,Linck,J.,otiationandshareholderwelfareinminorityfreeze-outdeals:areminorityshareholdersleftoutinthecold?JournalofFinancialEconomics81,681–nd,M.,Mehta,P.,Mullainathan,S.,ingouttunneling:rterlyJournalofEconomics117,121–,B.,Kraakman,R.,dLawReview109,1911–,F.,Hege,U.,ey,J.A.,Lease,R.C.,Smith,C.W.,lofFinancialEconomics20,267–,D.H.,Chen,X.Y.,Wan,H.L.,icResearch39/2,92–,Y.L.,Rau,R.,Stouraitis,A.,ing,proppingandexpropriation:lofFinancialEconomics82,343–,D.K.,McConnell,J.J.,lofFinancialandQuantitativeAnalysis38,1–,L.,Kling,G.,transfersandmarketreactions:,R.,Gordon,J.,,A.,Novaes,W.,l,G.A.,Poulsen,A.B.,epellentsandstockprices—lofFinancialEconomics19,127–,M.,Murphy,K.,lofPoliticalEconomy98,225–,M.,Wong,T.J.,ingandearningsmanagementthroughrelatedpartytransactions:n,S.,Boone,P.,Breach,A.,Friedman,E.,ategovernanceintheAsianfinancialcrisis,1997–lofFinancialEconomics58,141–n,S.,LaPorta,R.,Lopez-De-Silanes,F.,Shleifer,A.,anEconomicReview90,22–,T.,Palepu,K.,paffiliationprofitableinemergingmarkets:lofFinance55,867–,T.,Palepu,K.,ureofbusinessgroupsinemergingmarkets:yofManagementJournal43,268–285.

,/Pacific-BasinFinanceJournal16(2008)591–605605Koenker,R.,Hallock,K.,lofEconomicPerspectives15,143–a,R.,Lopez-de-Silanes,F.,Shleifer,A.,Vishny,R.W.,lofPoliticalEconomy106,1112–,Z.Q.,Wang,Z.W.,Sun,Z.,ingandownershiparrangement:lofAccountingResearch25/12,3–ell,J.,Servaes,H.,lofFinancialEconomics27,595–,Q.,lofAsianEconomics14,771–,D.H.,Shitu,D.N.,calresearchofChineselistedfirm\'conomics27/5,69–er,A.,Vishny,R.W.,lofPoliticalEconomy94,461–er,A.,Vishny,R.W.,lofFinance52,737–anian,G.,-siliconixfreeze-outs:theory,evidence,dLawSchoolDiscussionPaper,,Q.,Luo,D.,Wang,L.,tockholder\'stunneldigginganditbalancingstrength:,J.,etrica26,24–r,J.,lofFinancialEconomics58,187–,X.,Wang,K.,tiesinvestmentfund:investororspeculator?Pk,D.,lofFinancialEconomics40,185–a,R.,McElvey,W.,lofMathematicalSociology4,103–,J.W.,Yang,H.T.,uenceofmultiplemajorshareholders:l,J.,ofEconomicStudies62,161–185.

更多推荐

老款,奔腾,作者