law，finance and economic growth in china

《law，finance and economic growth in china》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

Law，financeandeconomicgrowthinchinaJournalofFinancialEconomics77(2005)57–116Law,??nance,andeconomicgrowthinChina$FranklinAllena,??,JunQianb,MeijunQianbaTheWhartonSchool,UniversityofPennsylvania,FinanceDepartment,Suite2300SH-DH,Philadelphia,PA19104,USAbFinanceDepartment,CarrollSchoolofManagement,BostonCollege,ChestnutHill,MA02467,USAReceived6January2003;receivedinrevisedform18May2004;accepted24June2004Availableonline1April2005AbstractARTICLEINPRESS .elsevier4>>/locate/econbase‘‘TheFutureofChineseManagement.’’FinancialsupportfromBostonCollegeandtheWhartonFinancialInstitutionsCenterisgratefullyacknowledged.Theauthorsareresponsibleforallremainingerrors.??0304-405X/$-seefrontmatterr2005ElsevierB.V.Allrightsreserved.doi:10.1016/j.j??neco.2004.06.010Correspondingauthor.Tel.:+12158983629;fax:+12155732207.E-mailaddress:allenf@wharton.upenn.edu(F.Allen).$Weappreciatedetailedcommentsfromananonymousrefereethatsigni??cantlyimprovedthepaper,andassistancefromQiaoYuandWuxiangZhuinconductingthe??rmsurvey.WealsoreceivedhelpfulcommentsfromErikBerglo¨f,DanBergstresser,LorenBrandt,Dong Chen,GregoryChow,JulanDu,JieGan,LiJin,SungWookJoh,SimonJohnson,RossLevine,LijunLin,AnthonyNeoh,TomRawski,YihongXia,andseminar/sessionparticipantsattheAmericanEconomicAssociationmeetings,theAsiaCorporateGovernanceConferenceinKorea,Dartmouth,FudanUniversity,Harvard,HEC-Montreal,InternationalMonetaryFund/WorldBank,LondonSchoolofEconomics,FederalReserveBankofNewYork,MassachusettsInstituteofTechnology,Rice,theRiksbank,theUniversityofHouston,UniversityofTexas-Austin,TsinghuaInternationalFinanceConferenceinBeijing,andWhartonConferenceonChinaisanimportantcounterexampletothe??ndingsinthelaw,institutions,??nance,andgrowthliterature:Neitheritslegalnor??nancialsystemiswelldeveloped,yetithasoneofthe fastestgrowingeconomies.Whilethelaw–??nance–growthnexusappliestotheStateSectorandtheListedSector,witharguablypoorerapplicablelegaland??nancialmechanisms,thePrivateSectorgrowsmuchfasterthantheothersandprovidesmostoftheeconomy’sgrowth.Theimbalanceamongthethreesectorssuggeststhatalternative??nancingchannelsandgovernancemechanisms,suchasthosebasedonreputationandrelationships,supportthegrowthofthePrivateSector.r2005ElsevierB.V.Allrightsreserved.JELclassification:O5;K0;G0Keywords:Lawand??nance;Economicgrowth;Privatesector;Corporate governance;Reputationandrelationshipsindustryand??rmlevel(e.g.,RajanandZingales,1998;JayaratneandStrahan,1996).ThethirdstrandofliteratureprovidesevidenceforthelinkandcausalityARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–116581Forexample,relativeto??rmsinFrench-origincountries,??rmsinEnglish-origincountrieshavemoredispersedshareholderownership(LaPorta,Lopez-de-Silanes,andShleifer,LLShereafter,1999),relymoreonexternalcapitalmarketstoraisefunds(LLSV,1997a),havehigherTobin’sQ(LLSV,2002),andcanenteranewmarketorindustrymoreeasily(Djankovetal.,2002).2Endeavorsbyresearchersincludeexaminingthedifferencebetweencontractinginstitutionsand propertyrightsinstitutions(e.g.,Johnsonetal.,2002;AcemogluandJohnson,2003),theendowmentofgeographyanddiseaseenvironmentinformercolonialcountries(e.g.,Acemogluetal.,2001;Becketal.,2003a),thelegalsystem’sabilitytoadapttoevolvingeconomicconditions(e.g.,Posner,1973;Becketal.,1.IntroductionSeveralrelatedstrandsofliteratureonlaw,institutions,??nance,andeconomicgrowthhaveemergedin??nancialeconomicsinrecentyears,andtheirimpactonotherareasofresearchhasbeensigni??cant.First,LaPorta,Lopez-de-Silanes,Shleifer,Vishny(LLSVhereafter)andothershaveproducedasubstantialbodyofempiricalevidencethatlinkstheoriginofacountry’slegalsystemtothecountry’sinstitutionsand??nancialandeconomic‘‘outcomes.’’OneofthecentralresultsofthisliteratureisthatcountrieswithEnglishcommon-laworigin(Frenchcivillaworigin)providethestrongest(weakest)legalprotectiontoboth shareholdersandcreditors(LLSV,1998,2000a).CountrieswithEnglishoriginalsoseemtohavebetterinstitutions,includinglesscorruptgovernments(LLSV,1999),moreef??cientcourts(Djankov,LaPorta,Lopez-de-Silanes,andShleifer,DLLShereafter,2003),andmoreinformativeaccountingstandards(LLSV,1998).Betterlegalprotectionandbetterinstitutions,inturn,leadtobetteroutcomesforthe??nancialsystem,bothattheaggregateand??rmlevels.1RelatedtotheLLSVresults,thereisarecentbodyofliteraturethatattemptstounderstandwhyandhowacountry’slegaloriginaffectsthecountry’sinstitutions,andhowlegaloriginandinstitutions,bothjointlyandseparately,affecteconomicand??nancialoutcomes.2Thesecondstrandofliteraturechampionsthe6><#00aa00'>viewthatthedevelopmentofa??nancialsystemthatincludesastockmarketandintermediationcontributestoacountry’soveralleconomicgrowth(e.g.,McKinnon,1973).Recently,researchers havestrengthenedthis<#00aa00'>viewbypresentingsupportingempiricalevidenceatthecountrylevel(e.g.,KingandLevine,1993;LevineandZervos,1998),aswellasatthe2003b),andreligionandculturalbeliefs(e.g.,Greif,1994;StulzandWilliamson,2003).amonglaw,??nance,andeconomicgrowthatcountry,industry,and??rmlevel(e.g.,Demirgu¨9><>c--KuntandMaksimovic,1998;Levine,1999;BeckandLevine,2002).However,manyoftheabovestudiesareatthecountrylevel,andtheytreateachcountryintheirsampleonanequal-weightbasis.Forexample,amongthecountriesintheLLSV(1998)sample,largediversecountriessuchasBrazilandIndiareceiveARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11659 thesameweightassmallhomogeneouscountrieslikeJordanandEcuador.Wemightexpectthatsmallhomogeneouscountriescouldhavemoreeffectivelegalsystemsbecausetheycanbecloselytailoredtothesecountries’needs.Moreover,mostofthestudiesexcludeoneofthemostimportantdevelopingcountriesintheworld,China.Inthispaper,wedemonstratethatChinaisasigni??cantcounter-exampletothe??ndingsoftheexistingliteratureonlaw,institutions,??nance,andgrowth.Despiteitspoorlegaland??nancialsystems,Chinahasoneofthefastestgrowingeconomiesintheworld.UsingPurchasingPowerParity(PPPhereafter),itpresentlyisthesecondlargesteconomy,andifcurrenttrendscontinue,willovertaketheU.S.andbecomethelargesteconomyintheworldintenyears.WeexaminethreesectorsoftheChineseeconomy:(1)theStateSectorincludesallcompaniessuchthatthegovernmenthasultimatecontrol(state-ownedenterprises,orSOEs);(2)theListedSectorincludesall??rmsthatarelisted onanexchangeandarepubliclytraded;and,(3)thePrivateSectorincludesalltheother??rmswithvarioustypesofprivateandlocalgovernmentownership.3We??ndthatthelaw-??nance-growthnexusestablishedbytheexistingliteratureworkswellfortheStateandListedsectors:Withpoorlegalprotectionofminorityandoutsideinvestors,(standard)externalmarketsareweak,andthegrowthofthese??rmsisslowornegative.However,thesize,growth,andimportanceofthesetwosectorsintheeconomyaredominatedbythoseofthePrivateSector.Inspiteofrelativelypoorerapplicablelegalprotectionandstandard??nancingchannels,thePrivateSectorhasbeengrowingmuchfasterthantheothersandhasbeencontributingtomostoftheeconomy’sgrowth.Ourconclusionfortheimbalanceamongthethreesectorsisthatthereexisteffective,alternative??nancingchannelsandcorporategovernancemechanisms,suchasthosebasedonreputationandrelationships,tosupportthe growthofthePrivateSector.Usingmeasuresfromtheexistingliterature,we??rst??ndthatChina’slawandinstitutions,includinginvestorprotectionsystems,corporategovernance,account-ingstandards,andqualityofgovernment,aresigni??cantlylessdevelopedthanmostofthecountriesintheLLSV(1997a,1998)andLevine(2002)samples.Wealso??ndChina’s??nancialsystemisdominatedbyalargebutunderdevelopedbankingsystemthatismainlycontrolledbythefourlargeststate-ownedbanks.ItsnewlyestablishedShanghaiStockExchange(SHSEhereafter)andShenZhenStockExchange(SZSEhereafter)havebeengrowingveryfastsincetheirinceptionin1990,buttheirscale3ThePrivateSectorincludesthefollowingtypesof??rms:(1)collective-andjointly-ownedcompanies,wherejointownershipamonglocalgovernment,communities,andinstitutionsisforged;and,(2)privately ownedcompanies(butnotpubliclylistedandtraded),wherecontrollingownerscanbeChinesecitizens,investors(orcompanies)fromTaiwanorHongKong,orforeigninvestors(orcompanies).SeeAppendixA.4formoredetails.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11660andimportancearestillnotcomparabletootherchannelsof??nancing,inparticularthebankingsector,fortheentireeconomy.Wenextexamineseparately??nancingchannels,corporategovernance,andthegrowthof??rmsineachofthethreesectors.TheStateSectorhasbeenshrinkingwiththeongoingprivatizationprocess,whichincludes??rmsgoingpublic.OurempiricalresultsontheListedSectorarebasedonasampleofmorethan1,100??rmslistedandtradedonSHSEandSZSE.First,we??ndthattheequityownershipis concentratedwithintheStatefor??rmsconvertedfromtheStateSector,andfounders’familiesfornonstate??rms(e.g.,Claessensetal.,2000,2002).Second,thestandardcorporategovernancemechanismsareweakandineffectiveintheListedSector.Finally,whenweexaminelisted??rms’dividendpoliciesandvaluationsandcomparethemtothoseintheLLSV(2000b,2002)sample??rms,we??ndthatboththedividendratioand??rmvalueofChinese??rmsarelowcomparedtosimilar??rmsoperatingincountrieswithstrongerinvestorprotection,consistentwithLLSVpredictions.MoreinterestingresultsarefoundforthePrivateSector.Ourevidenceismainlybasedonasurveyof17entrepreneursandexecutivesinZhejiangandJiangsuprovinces,twoofthemostdevelopedregionsinChina.First,thetwomostimportant??nancingchannelsforthese??rmsduringtheirstart-upandsubsequentperiodsare??nancialintermediaries,includingstate-ownedbanksandprivatecredit agencies,andfounders’friendsandfamilies.Firmshaveoutstandingloansfrommultiple??nancialintermediaries,withmostoftheloanssecuredby??xedassetsorthirdpartyguarantees.Duringa??rm’sgrowthperiod,fundsfrom‘‘ethnicChinese’’investors(fromHongKong,Taiwan,andothercountries)andtradecreditsfrombusinesspartnersarealsoimportantsources.Whenaskedabouttheprospectofgoingpublic,foundersandexecutiveslist‘‘accesstolargescaleoffunding’’and‘‘reputationincrease’’asthemostimportantbene??ts,and‘‘disclosureofvaluableinformationtocompetitorsandoutsiders’’and‘‘largeamountoffeespaid’’asthemostcriticaldisadvantagesofgoingpublic.Secondly,despitethealmostnonexistenceofformalgovernancemechanisms,alternativemechanismshavebeenremarkablyeffectiveinthePrivateSector.Perhapsthemostimportantoftheseistheroleofreputationandrelationships(Greif,1989,1993).Withoutadominantreligion,themostimportant forceshapingChina’ssocialvaluesandinstitutionsisthewidelyheldsetofbeliefsrelatedtoConfucius;thesebeliefsde??nefamilyandsocialordersandtrust,andaredifferentfromwesternbeliefsontheruleoflaw.Anotherimportantmechanismthatdrivesgoodmanagementandcorporategovernanceiscompetition.Giventheenvironmentoflowsurvivorshipduringearlystagesofa??rm’sdevelopment,??rmshaveastrongincentivetogainacomparativeadvantage.Thethirdimportantmechanismistheroleoflocalgovernments.Withintheregionsthatwitnessedthemostsuccessfuleconomicgrowthandimprovementinlivingstandards,properlymotivatedgovernmentof??cialssupportandparticipateinthegrowthofPrivateSector??rms.OurresultsonthedifferencesamongthethreesectorsinChinachallengethelaw-and-??nance<#00aa00'>viewthatitisthelegaloriginthatcausesthedifferencein??nancialsystems,the??nance-and-growth<#00aa00'>viewthatitisthe developmentofstockmarketsandabankingsystemthatcausesthedifferenceingrowthof??rmsandeconomies,andtransition(fromSocialist,centralplanningsystemstomarket-basedeconomies),availableto??rmsinChinaandothercountries.Section4examinestheListedSector.InSection5,we??rstprovideanecdotalandsurveyevidenceonPrivateSector??rms,ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11661andthendiscussalternative??nancingchannelsandgovernancemechanisms.Finally,weconcludeinSection6.AppendixAcontainsexplanationsofallthevariablesthatweuseinthepaper,andAppendixBprovidesdetailsofourempiricaltestsonthe ListedSector.2.EvidenceonChina’slegaland??nancialsystems,andgrowthinthethreesectorsInthissectionwe??rstprovideanassessmentofChina’sentireeconomy,andthenofthestatusofitslegaland??nancialsystems.WenextcompareChinatothecountriesstudiedintheexistingliterature,namely,theLLSVsampleandtheLevinesample.Finally,wecomparethegrowthintheState,Listed,andPrivatesectorsofChina.2.1.StatusofChina’seconomyTables1Aand1BillustrateChina’sstatusasoneofthemostimportantcountriesintheworld.Attheendof2002,Chinahadapopulationof1.28billionpeople,thelargestofanycountry.Usingsimpleexchangeratecalculations,orUS$1??RMB8.28yuan(inallcurrency-relatedcalculationsthroughoutthepaperunlessotherwisespeci??ed),we??ndthatChina’sGDPrankedsixth intheworld(leftcolumninTable1A).However,ifweusePPPtorecalculateGDPs,China’seconomyisthesecondlargestbehindonlytheU.S.(middlecolumnofTable1A).Moreover,includingEasternEuropeancountries,Russia,Vietnam,andChina(e.g.,McMillan,1997;McMillanandWoodruff,2002).Unlikeexistingresearch,ourpaperprovidesbothaggregateand??rm-levelevidenceonthe??nanceaspectsoftheChineseeconomy,andexamineswhyChinadiffersfromothercountriesstudiedinthestrandsofliteratureonlaw,institutions,??nance,andeconomicgrowth.Therestofthepaperisorganizedasfollows.Section2comparesChina’slegaland??nancialsystemstothoseofothercountries,anddiscussesitsgrowthintheState,Listed,andPrivateSectors.Section3presentsevidenceon??nancingchannelsthe<#00aa00'>viewsupportingthelaw–??nance–growthnexus.Moreover,thesuccessofthePrivateSectorinChinaalsochallengesthe<#00aa00'>viewthat propertyrightsandthelackofgovernmentcorruptionarecrucialindetermining??nancialandeconomicoutcomes.AlthoughourresultsarebasedonChina,similar‘‘substitutes’’basedonreputationandrelationshipsmaybebehindthesuccessofothereconomiesaswell,includingdevelopedeconomies.Thus,athoroughexaminationofthesesubstituteshasmoregeneralimplicationsandcanprovidevaluableguidanceformanyothercountries.SomeofourresultsareconsistentwithexistingresearchoneconomiesinwiththesamePPPapproachandassumingthattheU.S.economycontinuestogrowARTICLEIN PRESSTable1AComparisonofChinaandLLSVcountries:GDPandgrowth??Chinavs.LLSV-SampleCountries??RankGDPin2002GDPin2002usingPPP????AverageannualgrowthrateofGDPusingPPP(1990–2002)CountryGDP(US$Bil.)CountryGDP(Int’l$Billion)Country/groupofcountriesWeightedave.(%)1US(E)a10,416US(E)10,138China11.32Japan(G)a3,978China5,732Englishoriginb5.03Germany(G)1,978Japan(G)3,261Frenchoriginb3.94UK(E)1,552India(E)2,694Germanoriginb3.1 5France(F)a1,409Germany(G)2,171Scandinavianoriginb4.06China1,237France(F)1,5547Italy(F)1,180UK(E)1,5108Canada(E)715Italy(F)1,4819Spain(F)649Brazil(F)1,31110Mexico(F)637Russia1,141aNotes:E,F,GdenotestheEnglish,French,andGermanoriginofthecountry’slegalsystem.bSize-weightedaverageforcountriesinLLSVsample.Sourceforallcountries’GDP:WorldBank.??LegaloriginfollowsLLSVcategory.????TheGDPofeachcountryin2002isconvertedfromlocalcurrencytointernationalDollars,usethePurchasingPowerParity(PPP)conversionfactor.ThePPPconversionfactorisobtainedfromTheWorldBankDevelopmentIndicator(Table5.6,WorldBank.Fordetailsonhowtocalculatetheindicator,see ‘‘HandbookoftheInternationalProgram.’’UnitedNations,NewYork,1992).F.Allenetal./Journalo fFinancialEconomics77(2 005)57–11662ARTIC LEINPRESSTable1BComparisonofChinaandothermajoremergingeconomies:GDPandgrowthRankGDPin2002??GDPin2002onPPPbasis????AnnualgrowthrateofGDPusingPPP(1990–2002)CountryGDP(US$bil.)CountryGDP(Int’l$bil.)CountryGrowthrate(%)1China1,237China5,732China11.32Mexico(F)637India(E)2,694India(E)7.13India(E)515Brazil(F)1,311Pakistan(E)5.74Brazil(F)452Russia1,141Mexico(F)4.3 5Russia346Mexico(F)878Argentina(F)4.06SouthAfrica(E)104SouthAfrica(E)441Brazil(F)4.07Argentina(F)102Argentina(F)401SouthAfrica(E)3.58Pakistan(E)60Pakistan(E)291Russia??2.2Notes:‘‘E’’(‘‘F’’)denotesthelegaloriginofthecountryastheEnglishcommon-lawsystem(Frenchcivil-lawsystem).??GDP??guresarefromtheWorldBank.????SimilartoTable1A,thePPPconversionfactorisobtainedfromTheWorldBankDevelopmentIndicator(Table5.6,WorldBank.Fordetailsonhowtocalculatetheindicator,see‘‘HandbookoftheInternationalProgram.’’UnitedNations,NewYork,1992).F.Allenet al./JournalofFinancial Economics77(2005)57 –11663at4.7%peryearandtheChineseeconomyat11.3%,itwilltakeonlytenyearsbeforeChinaovertakestheU.S.tobethelargesteconomyintheworld.4ItmaybemoreusefultocompareChina’seconomicgrowthwithothermajoremergingeconomiesratherthanthemostdevelopedcountries,sinceChina’srapidgrowthonlystartedin1979.5InTable1BwecompareChinawiththesevenlargestemergingeconomiesintheworld.IntermsofPPP-adjustedGDP??guresin2002,ChinaismorethantwicethesizeofIndia,thesecondlargestemerging economy.IntermsoftheannualgrowthrateofPPP-adjustedGDPsduring1990to2002,ChinaChinacomparesfavorably,haveequalorbettermeasuresofcreditorandARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11664shareholderrights.TheoverallevidencethussuggeststhatthemajorityofLLSV-samplecountrieshavebettercreditorandshareholderprotectionthanChina.WeobviouslycannotdrawourconclusionsregardingthecomparisonoflegalsystemsbasedonTables7><2Aand2Balone.First,thescoresoncreditorandshareholderrightsmainlymeasuretheprotectionofownersofpubliclytradedcompanies,whicharerecentadditionstotheeconomy.Moreimportantly,these4AllofChina’sGNPandGNPgrowth??guresexcludeHongKong.In2003, despitetheimpactoftheSARSepidemic,thegrowthrateofChina’sGDPwas9.1%,thehighestamongthelargesteconomiesintheworld.SeeWorldBank’s‘‘WorldDevelopmentIndicators’’database(August2004)fordetails.ThePPPconversionfactorweuseisobtainedfromtheWorldBankDevelopmentIndicator(‘‘HandbookoftheInternationalProgram,’’UnitedNations,NewYork1992).ThegrowthrateoftheU.S.iscalculatedusingtheperiod1990–2002.5Measuredbysimpleexchangerates,China’sGDPin1980wasUS$180.6billionwhilein1990itreachedUS$368billion.AlsonotethattheexchangeratebetweentheRMBandUS$changedfromUS$1??4.25yuanto8.28yuanin1992,whichintroducedasigni??cantdownwardbiasforChina’sGDP??gurehasbeengrowingmuchfasterthanIndia,whichhasthesecondhighestgrowthrate duringthesameperiod.Moreover,China’spopulationgrowthduringthesameperiodwasslow,anditspercapitaPPP-growthrateisalsothehighestamongthegroupofemergingeconomies.WithChina’srecententranceintotheWTOandthelargepotentialmarketaccessitcanprovide,Chinaispoisedtoplayanincreasinglysigni??cantroleintheworldeconomy.2.2.LegalsystemWe??rstexaminemeasuresofChina’slegalsystemandcomparethemtotheaveragemeasuresofthe49countriesstudiedinLLSV(1998).SeeAppendixA.1foralistofthede??nitionsofthemeasuresusedinthepaper.Intermsofoverallcreditorrights(Table<2A),ChinafallsinbetweentheEnglish-origincountriesthathavethehighestmeasuresofprotection,andFrench-origincountriesthathavethepoorestprotection.China’sshareholderprotectionshowsasimilarpattern(Table2B).Becausethedistributionofthesemeasuresmaybeheavilyskewed towardthetailsduetoafew‘‘outlier’’countrieswithveryhighandlowscores,wealsoprovidethepercentageofcountriesinthesubsamplesandtheentiresampleofLLSVcountriesthathaveequalorhighermeasuresthanChina’s(numbersinbracketsinTables<2Aand2B).AlmosthalfofthecountriesintheFrench-originsubsample,againstwhichin1992.ThisiswhyusingPPP-adjusted??gurestomeasureGDPanditsgrowthismoreappropriate.ARTICLEINPRESS Table<2AAcomparisonofcreditorrights:ChinaandLLSVcountriesCountryEnglishoriginaverageFrenchoriginaverageGermanoriginaverageScandinavianoriginaverageLLSVsampleaverageChina Noautomaticstayonassets0.720.260.670.250.490Securedcreditors??rstpaid0.890.65110.810Restrictionsforgoingintoreorganization0.720.420.330.750.551Managementdoesnotstayinreorganization0.780.260.3300.451(Overall)Creditorrights??3.111.582.3322.32(78%)a(53%)a(83%)a(75%)a(68%)aLegalreserverequiredas%ofcapital0.010.210.410.160.150 Source:China—BankruptcyLawofChina(2000);LLSVcountries—LLSV(1998).??Equalsthesumofthescoresofthefourcategoriesabove,where1??Creditorprotectionisinthelaw,0otherwise.aNumbersinthebracketindicatepercentageofcountriesinthesubsamplewhosemeasureishigherorequaltotwo(China’soverallmeasure).F.Allenetal./Jou rnalofFinancialEconomics 77(2005)57–11665 ARTICLEINPRESSTable2BAcomparisonofshareholderrightsCountryEnglishoriginaverageFrenchoriginaverage GermanoriginaverageScandinavianoriginaverageLLSVsampleaverageChinaOneshare—onevote0.170.290.3300.221Proxybymailallowed0.390.0500.250.180Sharesnotblockedbeforemeeting10.570.1710.710Cumulativevoting/Proportionalrepresentation0.280.290.300.270 Oppressedminority0.940.290.500.531Preemptiverighttonewissue0.440.620.330.750.531Percentageofsharecapitaltocallanextraordinaryshareholdermeeting0.090.150.050.10.110.1Antidirectorrights??42.332.33333(94%)a(45%)a(33%)a(75%)a(65%)aMandatorydividend00.11000.050Source:China—CompanyLawandCommercialCodesofChina(2000);LLSVcountries—LLSV(1998).??OverallscoreisthesumofthescoresonRows(2),(3),(4),(5),and(7),wherescore??1,whentheprotectionisinthelaw;0otherwise.aNumbersinthebracketindicatepercentageofcountriesinthesubsamplewhosemeasureishigherorequaltothree(China’soverallmeasure).F .Allenetal./JournalofFin ancialEconomics77(200 5)57–11666ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11667scoresmeasurethelegalsystemonpaper,notinpractice:WhiletheChinesegovernmenthasadoptedvariousprotectionmeasuresintothelaw,fromwhichwe calculatethescores,onecanarguethatamoreimportantmeasurewouldbelawenforcement.Table2<>Cprovidessomeevidenceinthisregard,withallthemeasures(includingthoseforChina)drawnfromindependentinternationalratingagencies.Fortwokeycategoriesoflawenforcement,theruleoflawand(government)corruption,China’smeasuresaresigni??cantlybelowallaveragemeasuresofLLSV-samplecountries,regardlessoftheirlegalorigins.ThissuggeststhatthescoresofcreditorandshareholderprotectionofChinainTables<2Aand2Barenotreliable.WealsocompareChina’slegalsystemtothoseofotheremergingcountriesinTable2D,similartothegrowthcomparisonabove.China’scorruptionindexistheworstamongthesevendevelopingcountries,whileitsmeasureofantidirectorrights(creditorrights)isonlyhigherthanthatofIndiaandMexico(ArgentinaandMexico).Hence,thedevelopmentofChina’slegalsystemisnotaheadofanyofthe othermajoremergingeconomies,anditisclearlydominatedbythosethathaveEnglishcommon-laworigin(India,Pakistan,andSouthAfrica).Inordertohaveaneffectivelawenforcementsystem,acountrymusthaveanindependentandef??cientjudicialsystemwithasuf??cientsupplyofquali??edlegalprofessionals.First,DLLS(2003)comparetheef??ciencyandformalismofthejudicialsystemacross109countriesincludingChina.Theresultsarebasedonhowtwospeci??<>ctypesofdisputes,theevictionofatenantandcollectionofabouncedcheck,areresolvedinacountry’sjudicialsystem.SincebothtypesofdisputesarerareeventsinChina,astherealestatemarket(includingtherentalmarket)andtheuseofpersonalchecksareunderdevelopedandlimitedtoafewlargecities,theirresultsarenotverymeaningfulforChina.Ontheotherhand,theMinistryofJusticeofChinastatesthatthereare110,000lawyersand9,000law??rmsasof2002,whileOrts(2001)estimatesthatthereare150,000lawyersinChina,roughly thesamenumberoflicensedattorneysasinthestateofCalifornia.Lawyersrepresentonly10–25%ofallclientsincivilandbusinesscases,andevenincriminalprosecutions,lawyersrepresentdefendantsinonlyhalfofthecases.Amongtheapproximately??vemillionbusinessenterprisesinChina,only4%ofthemcurrentlyhaveregularlegaladvisers.Moreover,onlyone-??fthofalllawyersinChinahavelawdegrees,andevenalowerfractionofjudgeshaveformallystudiedlawatauniversityorcollege.Needlesstosay,itwillbealongtimebeforeChinahasastronglegallaborforce.AnotherreasonthatmanynewlawsarenoteffectivelyenforcedinChinaistheintrinsiccon??ictofinterestbetween‘‘fairplay’’inpracticinglawandthemonopolypowerofthesinglerulingparty,especiallyincasesinwhichgovernmentof??cialsortheiraf??liatesareinvolved.Consistentwiththisargument,LaPorta,Lopez-de-Silanes,Pop-Eleches,andShleifer(LLPS,2004)??ndthatChinaranks amongtheworstcountriesintermsofpoliticalfreedomaswellastheprotectionofpropertyrights.Theyalso??ndapositivecorrelationbetweenpoliticalfreedom(constitutionalrules)andmeasuresofeconomicfreedom(propertyrights,proceduresofstart-up??rm)acrosscountries,andthatjudicialindependenceaccountsforthepositiveeffectofcommon-lawlegaloriginoneconomicfreedom.However,thefactthatChinascoresextremelypoorlyonbothpoliticalandeconomicfreedomsandyetenjoyedARTICLEINPR ESSTable2<>CAcomparisonoflawenforcementCountryEnglishoriginaverageFrenchoriginaverageGermanoriginaverageScandinavianoriginaverageLLSVsample averageChinaEf??ciencyofjudicialsystem8.156.568.54107.67N/aRuleoflaw6.466.058.68106.855Corruption7.065.848.03106.902Riskofexpropriation7.917.469.459.668.05N/aRiskofcontractrepudiation7.416.849.479.447.58N/aAccountingratingonaccountingstandards69.6251.1762.677460.93N/aSource:China—Internationalcountryrisk(ratingagency);LLSVcountries—sameasabove.F.Allen etal./JournalofFinan cialEconomics77(2005)5 7–11668ARTICLEINPRESSTable2D Acomparisonoflegalsystems:ChinaandothermajoremergingeconomiesEf??ciencyofjudicialsystemRuleoflawCorruptionAnti-directorrightsOneshare-onevoteCreditorrightsAccountingstandardsChinaN/a52312N/aIndia(E)84.174.5820457Pakistan(E)53.032.98414N/a S.Africa(E)64.428.9240470Argentina(F)65.356.0240145Brazil(F)5.756.326.3231254Mexico(F)65.354.7700060Source:China—Internationalcountryrisk(ratingagency);allothercountries—LLSVsources;‘‘E’’(‘‘F’’)denotesthelegaloriginofthecountryastheEnglishcommon-lawsystem(Frenchcivil-lawsystem).F.Allenetal. /JournalofFinancialEco nomics77(2005)57–1 1669ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11670oneofthefastesteconomicgrowthratescastsdoubtontheimportanceofpoliticalfreedomandeconomicfreedomasmeasuredinLLPS.Finally,wecommentonthecurrentstatusofChina’saccountingsystem.Thereformstartedin1992,withtheenactmentofregulationsgoverningenterpriseswithforeigninvestment.Sincethen,theAccountingStandardsforBusinessEnterprisesofChina,togetherwiththe13-industryregulationboard,havebeentryingtomoveChina’saccountingpracticeintheListedSectortowardtheIAS(International AccountingStandards).However,themostglaringprobleminChina’saccountingsystemisthelackofindependent,professionalauditors,similartothesituationforlegalprofessionals.ThisimpliesthattheproposedIAS-basedstandardsmaybecounterproductivewithinChina’scurrentinfrastructure:Withfewauditorsunder-standingandenforcingthenewstandards,andgiventhelackofaneffectivejudicialsystem,embezzlementofcompanyassetsandotherformsoffraudmayoccurmorefrequentlyunderIAS-basedstandards,ascomparedtoanalternativesystemwithamuchsimplersetofaccountingstandards(e.g.,Xiang,1998).2.3.FinancialsystemWe??rstexamineChina’s??nancialsystemattheaggregatelevel,includingbothits??nancialmarketsandbankingsystem.Wethenexamineitsstockexchangesinmoredetailandbrie??ydiscussitsventurecapitalmarkets.Finally,weexamineproblemsinthebankingsector. InTable3wecompareChina’s??nancialsystemtothoseoftheLLSV-samplecountries(LLSV,1997a,1998),usingmeasuresfromLevine(2002).We??rstcomparethesizeofacountry’sequitymarketsandbanksrelativetothatcountry’sGDP.China’sstockmarkets,whichhavebeengrowingveryrapidlysince1990,aresmallerthanthoseofmostoftheothersamplecountries,bothintermsofmarketcapitalizationandthetotalvaluetradedasafractionofGDP.Noticethat‘‘totalvaluetraded’’isabettermeasurethan‘‘marketcapitalization’’becausethelatterincludesnontradableshares,whiletheformermeasuresthefractionoftotalmarketcapitalizationtradedinthemarkets,orthe‘‘??oatingsupply’’ofthemarket.WefurtherdiscussthisissueinTable4AofthissectionandSection4below.Incontrast,China’sbankingsystemismuchmoreimportantintermsofsizerelativetoitsstockmarkets,withitsratiooftotalbankcredit toGDP(1.11)higherthaneventheGerman-origincountries(withaweightedaverageof0.99).However,whenweconsiderbankcreditissued(orloansmade)tothePrivateSectoronly,China’sratiodropssharplyto0.24,suggestingthatmostofthebankcreditisissuedtocompaniesintheStateandListedSectors.Moreover,China’sbankingsystemisnotef??cient:Itsoverheadcosttototalassets(0.12)ismuchhigherthantheaverageofFrench-origincountries(0.05),thenext-highestgroupofcountries.Chinaalsohasthelowestscoresforboth‘‘Structureactivity’’and‘‘Structuresize’’(secondpanelofTable3),suggestingthatitsbankingsectorismuchlargerthanits??nancialmarkets,andthisdominancebythebanksovermarketsisstrongerthantheaverageofallLLSV-samplecountries.Intermsof‘‘Structureef??ciency,’’whichdenotestherelativeef??ciencyofmarketsvs.banks,Chinahasthehighestscore, ARTICLEINPRESSTable3Acomparisonof??nancialsystems:bank-vs.market-basedmeasures(value-weightedapproach)MeasuresEnglishorigin??Frenchorigin??Germanorigin??Scandinavianorigin?? SampleaverageChinaBankandmarketsizeBankcredit/GDP0.620.550.990.490.731.11(0.24)aOverheadcost/Banktotalassets0.040.050.020.030.030.12Totalvaluetraded/GDP0.310.070.370.080.270.11Marketcapitalization/GDP0.580.180.550.250.470.32Structureindices:Marketsvs. banks????Structureactivity??0.76??2.03??1.14??1.83??1.19??1.07(0.46)aStructuresize??0.10??1.05??0.77??0.69??0.55??1.24(0.29)aStructureef??ciency??4.69??6.00??5.17??6.17??5.17??1.48(??3.07)aStructureaggregate1.21??0.050.660.130.72N/aStructureregulatory7.028.2110.157.728.9516Financialdevelopment(bankingandmarketsectors)Financeactivity??1.18??3.38??0.84??2.86??1.58??0.85(??2.38)aFinancesize5.104.295.224.604.95??1.02(??2.55)a Financeef??ciency2.180.442.851.042.01??0.60(1.14)aFinanceaggregate1.230.131.470.481.05N/ANotes:AllthemeasuresforcountriesotherthanChinaaretakenfromLevine(2002);measuresonChina(inTable3)arecalculatedusingde??nitionsfromLevine(2002)(seeAppendixA.2forlistofde??nitions).Sources:AlmanacofChina’sFinanceandBanking(2000);ChinaStatisticalYearbook(2000).??Thenumericalresultsforcountriesofeachlegalorigingrouparecalculatedbasedonavalue-(GDPofeachcountry)weightedapproach.????Measuringwhetheracountry’s??nancialsystemismarket-orbank-dominated;thehigherthemeasure,themorethesystemisdominatedbymarkets.aNumbersinbracketindicatebankcreditissuedtoonlyprivatesectors(insteadoftotalbankcredit).F.A llenetal./JournalofFinan cialEconomics77(2005) 57–11671ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11672Table4AAcomparisonofthelargeststockmarketsintheworld(2002)RankStockmarketTotalmarketcapConcentrationTurnovervelocity(US$billion)(%)(%) 1.NYSE9,01561.394.82.Tokyo2,09560.667.93.Nasdaq1,99463.1319.5??4.London1,80084.597.35.Euronext1,53872.3153.66.DeutscheBo¨rse68672.0125.17.Toronto57067.867.98.Swiss54781.2138.69.Italian47766.1120.710.China(HongKong)46383.039.711.China(domestic)46329.4224.2Notes:1.All??gures(exceptthoserelatingtoChina’sdomesticexchanges) arefrom//.??bv>,thewebsiteoftheinternationalorganizationofstockexchanges.TheChinesedataisfrom//.csrc.gov,thewebsitefortheChinaSecurityRegulationCommittee(CSRC).suggestingthatitsstockmarketsareactuallyrelativelymoreef??cientthanbankscomparedtoothercountries.ThisresultismostlydrivenbytheextremelyhighoverheadcostsofChina’sbankingsystem.Finally,wecomparethedevelopmentoftheentire??nancialsystem(‘‘Financialdevelopment’’),includingbothbanksandmarkets(lastpanelofTable3).Givenallothercountries’measuresarebasedonprivatebankcreditonly,ifweonlyincludeChina’sprivatebankcredit,we??ndthatChina’soverall??nancialmarketsize(‘‘Financeactivity’’and‘‘Financesize’’)issmallerthantheLLSV-sampleaveragelevel,andisonlyhigherthantheFrench-origincountries’average.Intermsofthe ef??ciencyofthe??nancialsystem,China’smeasureisbelowallsubsamplesofLLSVcountries.Basedontheaboveevidence,wecanconcludethatChina’s??nancialsystemisdominatedbyalargebutinef??cientbankingsector.Financialmarkets.China’sdomesticstockexchanges,SHSEandSZSE,withtheircombinedtotalmarketcapitalization,includingnon-tradableshares,rankeleventhamongthelargeststockexchangesintheworldattheendof2002(Table4A).Inaddition,theHongKongStockExchange(HKSEhereafter),whereselected??rmsfromMainlandChinacannowbelistedandtraded,isrankedtenthintheworld.Ifwerankthecombinedsizeofallstockexchangesinacountry,Chinawouldrank??fth,behindonlytheU.S.,Japan,theU.K.,andFrance.2.All??guresrelatetotheperiodof01/01/2002to12/31/2002.3.Concentrationistheratioofthecombinedmarketcapitalization(includingnontradableshares)of??rms rankedinthetop5%bycapitalizationoverthetotalcapitalization(includingnontradableshares)ofall??rmslistedontheexchange.4.Turnovervelocityisthetotalturnoverfortheyearexpressedasapercentageofthetotalmarketcapitalization.??TurnovervelocityforNasdaqincludesdoublecounting;theactual??gureshouldbehalfofthereported??gure.AsfastasthegrowthofChina’sstockmarketshasbeen,thesemarketsarenotef??cientinthatpricesandinvestorbehaviordonotre??ectfundamentalvaluesoflisted??rms.InTable4A,‘‘Concentration’’measuresthefractionoftotalmarket ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11673capitalizationofanexchangethatiscomingfromthecombinedcapitalizationofthelargest??rmsrankedinthetop5%(bycapitalization).Thedominanceoflarge-capstocksinChinaisthelowestamongmajorstockexchangesintheworld,withitsconcentrationratioof29.4%lessthanhalfofthatofTokyo,whichhasthesecond-lowestconcentration.Ontheotherhand,medium-andsmall-capstocksaretradedextremelyfrequentlyinChina,asshownbythehigh‘‘TurnoverVelocity,’’de??nedasthetotalturnoverfortheyearexpressedasapercentageoftotalmarketcap.China’svelocityof224.2%isevenhigherthanthatofNASDAQ,6withthewell-knowntradingpatternsofmanysmallandmediumtechnologystocks(theconcentrationratioofNASDAQis63.1%).Consistentwithour??ndings,Morcketal.(2000)??ndthatstock pricesaremoresynchronousinemergingcountries,includingChina,thanindevelopedcountries.Theycontributethisphenomenontopoorminorityinvestorprotectionandimperfectregulationofmarketsinemergingmarkets.Oneexampleistherestrictiononshort-sales.Brisetal.(2003)??ndthatlimitingshort-salescontributestothehighco-movementofstockprices,butdoesnottendtoincreasetheprobabilityofamarketcrash,ascommonlyfearedbygovernmentsinemergingcountries.Theinef??cienciesintheChinesestockmarketscanbeattributedtopoorandineffectiveregulation.Basedonastudyofsecuritieslawswiththefocusonthepublicissuanceofnewequityin49countries(Chinanotincluded),LaPorta,Lopez-de-Silanes,andShleifer(LLShereafter,2003)??ndthatprivateenforcementoflawsthroughdisclosureandliabilityrulesissuperiortostrongregulationbythe governmentinpromotingstockmarketdevelopment.GivenChina’spoordisclosurerules,accountingstandards,andjudicialsystems,theLLS(2003)resultcanbeusedtoexplainthestatusofChina’sstockmarkets.Toimprovethequalityofgovernmentregulation,Glaeseretal.(2001)arguethatregulatorsmustbeproperlymotivated.TheconcentrationandturnovervelocityofChina’smarkets(Table4A)wereactuallyevenhigherinthelate1990s,andtheimprovementisinpartduetoadvancesinthequalityofregulation.Next,webrie??yexaminetheroleof??nancialmarketsinhelping??rmsraisefunds(Table4B).Boththescaleandrelativeimportance(comparedwithotherchannelsof??nancing)ofChina’sexternalmarketsarenotsigni??cant.Forexample,fortheratioofexternalcapitalandGNP,theLLSV(1997a)sampleaverageis40%,comparedtoChina’s16%(usingonlythe??oatingsupplyorvaluetradedpartofthestockmarket,ratherthanthetotalmarketcap);fortheratiooftotaldebt (includingbankloansandbonds)toGNP,theLLSV-sampleaverageis59%,comparedtoChina’s35%.However,ifweincludealldebt,includingbankloansissuedtoallsectorsincluding6TheactualturnovervelocityofNasdaqshouldbehalfofthereported??gure,319.5%.ThisisbecauseunlikeNYSEandmostotherexchangesaroundtheworld,Nasdaqdealersreportboththebuyandselltradesseparately,whichleadstodoublecountinginthecalculationofvelocity.SeeAtkinsandDyl(1997)formoredetails.ARTICLEINPRESSTable4BAcomparisonofexternalcapitalmarkets(mean)CountryEnglish originFrenchoriginGermanoriginScandinavianoriginaverageLLSVsampleChina(2002)F.Allenetal./JournalofFinancialEconomics77(2005)57–11674theStateSector,thedebt/GNPratioincreasesto79%,suggestingthat themajorityofdebtdoesnotgothroughthecapitalmarkets.Weprovide??rm-levelevidenceon??nancinginSection4below.Finally,webrie??ydiscussChina’sventurecapitalmarkets,whichshouldberegardedaspartofthe??nancialmarketsratherthantheintermediationsector(e.g.,AllenandGale,2000a).ItisoftenarguedthatoneofthereasonstheU.S.hasbeensosuccessfulindevelopingnewindustriesinrecentyearsistheexistenceofastrongventurecapitalsector(e.g.,KortumandLerner,2000).ConsistentwithourpreviousaverageaverageaverageaverageExternalcapital/GNP0.600.210.460.300.400.49(0.16)??Domestic??rms/Pop35.4510.0016.7927.2621.590.93IPOs/Population2.230.190.122.141.020.05Totaldebt/GNP0.680.450.970.570.590.35 GDPgrowth(one-year)4.303.185.292.423.796.77Ruleoflaw6.466.058.6810.006.855Antidirectorrights3.391.762.002.502.443Oneshare—onevote0.220.240.330.000.221Creditorrights3.111.582.332.002.302Sources:LLSV(1997a)paper;AlmanacofChina’sFinanceandBanking(2003).??Externalcapital/GNPratiousingthe??oatingsupplyorvaluetradedportionofthemarketcapitalization.??ndings,China’sventurecapitalindustry,sinceitsinceptioninthe1980s,isunderdevelopedanditsroleinsupportingthegrowthofyoung??rmsisverylimited.Moreover,basedoninterviewsconductedwith36venturecapitalistsin24venturecompanies,BrutonandAhlstrom(2002)??ndthatthelimitedformalrulesand regulationsareoftenineffective,whilealternativemechanismsbasedonreputationandrelationshiparethenorminallstagesandphasesoftheindustry.Insummary,theoverallevidenceonthecomparisonofChinaandothercountries’externalmarketsisconsistentwithLLSV(1997a,1998)predictions:Withanunderdevelopedlegalsystem,thefactthatChinahassmallexternalmarketscomesasnosurprise.Fig.1comparesChina’slegalsystemandexternal??nancialmarketstothoseofLLSVcountries.Thehorizontalaxismeasuresoverallinvestorprotectionineachcountry,whiletheverticalaxismeasuresthe(relative)sizeandef??ciencyofthatcountry’sexternalmarkets.7CountrieswithEnglishcommon-lawsystems7FollowingLLSV,thescoreonthehorizontalaxisisthesumof(overall)creditorrights,shareholderrights,ruleoflaw,andgovernmentcorruption.Thescoreoftheverticalaxisindicatesthedistanceofa country’soverallexternalmarketsscore(externalcap/GNP,domestic??rms/Pop,IPOs/Pop,Debt/GNP,andLogGNP)tothemeanofallcountries,withapositive(negative)??gureindicatingthatthiscountry’soverallscoreishigher(lower)thanthemean.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11675Singapore(E)HK(E)Malaysia(E)US(E)SouthAfrica(E)Japan(G)UK(E) 2468ExternalMarket(Frenchcivil-lawsystems)lieinthetop-rightregion(bottom-leftregion)ofthegraph,whileChinaisplacedclosetothebottom-leftcornerofthegraph.Bankingsector.China’sbankingsectorisdominatedbyfourlarge andinef??cientstate-ownedbanks.LLS(2002)showthatthegovernmentowns99.45%ofthe10largestcommercialbanksinChinain1995(100%in1970);thisownershiplevelisoneofthehighestintheirsampleof92countries.Moreover,theLLSresultonthenegativerelationbetweengovernmentownershipofbanksandthegrowthofacountry’seconomyseemstoapplytoChina’sStateSectorandthestatusquoofitsbankingsector.However,highgovernmentownershiphasnotsloweddownthegrowthofthePrivateSector.ThemostglaringproblemforChina’sbankingsectoristheamountofnonperformingloans(NPLshereafter)withinthefourlargeststate-ownedbanks.AlargefractionofthesebadloansresultedfrompoorlendingdecisionsmadeforSOEs,someofwhichwereduetopoliticalorothernoneconomicreasons.TheTaiwan(G) Thailand(E)France(F)German(G)Brazil(F)Argentina(F)Mexico(F)ChinaIndia(E)Pakistan(E)Indonesia(F)SouthKorea(G)-4-20101214161820222426InvestorProtection Fig.1.Comparisonoflegaland??nancialsystems.ItcomparesChina’slegalsystemandexternal??nancialmarketstothoseofLLSVcountries(LLSV,1997a,1998).FollowingLLSV(1997a,1998),thescoreonthehorizontalaxismeasuresoverallinvestorprotectioninacountry.Itisthesumof(overall)creditorrights,shareholderrights,ruleoflaw,andgovernmentcorruption.Theverticalaxismeasuresthe(relative)sizeandef??ciencyofthatcountry’sexternalmarkets.Thescoreofacountrymeasuresthedistanceofthecountry’soverallexternalmarketsscore(externalcap/GNP,domestic??rms/Pop,IPOs/Pop,Debt/GNP,andLogGNP)tothemeanofallcountries,withapositive(negative)??gureindicatingthatthiscountry’soverallscoreishigher(lower)thanthemean. additionalproblemisthatdataavailabilityonNPLsislimited,whichcanbeviewedasastrategicdisclosuredecisionofthegovernment.However,thislackofdisclosureofNPLsonlyfuelsspeculationsthattheproblemmustbesevere.Forexample,Lardy(1998)arguesthatifinternationalstandardswereused,theexistingNPLswithinthestate-ownedbanksasofthemid-1990swouldmakethesebanks’totalnetworthnegative.Tables5Aand5BcompareNPLsandbankingsystempro??tabilityinChinaandsixothermajorAsianeconomiesinrecentyears.InformationonChina’sNPLs??rstbecameavailablein1998,butthe??guresin1998and1999inTable5Aprobablysigni??cantlyunder-estimatetheactualsize.Duringtheperiodof2000–2002,ChinahasthelargestamountofNPLsamongthesevenAsianeconomies,eitherasafractionoftotalnewloansmadebyallbanksorasafractionofGDPinagivenyear. ThiscomparisonincludestheperiodduringwhichAsiancountriesrecoveredfromthe1997??nancialcrisis,andtheperiodduringwhichtheJapanesebankingsystemARTICLEINPRESSTable5AAcomparisonofnonperformingloansofbankingsystems199719981999200020012002ChinaN/a2.0(2.2)9.5(10.6)18.9(24.9)16.9(22.7)12.6(15.2)HongKong1.3(3.0)4.3(10.2)6.3(13.9)5.2(12.6)4.9(12.9)3.7(9.6)IndiaN/a7.8(1.6)7.0(1.6)6.6(1.6)4.6(1.7)2.2(0.8)Indonesia0.3(0.2)11.8(4.6)8.1(2.0)13.6(3.2)9.9(2.2)4.5(0.9)Japan2.7(5.4)5.1(10.8)5.3(10.9)5.8(11.5)9.2(15.3)7.4(12.8)SouthKorea2.9(5.1)4.8(6.3)12.9(12.9)8.0(8.6)3.4(3.4)2.5 (2.6)Taiwan2.4(3.2)3.0(3.9)4.0(5.7)5.2(7.6)6.2(9.4)4.1(5.2)Notes:NPLismeasuredas%oftotalloansmade,andas%ofGDP(numbersinbrackets).BoththeloanandNPLaretheaggregateofallbanksinacountry.Source:TheAsianBankerdatacenter2003.//.thesianbanker>F.Allenetal./JournalofFinancialEconomics77(2005)57–11676Table5BAcross-countrycomparisonofbankingsystempro??tabilityThepro??tabilityismeasuredasthereturnonaverageequity(ROAE),andreturnonaverageassets(ROAA).Thelatterispresentedinthebrackets.199719981999200020012002China6.6(0.2)4.0(0.2)3.2(0.18)3.9(0.2)3.5(0.2)4.16(0.2) HongKong18.7(1.8)11.0(1.0)18.2(1.6)18.8(1.6)15.7(1.4)15.6(1.4)India17.0(0.9)9.7(0.5)14.2(0.7)10.9(0.5)19.2(0.9)19.6(1.0)Indonesia??3.8(??0.3)N/aN/a15.9(0.3)9.7(0.6)21.1(1.4)Japan??18.6(??0.6)??19.2(??0.7)2.7(0.1)??0.7(0.0)??10.4(??0.5)??14.5(??0.6)SouthKorea??12.5(??0.6)??80.4(??3.0)??34.0(??1.5)??7.0(??0.3)15.8(0.7)13.1(0.6)Taiwan11.2(0.9)9.5(0.8)6.9(0.6)5.1(0.4)4.0(0.3)??5.2(??0.4)Source:TheAsianBankerdatacenter2003.//.theasianbanker>wasdisturbedbytheprolongedNPLproblem.Moreover,thepro??tabilityofChina’sbankingsystem,measuredbythereturntoequityorassets,isalsoamongthelowest8%to60%.Second,state-ownedbankshaveimprovedtheirloanstructureby ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11677increasingloansmadetoindividuallenderswhilebeingmoreactiveinriskmanagementandmonitoringofloansmadetoSOEs.Forexample,theratioofconsumerlendingtototalloansmadeforthefourstate-ownedbanksincreasedfrom1%in1998to10%in2002.Third,therehasbeenaboomintheentryandgrowthofnon-state??nancialintermediaries,andthistrendisexpectedtocontinuewithmoreforeignbanksenteringthedomesticcreditmarketsasaresultofChina’sentranceintotheWTO.In1997,totalnewloansmadebythefourlargeststate-ownedbanksaccountedformorethan75%ofallnewloans,whilenewloansmadeby‘‘shareholding’’banksaccountedforlessthan7%.In2001,theshareofnewloansmadebystate-ownedbanksdroppedto49%whilethefractionofnewloansmadeby shareholdingbanksroseto23.5%.AlltheabovefactstakentogethercanexplainwhyNPLshavebeenfallinginrecentyears,asre??ectedinTable5A.Tosummarize,thecontinuingeffortofreformingandimprovingthebankingsystemisoneofthemostimportanttasksforChinainthenearfuture.Infact,Chinarecentlyannouncedthatitscentralbankwillinjectforeigncurrencyreservesintotwoofthebigfourstate-ownedbanks,toimprovetheirbalancesheetsandenhancethelikelihoodthatthesebankscangopublicbytheendof2004.Similarfundinjectionplansfortheothertwostate-ownedbanksarealsointheworks.GiventhatChina’stotalforeignexchangereserveisUS$400billionwhilethetotalamountofNPLsasof2002is15%ofGDP,orUS$188billionusingtheUS$1??8.26RMBexchangerate,theforeignreserveitselfshouldbemorethanenoughtoremovetheNPLsoffthebooksofallthebanksinChina.8Whetherthegovernmentwilldoexactlythis remainstobeseen,butitisclearthattheultimatesourceofeliminatingNPLsliesinoveralleconomicgrowth.Aslongastheeconomymaintainsitsstronggrowthmomentumsothatthegovernment’staxableincomealsoincreases(e.g.,SachsandWoo,1997;Rawski,2002),thegovernmentcanalwaysassumetheremainderoftheNPLswithoutsigni??cantlyaffectingtheeconomy.2.4.GrowthintheState,Listed,andPrivatesectorsTable6AcomparesthegrowthofindustrialoutputproducedintheStateandListedsectorsvs.thatofthePrivateSectorfrom1996to2002.ThePrivateSector8Onepotentialriskforusingforeignreservesisthepressureofsigni??cantRMBappreciationrelativetoUS$andotherforeigncurrencies.See,forexample,Broda(2004)andAlesinaetal.(2002)ontheinthesamegroupofeconomies(Table5B).InrecentyearstheChinesegovernmenthastakenactivemeasurestoresolvethis problem.First,fourstate-ownedassetmanagementcompanieswereformedwiththegoalofassumingtheseNPLsandliquidatingthem.Informationfromthesecompanies’auctiondatashowsthatthecashrecoveryonthebadloansrangesfromrelationshipbetweenexchangeregime,trade,andeconomicperformanceindevelopingcountries.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11678Table6AGrowthratesoftheState,Listed,andPrivatesectorsInthistable,PanelAdisplaysthegrowthrateof‘‘industrialoutput’’forthetwosectorsinChina.TheStateandListedsectorsincludesstate-ownedandpubliclytradedcompaniessuchthatthegovernment holdscontrollingshares.ThePrivateSectorconsistsof??rmswithallothertypesofownershipstructures.DatasourceforthistableistheChineseStatisticalYearbook2000,2001,2002,and2003.Foreachsector,wealsocalculatetheweighted-averagegrowthrateacrosstheselectedownershiptypes.PanelBdisplaystheaveragegrowthrateof‘‘investmentin??xedassets’’forthetwosectors.Growthrate(%)PanelA:industrialoutputPanelB:investmentin??xedassetsYearState&ListedsectorsPrivateSectorState&Listed sectorsPrivateSector??199615.917.410.217.31997??0.618.99.06.11998??6.510.217.49.019995.86.83.87.5200014.024.23.511.420014.69.96.712.620026.512.57.216.8Ave.Annualrate(95–02)5.414.38.211.5Sources:ChinaStatisticYearbooks2000–2003.??Includesforeign-ownedcompanies,companiesownedbyinvestorsfromTaiwanandHongKong,andTVEs. dominatestheStateandListedsectorsintermsofboththesizeoftheoutput,andthegrowthtrend:Totaloutputin1999isUS$1200billionforthePrivateSector,whileitisaroundUS$400billionintheStateandListedsectorscombined;thePrivateSectorgrewatanannualrateof14.3%between1996and2002,whilethecombinedStateandListedsectorsgrewat5.4%duringthesameperiod(PanelA).Inaddition,thegrowthratesforinvestmentin??xedassetsofthesesectorsarecomparable(PanelB),whichimpliesthatthePrivateSectorismoreproductivethantheStateandListedsectors.Finally,therehasbeenafundamentalchangeamongtheState,Listed,andPrivatesectorsintermsoftheircontributiontotheentireeconomy:TheStateSectorcontributed76%ofChina’stotalindustrialoutputin1980,butin1996itonlycontributed28.5%;in1980,individuallyowned??rms,whichareasubsetofPrivateSector??rms,werenegligible,butin1996theycontributed15.5%oftotalindustrialoutput;theabovetrendofthe PrivateSectorreplacingtheStateSectorwillcontinueinthenearfuture.Table6Bpresentsthenumberofnonagriculturalemployeesinthethreesectors.ThePrivateSectorisamuchmoreimportantsourceforemploymentopportunitiesthantheothertwosectors.Overtheperiodfrom1995to2002,thePrivateSectoremployedanaverageofover70%ofallnonagriculturalworkers,whiletheTownshipVillageEnterprises(TVEshereafter),alsoasubsetofPrivateSector??rms,arebyfarthemostimportantemployerforworkersfromtheruralareas.Moreover,thenumberofemployeesworkinginthePrivateSectorgrewatarateof1.5%peryearoverthisseven-yearperiod,whilethelaborforceintheStateandListedsectorsretracted.ThesepatternsareparticularlyimportantforChina,givenitsvastARTICLEINPRESS F.Allenetal./JournalofFinancialEconomics77(2005)57–116793.Firms’??nancingsources:Aggregateevidenceandcross-countrycomparisonsInthissectionwecompare,attheaggregatelevel,how??rmsraisefundsinChinaandinLLSV-samplecountrieswiththeemphasisonemergingeconomies.Itisthenworthwhiletostudywhatotherchannelsof??nancingareplayingtheroleofsubstitutingforexternalcapitalmarketsandstandard,textbook??nancingchannels.populationandpotentialproblemofunemployment.Botero,Djankov,LaPorta,Lopez-de-Silanes,andShleifer(BDLLS,2003)comparelaborlawsandsocialsecuritysystemsacross85countriesincludingChina,and??ndthatFrenchlegalorigin,socialist,andpoorcountrieshavehigherlevelsoflaborregulationthanEnglishcommon-lawandrichcountries.TheirevidenceonChinaexcludesthelaborforceintheruralareas.GiventheimportanceofTVEsintermsof employment,thislimitstheapplicationoftheirresultstoChina.Table6BEmploymentintheState,Listed,andPrivatesectorsYear1995199619971998199920002001200295–02AnnualgrowthratePanelA:numberofemployees(million)??State&Listedsectors1151161159489858177??5.7%PrivateSector2212332332352402332452461.5%PanelB:Percentageoftotalemployeesbelongingtoeachsector(%)State&Listedsectors34.333.333.028.727.226.824.923.8PrivateSector65.766.767.071.372.873.275.176.2Source:ChinaStatisticYearbooks2000–2003.??Indicatesnon-agriculturalemployees.3.1.China’smostimportantfinancingchannelsThefourmostimportant??nancingsourcesforall??rmsinChina,in termsof??xedassetinvestments,are:(Domestic)bankloans,??rms’self-fundraising,statebudget,andforeigndirectinvestment.Byfarthetwomostimportantsourcesof??nancingchannelsareself-fundraisingandbankloans.ConsistentwithpreviousevidenceonChina’sbankingsector,bankloans,includingloansfromthenonstatebanks,providealargeamountoffundsto??rms,andconstitutealargefractionof??rms’total??nancingneeds.Forexample,??rmsintheStateSectorrelyonbankloanstoraisemorethan25%oftheirtotal??nancingneeds.Asimilarpatternholdsforjointly-andcollectively-ownedcompanies,bothofwhichbelongtothePrivateSector.Oursurveyevidencebelow(Section5)alsoindicatesthatbankloansareimportant??nancingsourcesforthePrivateSector,especiallyduringthe??rms’start-upperiod.Self-fundraisingincludesproceedsfromcapitalraisedfromlocalgovernments(beyondthestatebudget),communities,otherinvestors,internal ??nancingchannelssuchasretainedearnings,andallotherfundsraiseddomesticallybythe??rms.Sinceourcurrentdatasource,theChinaStatisticalYearbook(2000–02),doesnotprovidethebreakdownsof‘‘self-fundraising,’’weonlyhavethetotal??guresinsubsequenttablesandgraphs.Thesizeoftotalself-fundraisingofall??rmsgrewatanaverageannualrateof14%overtheperiodof1994to2002.Attheendof2002,totalself-fundraising(for??xedassetinvestment)reachedUS$275.5billion,comparedtoatotalofUS$106.6billionfordomesticbankloansforthesameyear.ItisimportanttopointoutthatequityARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11680andbondissuance,whichareincludedinself-fundraising,applyonlytotheListed Sector,andaccountforasmallfractionofthiscategory.Moreover,self-fundraisingisthemostimportantsourceof??nancingformanytypesof??rms.Forexample,individuallyowned??rms(PrivateSector),notsurprisingly,relymostlyonself-fundraising(about90%oftotal??nancing).Interestingly,evenforstate-orquasi-state-ownedcompanies,self-fundraisingisalsoimportantinthatitcapturessomewherebetween45%and65%oftotal??nancing.Statebudgetandforeign(direct)investmentaretheothertwoimportant??nancingsources.Aswasthecaseforallsocialistcountries,Chinausedtorelyonacentralplanningsystemtoallocatethestatebudgettomostofthecompaniesinthecountry.Butthestatebudgetnowonlycontributes10%ofstate-ownedcompanies’totalfunding.Ontheotherhand,foreigninvestmentiscomparabletothestatebudget,bothintermsofaggregatesizeandintermsoftherelativeimportancein??rms’??nancing.Thisevidencecon??rmsthatChinahasevolvedfroma centrallyplanned,closedeconomytowardanopenmarketeconomy.Withtheknowledgeonthefour??nancingchannelsattheaggregatelevel,wenowfocusondifferenttypesof??rms’??nancingdecisions.Theresultsarepre-sentedinFigs.<2A–<>C.Inallofthese??gures,eachofthefourconnectedlinesrepresentstheimportanceofaparticular??nancingchanneloverthetimeperiod1994–2002,measuredbythepercentageof??rms’total??nancingcomingfromthischannel.First,Fig.<2AandBillustratehow??rmsintheListedSectorandtheStateSector??nancetheirinvestment(for??xedassets).WhiletheListedSectorhasbeengrowingfast,SOEsareonadownwardtrendasprivatizationofthese??rmsisstillinprogress.Around30%ofpubliclytradedcompanies’fundingcomesfrombankloans,andthisratiohasbeenverystabledespitethefastgrowthofthestockmarkets(Fig.<2A). Around45%oftheListedSector’stotalfundingcomesfromself-fundraising,includinginternal??nancingandproceedsfromequityandbondissuance.Moreover,Fig.2.(A)FinancingsourcesfortheListedSector;(B)??nancingsourcesfortheStateSector;(<>C)??nancingsourcesforthePrivateSector.A–<>Cexamine??nancingsources(fortheinvestmentof??xedassets)ofdifferenttypesof??rmsinChina.Inallthree??gures,eachofthefourconnectedlinesrepresentstheimportanceofaparticular??nancingchanneloverthetimeperiod1994to2002,measuredbythepercentageof??rms’total??nancingcomingfromthischannel.(A)presents??nancingsourcesfor??rmsintheListedSector(publiclylistedandtraded),(B)presentsresultsfor??rmsintheStateSector(state-owned??rms),while(<>C)presentsresultsfor??rmsinthePrivateSector(allother??rms). ARTICLEINPRESS0%10%20%30%40%50%60%199419951996199719981999200020012002CalendarYearPercen tageStateBudgetBankLoanForeignInvestmentSelf-fundrasing0%10%20%30%40%50%60%199419951996199719981999200020012002CalendarYearPe rcentageStateBudgetBankLoanForeignInvestmentSelf-fundraising0%10%20%30%40%50%60%70%199419951996199719981999200020012002 CalendarYearPercentageStateBudgetBankLoanForeignInvestmentSelf-fundraising(A)(B)(<>C)F.Allenetal./JournalofFinancialEconomics77(2005)57–11681ARTICLEINPRESS F.Allenetal./JournalofFinancialEconomics77(2005)57–11682equityandbondsales,whichrelyontheuseofexternalmarkets,onlyconstituteasmallfractionoftotalfundsraised,comparedtointernal??nancingandotherformsoffundraising.Combinedwiththefactthatself-fundraisingisalsothemostimportantsourceof??nancingfortheStateSector(Fig.2B),wecanconcludethatalternativechannelsof??nancingareimportantevenfortheStateandListedSectors.Next,weconsiderhow??rmsinthePrivateSectorraisefunds(Fig.2<>C).Self-fundraisinghereincludesallformsofinternal??nance,capitalraisedfromfamilyandfriendsofthefoundersandmanagers,andfundsraisedintheformofprivateequityandloans.Clearly,thiscategoryisbyfarthemostimportantsourceof??nancing,accountingforcloseto60%oftotalfundsraised.Moreover,since??rmsinthissectoroperateinanenvironmentwithpoorerlegaland??nancialmechanismsand regulationsthanthose??rmsintheStateandListedSectors,all??nancingsourcesprobablyworkdifferentlyfromhowtheyworkintheStateandListedSectors,andthoseindevelopedcountries.InSection5below,wepresentdetailedevidenceonhowdifferenttypesofself-fundraisinghelpPrivateSector??rmsatvariousstages.3.2.ComparingfinancingchannelsinemergingeconomiesWebrie??ycompare??nancingchannelsattheaggregatelevelinChinaandothermajoremergingeconomies.Inparticular,werelatetheaggregate??nancingchannelswiththegrowthoftheeconomyduringdifferentgrowthperiods,inordertodeterminewhethertheChineseexperiencein??nancingisunique.First,Fig.3Acomparesthedevelopmentofstockmarketsattheaggregatelevel,whileFig.3Dcomparesthegrowthratesof(PPP-adjusted)GDP.BothTaiwanandSouthKoreaexperiencedhighGDPgrowthinthe1970sandearly1980s,whilethetotalmarket capitalizationoftheirrespectivestockmarketsaccountedforlessthan20%oftheirGNPduringthesameperiod,andthegrowthofstockmarketsdidnottakeoffuntilthemid-tolate-1980s.Fig.3Bcomparesthegrowthofcorporatebondmarkets:SouthKoreahasthefastestgrowthpath,whileinTaiwanandChina,thecorporatebondmarketsseemtolagthedevelopmentofstockmarkets.Finally,Fig.3<>CcomparestotalequityissuanceincludingIPOsandSEOs.WiththeexceptionofSouthKorea,Chinaseemstobeonasimilarpaceintermsofsizeofequityissuance(asafractionofGNPinagivenyear)withTaiwan,India,andBrazil.FromtheabovecomparisonsitisclearthatthedevelopmentofChina’sexternalmarketsrelativetoitsoveralleconomicgrowthisnotdramaticallydifferentfromotheremergingcountries.Oneofthecommonpatternsisthatthedevelopmentofexternalmarketstrailsthatofthegrowthoftheoveralleconomy.Thisisnotsurprisinggiventhatthedevelopmentofthesemarketsrequiresa minimumef??ciencyforacountry’sinstitutionsincludingthelegalsystem,accountingstandards,andthedevelopmentofassociatedprofessionals.Bycontrast,duringearlystagesofeconomicgrowth,alternativeinstitutionsandmechanismsalonecansupportthegrowthof??rmsandtheoveralleconomy,asisthecaseforChinabasedonourevidence.Perhapssimilarinstitutionshaveworkedwellinotheremerginganddevelopedeconomies,andfutureresearchcandeterminewhethertheChineseexperiencealsooccursinothercountries.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11683180%Taiwan KoreaChinaIndiaBrazilIn30%TaiwanKoreaChinaIndonesiadonesia4.EvidenceontheListedSectorInthissection,wefocusonpubliclytradedcompaniesandexaminetheir??nancingandinvestmentdecisions.AsstatedintheIntroduction,wewanttodrawgeneralconclusionsonwhethertherearefundamentaldifferencesbetweentheChinese??rms0% 20%40%60%80%100%120%140%160%197019731976 19791982198519881991199419972000Year Market<>cap./GNP0%5%10%15%20%25%19 7019731976197919821985198819911994 19972000YearCorporatebond/GNP0%1% 2%3%4%197019731976197919821985 19881991199419972000YearIPO&SEO/GNP TaiwanKoreaChinaIndiaBrazil(A)(B)(<>C)-15%-10%-5%0%5%10%15% 20%25%1981198419871990199319961999 2002YearGDP(ppp)growthrateBrazilChinaIndiaIndonesiaKorea Taiwan(D)Fig.3.Comparing??nancingchannelsinemergingeconomies:(A)Stockmarketcapitalization;(B)corporatebondmarket;(<>C)equityissuance;(D)GDPgrowthrates.(A)comparesthetimeseriesofstockmarketcapitalization/GNPratiosacrosssixemergingeconomies,(B)presentsthetimeseriesoftheratiosoftheamountofcorporatebondsoutstanding/GNP,while(<>C)presentsthetimeseriesofIPOandSEOissuance(inagivenyear)/GNP.Thecalculationsforalltheratiosinthesethree??guresarebasedonlocalcurrenciesofacountryinagivenyear.(D)comparestimeseriesofthegrowthratesofGDP,andthegrowthratesarecalculatedusingPPP-adjustedGDP??guresinordertoavoidbiasescausedbydifferent currencypolicies.and??rmsstudiedinpreviouspapers(LLS,1999;LLSV,1997a,2000b,2002).Beforedoingso,we??rstlookattheuniqueownershipstructureandcorporategovernancemechanismsinChinese??rms.4.1.Typesofstocks,ownershipstructure,andcorporategovernanceListed??rmsinChinaissuebothtradableandnontradableshares(Table7A).Thenontradablesharesareeitherheldbythestate/governmentorbyotherlegalentities(i.e.,otherlistedornonlisted??rmsororganizations).Amongthetradableshares,ClassAandBsharesarelistedandtradedineithertheSHSEorSZSE,whileClassA(B)sharesareissuedtoChineseinvestors(foreigninvestorsincludingthosefromTaiwanandHongKong).Finally,ClassHsharescanbelistedandtradedontheHKSEandareissuedbyselected‘‘RedChip’’Chinesecompanies. Table7Bdemonstratesthatnontradablesharesconstituteamajorityofallsharesandmostofthesesharesareheldbythestate,whilethemajorityoftradablesharesARTICLEINPRESSTable7ATypesofcommonstockissuedinChinaTradableontheexchanges?De??nitionNo(Privateblocktransferpossible)State-ownedshares????Sharesthatarecontrolledbythecentralgovernmentduringtheprocessinwhich??rmsare convertedintoalimitedliabilityincorporationbutbeforetheyarelisted.AllthesesharesaremanagedandrepresentedbytheBureauofNationalAssetsManagement,whichalsoappointsboardmemberson??rms’boards.Entrepreneur’ssharesSharesreservedfor??rms’foundersduringthesameprocessdescribedabove;differentfromF.Allenetal./JournalofFinancialEconomics77(2005)57–11684sharesthatfounderscanpurchaseandsellinthemarkets.ForeignownersSharesownedbyforeignindustrialinvestorsduringthesameprocess.LegalentityholdersSharessoldtolegalidentities(suchasother companies,listedornon-listed)duringthesameprocess.EmployeesharesSharessoldto??rm’semployeesduringthesameprocess.Yes(Newissuedshares)AsharesChinesecompanieslistedinShanghaiorShenzhenStockExchanges,andsharessoldtoChinese(citizen)investors.BsharesChineseCompanylistedinSHSZorSZSE,butsharesaresoldtoforeigninvestors.HsharesChineseCompanylistedinHongKong(sharescanonlybetradedontheHKExchangebutcanbeheldbyanyone). ????Therearesubcategoriesunderthisde??nition.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11685Table7BTradablevs.nontradablesharesforChina’slistedcompaniesYearState/totalsharesNontradablea/totalsharesTradable/totalsharesA/totalsharesA/tradable shares??19920.410.690.310.160.5219930.490.720.280.160.5719940.430.670.330.210.6419950.390.640.360.210.6019960.350.650.350.220.6219970.320.650.350.230.66areAshares.Table7<>CprovidessomeevidenceontherelationbetweenownershipandcontroloftheBoardofDirectors.Informationprovidedhereisbasedonasurveyofcorporategovernancepracticesamong257companieslistedontheSHSEconductedin2000bytheResearchCenterofSHSE.ConsistentwithTables7Aand7Bandthe‘‘one-share,one-vote’’schemeadoptedby??rmsintheListedSector,stateandlegalpersonshareholdersappointmostoftheboardmembers,while theotherdirectorsareappointedbythegovernment.ThestandardcorporategovernancemechanismsarelimitedandweakintheListedSector(e.g.,SchipaniandLiu,2002).First,listed??rmsinChinahaveatwo-tierboardstructure:TheBoardofDirectorsandtheBoardofSupervisors.Thesupervisorsofalisted??rm,rankingabovethedirectors,areusuallyeitherof??cialsTable7<>COwnershipandcontrolinlisted??rmsofChinaCompanyownershipandcontrol(%)ShareholdertypeOwnershipControl(boardseats)State2421Legalperson4448Employees23 Tradableshares304Total10076Source:Table4.6,p.83,CorporategovernanceandenterprisereforminChina,buildingtheinstitutionsofamodernmarket,2002,WorldBankpublication.19980.340.660.340.240.7119990.360.650.350.260.7520000.390.640.360.280.8020010.390.640.360.290.802002n/a0.650.350.260.742003n/a0.640.350.270.76Source:ChinaSecurityRegulationCommitteeReports(2000)and//.csrc.govaNontradablesharesinclude‘‘state-owned’’and‘‘sharesowned bylegalentities’’.??TradablesharesincludeA,B,andHshares.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11686chosenfromgovernmentbranchesorexecutivesfromtheparentcompanies,whiletheBoardofDirectorsiscontrolledbythe??rm’sparentcompanies.Notalldirectorsareelectedbytheshareholders,andtherestarenominatedandappointedbythe??rm’sparentcompaniesandthenominationprocessisusuallykeptsecret(Table7<>C).Incentivepayisrarelyexplicitlyspeci??edinthedirectors’compensationpackages,whiletheconsumptionofperks,suchascompanycars,isprevalent.Theexternalgovernancemechanismsarealsoweak.First,theexistingownership structure,characterizedbycross-holdingsofsharesamonglistedcompaniesandinstitutions,makeshostiletakeoversvirtuallyimpossible.Secondly,institutionalinvestorsdonothaveastrongin??uenceonmanagementoronthestockmarket,astheyareaveryrecentadditiontothesetof??nancialinstitutionsinChina.Moreover,ineffectivebankruptcyimplementationmakesthethreatandpenaltyforbad??rmperformancenoncredible.TheWorldBank’scross-countryinformationontheef??ciencyofbankruptcyprocedures,whichisbasedonsurveysoflawyersandbankruptcyjudgesaroundtheworld,indicatesthatChina’s‘‘goalsofinsolvency’’indexisequaltothemedianofthesampleof108countries.Finally,thegovernmentplaysthedualrolesofregulatorandblockholderofmanylisted??rms.TheChinaSecuritiesRegulationCommittee(CSRC)isthecounterpartoftheSECintheU.S.,anditsmainroleistomonitorandregulatestockexchangesandlistedcompanies,whilethegovernmentexercisesshareholder controlrightsinlisted??rmsmainlythroughstate-ownedassetmanagementcompanies,whichholdlargefractionsofthestateshares.However,sincethetopof??cialsoftheseassetmanagementcompaniesareelectedbythegovernment,itisdoubtfulthattheydiligentlypursuetheir??duciaryroleascontrolshareholders.Moreover,thegovernment’sdualrolescanleadtocon??ictinggoalsindealingwithlisted??rms,whichinturnweakenstheeffectivenessofbothofitsroles.4.2.Evidenceonownership,financing,dividendandvaluationInthissection,weexamineandcomparevariouscharacteristicsoflisted??rmsinChinawiththoseofothercountries.OurresultsonChina’sListedSectorarebasedonasample(paneldata)ofmorethan1,100listed??rmsthatwecollectfromSHSE,SZSE,andthe‘‘AsiaEmergingMarketDatabase’’oftheTaiwanEconomicJournal,fortheperiod1992to2000.Table8Apresentsthesummarystatisticsfora ‘‘snapshot’’ofthesample??rmsattheendof2000.FromPanelA,theaveragemarketcapisUS$448million(medianisUS$355million),andtheaverageleverageratio,measuredbytheratiooflong-termdebtandcommonequity,is32%(medianis9%).Inshort,thesearelarge??rmsoperatinginvirtuallyallindustries.PanelBcompareslisted??rmsconvertedfromtheStateSectortothosenonstate??rms.First,80%ofthesampleoflisted??rmsusedtobestate-owned(921outof1163??rms).Second,thetwogroupsof??rmsaresimilarintermsofmostofthe??nancialratiosexceptforleverage:Firmsthatusedtobestate-ownedhavemuchhigherleveragethantheothergroup,partiallyduetothelargeamountofbankloansaccumulatedinthese??rmspriortotheirIPO.AR TICLEINPRESSTable8ASummarystatisticsoflisted??rms(inUS$millions)DatasourceforTables8A,8B,and8<>C(alsoempiricaltestsinAppendixB):??rmsarelistedinSHSEandSZSE(asofDecember2000).DataaredownloadedfromTaiwanEconomicJournal’s‘‘AsiaemergingmarketDatabase’’(//.tei>.tw/).PanelA:keyfinancialitemsandratios(wholesample)MeanMedianMinMaxStd.DevNumberofobsMarketcap.(US$mil)448.2354.90.08,190.2513.91174 LTdebt/commonequity0.30.10.06.90.6981Netincome99.6502.0??1,215.921,718.6721.0979EPS0.20.2??3.21.60.4979Proceedsfromstocksales163.60.0??290.829,379.2987.0975(272)aDividend50.818.40.08,106.0270.2979(617)aRetainedearnings26.433.2??2,125.72,210.18234.4979(951)aBondsissue0.80.00.0521.017.3975(6)aLongtermborrowing634.9233.10.0157,053.15,073.7974(895)aPanelB:listedfirmsconvertedfromSOEsvs.nonstatefirmsTypesoflisted??rmsandsamplesizeMarketCap.(US$mil.)Tobin’sQDividend/earningsDividend/netsalesL-TDebt/bookequityReturnonassetsPreviouslySOEs(921)490.620.500.480.060.350.028Previouslynon-SOEs(242)454.940.510.110.060.240.028Differenceinmeans(t-test)1.03??0.190.85??0.083.00??0.004aNumberofnonzeroobservations. ??Signi??cantat1%.F.Allenetal./Journalof FinancialEconomics77(2 005)57–11687Table8Bcomparestheownershipstructureofthese??rmstothosefromtheLLS(1999)sample,whichincludesover1,000listedcompaniesfrom33countries.The mainresultofLLS(1999)isthatcountriesthatprotectminorityshareholderspoorly(strongly)tendtohavemoreconcentrated(dispersed)ownership,asshowninthe??rsttwopanelsofTable8B.Theownershipstructureoflisted??rmsinChina,showninPanel<>CofTable8B,isconsistentwiththepredictionofBurkartetal.(2003),andclosertothatofotherAsian??rmsdocumentedinClaessensetal.(2000)thantotheLLS(1999)results.Thedominantownerof60%ofoursample??rmsisthe(central)government,whilefor13.6%of??rms,thedominantownerisfounders’families.Wealso??ndthatfor24.17%(1.83%)of??rms,thedominantshareholderisa??nancialARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11688company(anotherlisted??rm).Sincewedonothaveownershipdataforthis??nancialcompany(listed??rm),wedonotknowwhetherthiscompany(listed??rm)iswidelyheldornot.Butgiventhefactthatstateownershipisprevalentin listed??rmsandbanks,itisreasonabletoassumethattheyarenotwidelyheld.Finally,only0.44%ofall??rmsarewidelyheldsothatnoshareholderownsmorethan10%ofstocks.Table8<>Cprovidessomeevidenceon??nancingsourcesatthe??rmlevel.Theratiosforallthecountries(exceptforChina)inthetablearetakenfromLLSV(1997a).9TheevidenceinTable8<>Cisconsistentwithpreviousevidenceattheaggregatelevel:Intermsoftotalequity,thelistedChinesecompaniesdonotrelyonexternalmarketsasmuchastheircounterpartsinLLSVcountries,buttheydorelymoreheavilyondebt,andinparticularbankdebt,than??rmsinLLSV-samplecountries.Finally,weexaminedividendpoliciesandvaluationsoflisted??rmsinChina,andcomparetheseto??rmsstudiedbyLLSV(2000b,2002).Makingthemostoutoftheavailabledata,10weperformthreedifferentsetsofempiricaltests and??ndsimilarresults.DetaileddescriptionsofthesetestsarepresentedinAppendixB.First,LLSV(2000b)??ndthat??rmsincountrieswithpoorerprotectionofoutsideshareholderstendtohavelowerdividendratiosduetomoresevereagencyproblems.Usingthedividend-to-earningsratioasaproxyfordividendpolicy,we??ndthatonaverageChinese??rmstendtounderpaydividendstotheirshareholderscomparedto??rmsincountriesstudiedinLLSV(2000b).Second,LLSV(2002)??ndthat??rmsincountrieswithpoorerprotectionofoutsideshareholderstendtohavealowerTobin’sQ,measuredbythemarket-to-bookassetratios.WhenweexaminetheTobin’sQoflisted??rmsinChina,wecannotrejectthehypothesisthatonaveragetheirTobin’sQislowercomparedtocountrieswithbettershareholderprotection(LLSV,2002).9InLLSV(1997a),aratio(e.g.,marketcap/sales)foracountryisobtainedby??rst??ndingthemedianof thisratioacross??rmswithinvariousindustries,andthenbytakingtheaverageofthemediansacrossindustries.Asimilarprocedureistakento??ndtheratiosforChinausingourdatasetoflisted??rms.Finally,wetaketheaverage(median)ratiosacrossgroupsofcountriesaccordingtotheirlegalorigins,andcomparethemtothoseofChina.10Thedatasetsthatweemployinclude:(1)accountingand??nancialinformationfor1,100+listed??rmsfromChina(1990–2000);and,(2)LLSV(2000b,2002)resultsarebasedoninformationforover4,100??rmsfrom33countries(1989–1994),whiledetailed??rm-leveldataforLLSV-sample??rmsarenotavailabletous;however,wedohavetheircross-sectionalsummarystatisticsbycountry,aswellastheregressionresultsacrosscountries. ARTICLEINPRESSTable8BComparingownershipstructureoflisted??rmsPanelsAandBaretakenfromLLS(1999).InPanel<>C,the??rstrowistheaverageoftheAsiancountriesincludedinClaessensetal.(2000),excludingJapan.ThelastrowforChinaincludesoursampleof1,147listed??rms.CountryWidelyheldStateFamilyWidelyheld??nancialWidelyheldcorporationPanelA:LLS(1999)samplewithlargefirms Highantidirectoraverage34.1715.8330.425.05.83Lowantidirectoraverage16.0023.6738.3311.02.00Sampleaverage24.0020.1934.818.33.70PanelB:LLS(1999)samplewithmediumfirmsizeHighantidirectoraverage16.6710.3350.925.831.67Lowantidirectoraverage6.0020.8753.806.672.67Sampleaverage10.7416.1952.526.302.22Panel<>C:AsianfirmsAsia(noJapan,fromClaessensetal.,2000)3.099.3659.369.6618.55China(ourcalculations)0.446013.561.83??24.17????Notes:(1)‘‘Widelyheld’’??rmsarede??nedasnolargeshareholderholdsmorethan10%ofshares.‘‘State’’(‘‘family’’)??rmsarethosewiththecontrolling shareholderbeingthestate(afamily).‘‘Widelyheld??nancial’’(‘‘widelyheldcorporation’’)arethosewiththecontrollingshareholderbeingawidelyheld??nancialcompany(widelyheldcorporation).??FortheseChinese??rms,weidentifythedominantshareholdertobea??nancialcompany,butwearenotsurewhetherthe??nancialcompanyiswidelyheldornot.????FortheseChinese??rms,weidentifythedominantshareholdertobeanotherlistedandtradedcorporation,butwearenotsurewhetherthiscorporationiswidelyheldornot.F.Alleneta l./JournalofFinancialE conomics77(2005)57–1 1689ARTICLEINPRESSTable8<>CExternalfundingat??rmlevelF.Allenetal./JournalofFinancialEconomics77(2005)57–116905.EvidenceonthePrivateSectorInthissectionwestudyhow??rmsinthePrivateSectorraisefunds,theirvariousgrowthpaths,andthealternativemechanismsemployedbyownersthatcansubstituteforformalcorporategovernancemechanisms.Duetodatalimitations,muchofthisevidenceisbynecessityanecdotalorbysurvey.11Some ofourevidenceOverall,becauseinvestorprotectionisweak(andtheagencyproblemissevere)intheListedSectorinChina,boththedividendratioandTobin’sQarelowcomparedtosimilar??rmsoperatingincountrieswithstrongerinvestorprotection.Theseresultscon??rmthatLLSVpredictionsworkwellforChina’sListedSector,whichincludesmany??rmsconvertedfromtheStateSector,andisalsoconsistentwithevidencepresentedinFig.1.CountryEnglishoriginaverageFrenchoriginaverage GermanoriginaverageScandinavianoriginaverageLLSVsampleaverageChinaMarketcap/sales0.690.510.630.370.580.06Marketcap/cash??ow5.163.857.483.254.770.52Debt/sales0.260.270.30.280.270.67Debt/cash??ow2.012.063.182.422.245.34 Sources:LLSVcountries—WorldScopeandLLSV(1997a);dataforChinaisbasedonapanelof7,377??rm-year(1,174listed??rms,1992–2000)observations,witheachratiobeingthemeanofthepooledpanelof??rmsduringthesametimeperiod.coincideswiththeanecdotalevidenceinNaughton(1995),whileMcMillanandNaughton(1992)alsomakesimilarargumentsregardingtheroleofalternativemechanismsinsupportingthegrowthofnonstate??rmsinChina.Unliketheirwork,wealsoprovide??rm-levelsurveyevidence.We??rstpresentanecdotalevidenceon??rmsintwohighlysuccessfulregionsinSection5.1,thenpresentevidencebasedonasurveyof17??rmsinJiangsuandZhejiangprovincesinSection5.2.Finally,Section5.3providesdiscussionsofourevidence.5.1.AnecdotalevidenceintwosuccessfulregionsWenZhou.Wenzhou,acityintheZhejiangprovince,isthehomeofsome oftheearliestandmostsuccessful??rmsofthePrivateSector.Entrepreneursintheregion11All??rmsincludingPrivateSector??rmsmustdiscloseaccountingand??nancialinformationtothelocalBureauofCommerceandIndustry,andmostofthereportsareaudited.However,thesedataarethenaggregatedintotheStatisticalYearbookwithoutany??rm-levelpublications.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11691areknownfortheirkeenbusinesssenseandinnovation,aswellassharpmanagementskills(e.g.,McMillanandWoodruff,2002).Theyusuallystarttheirfamily-runbusinessesintownshipswithasimilarproductemphasis,inordertohave easyaccesstothenecessarytechnology,humancapital,andpotentialclientsandpartners.Thusweobservespecializationbyregions(e.g.,TownAproducesshoes,TownBradioparts,etc.).Thisspecializationcanbearesultoftheattemptof??rmstosignaltopotentialcustomersthattheyarecompetitivebylocatingthe??rminaregion??lledwithother??rmsproducingandsellingsimilarproducts.Duringrecentyears,certaindevelopedareashaveshiftedproductemphasisfromlabor-intensiveproductssuchasclothestomorehigh-techproductssuchascomputerparts.Thefailurerateforstart-upsinmostindustriesishigh.Newproductstrategiesoftenstartwithmimickingsuccessfulorpopularproducts.Patentlawsaredif??culttoimplement,andoftendisputesaresettledamongtheentrepreneursthemselves,similartotheevidencefoundinVietnambyMcMillanandWoodruff(1999b).Tobetterovercomethisproblem,someentrepreneursexpendeffortandmoneyto ensurethatthekeypartsoftheirnewproductsaredif??culttodisassembleandcopy.Anotherproductstrategyformanyentrepreneursisthattheyoftenaimat‘‘exporting’’theirproductstootherregions,includingforeigncountries,insteadofsellingthemlocally.Kunshan.KunshanCounty,whichisinJiangsuprovinceandisclosetoShanghai,isfamousforattractingforeigndirectinvestment,especiallyfromTaiwaneseinvestors.Someofthemosteffectivegovernmentpolicieshaveincludedsettingupspecialdevelopmentzoneswithfavorablelandandtaxpolicies.In1997,Kunshansetupahigh-techdevelopmentzone,inwhichenterprises,intheownershipformofjointventures,cooperatives,andventuressolelyownedbyforeigninvestors,cantakefulladvantageofataxwaiverandtaxreductionforinitialperiods.Firmswhosehigh-techproductsareexport-orientedcanenjoyevenmoretaxadvantages.Thereisalsoacenterinaspecialzoneestablishedbythelocalgovernment; thiscenteractsastheliaisonbetweenthelocalgovernment,entrepreneurs,andforeigninvestors,andastheregulatoraswellasserviceproviderforenterprisesoperatinginthezone.Enterprisesinthezonearerequiredtoreporttheiroperatingand??nancialinformationto,andareregulatedbythecenter,buttheyunderstandthatthecenterwillalmostneverinterferewiththeirinternaldecisions.Thecenter’sof??cialsaremainlyfromthelocalgovernment.Thehigh-techdevelopmentzonehasgrownveryfastsinceitsinceptionin1997.Duringtheearlystageoftheabovespecialzone,investorsfromTaiwanwerewillingtocommittheircapitaltothesestart-upsandre??nancethemwhennecessary.ThereasonthatmanyinvestorsarefromTaiwanisnocoincidence:ManypeopleinKunshanhaverelativesinTaiwanandthroughtheminvestorsobtainedinformationontheinvestmentopportunities.TheTaiwaneseinvestorsalsocametounderstand thatalthoughtherewerealmostnoformalinvestorprotections,localgovernmentof??cialshaveanincentivetocooperatewiththedevelopmentofthespecialzoneandtrytocreateaneconomicboominthelocaleconomy.Thisisthecasebecauseaboomingeconomycangreatlyenhancethechanceofanof??cialbeingpromoted,inadditiontoparticipatinginpro??t-sharing.Duringtheearlystageofdevelopment,Taiwaneseinvestorsdidnotstayintheareaastheyoftendonow.Asaresult,therewasvirtuallynomonitoringoftheentrepreneurs,andtherewascompleteseparationofownershipandcontrol.5.2.SurveyevidenceIndesigningoursurvey,wefollowGrahamandHarvey(2001,surveyofUSCFOs),Johnsonetal.(2002,surveyofEasternEuropecountries),and McMillanandWoodruff(1999a,bsurveysonVietnam).(Thesurveyquestionsandthetabulationofanswersareavailableat//2.bc.edu/??qianju/research.html.)AsTable9shows,amongthe17??rmsthatwesurveyedandwhichprovidedusdetailedanswerstoourquestions,one??rmisfromsuburbanShanghai,threearefromJiangsuprovince,andtheremainingthirteenarefromZhejiangprovince.ARTICLEINPRESSAgeofthe??rm11.43.0027.006.7F.Allenetal./JournalofFinancialEconomics77(2005)57–11692Noofemployees1634.390.005552.002107.8Size(totalassetsinmil.US$)55.30.60337.3082.7D=Eratio2.10.3814.953.4Netincome(inmil.US$)2.50.209.002.8Returnonassets0.10.000.340.1 These??rmsoperateinawiderangeofindustries.Theaverageageofthe??rmisover11years,andtheyemployanaverageofover1,600employees.Theaveragesizeof(book)assetsisUS$55million,withaveragereturnonassetsbeing10%.Finally,onaverage??rmsarehighlylevered,withtheaverage(privateandbank)debtto(private)equityratioreaching2.1.Fig.4Aprovidesmorebackgroundinformationforthesurvey??rms.Therearesigni??cantvariationsinthepastperformanceandtheexpectedfutureperformance(toptwohistograms)of??rms.Intermsofownershipstructure(secondpanelofhistograms),bothatstart-upandatthepresenttime,thetwodominantformsare‘‘founderandfamily’’,and‘‘shareholding’’,whichresemblesaprivateequitystructure.Around35%ofthefoundersofoursample??rmsworkedinTVEspriortostartinguptheirown??rms(bottomhistogram),while23%(18%)ofthefoundersworkedinSOEs(governmentagencies).TheexperiencefromtheState SectororotherPrivateSector??rmsisvaluablefortheentrepreneurs,astheynotonlygainedknowledgeonhowtorunaprivate??rm,butalsolearnedhowtodealwithgovernmentof??cials.Table9Summarystatisticsforsurvey??rms(asofDecember2002)Thesampleincludes17??rms:onefromShanghai,threefromJiangsuProvince,and13fromZhejiangProvince.Thesamplecovers??rmsintheindustryofchemicalproducts(3),fabricmakingandprinting(3),metalproducts(2),medicalandhealthproducts(2),realtymanagement(2),autorepairing(1),foodprocessing(1),agricultureproductprocessing(1),electronicproducts(1),andhandcraftandartproducts(1).Some??rmsareinmultiplebusinesslines. MeanMinMaxStd.DevARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11693Annualreturninthepast5years<00.1~0.20.2~0.3>0.320%30%40%Expectedannualreturninthenext3years <00~0.10.1~0.2>0.310%20%30%40%Financingchannels.Fig.4Bpresentsevidenceofthe??nancingchannelsofthe??rms.First,itisnotsurprisingthatduringthestart-upstage,fundsfromfounders’familyandfriendsareanimportantsourceof??nancing(top-lefthistogram).Moreover,fundsfromfriends,intheformofprivateloansandequity,arealsoveryimportantduringthe??rm’ssubsequentgrowthperiod(top-righthistogram).Insomecasestherearenoformalwrittencontractsbetweenthefriends/investorsandthe entrepreneurs,implyingthatreputation-andrelationship-basedimplicitcontractualFounder'spreviousexperienceSOECollectiveJVOthersGOVTTVEPrivateFirm0%5%10%15%20% 25%30%35%40%OwnershipwhenstartupPartnerTVEShare-holdingOthersGOVTForeignFounder/family 0%10%20%30%40%CurrentownershipPartnerForeignTVEShare-holdingOthersGOVTFounder/family0%10% 20%30%40%0~01.0%10%0.2~0.30%(A)Fig.4.(A)Backgroundinformationonsurvey??rms;(B)??nancingchannelsofsurvey??rms(shownonthenextpage);(<>C)governancemechanismsofsurvey??rms(shownonthenextpage).Inallofthehistogramsin(A),theverticalaxisrepresentsthepercentageof??rms’managers/founderswhoprovidethesame answerforaparticularquestioninthesurvey.Inthebottomthreehistogramsin(A),‘‘GOVT’’standsfor??rmsthathavelocalgovernmentasthemajorityowner;while‘‘JV’’standsforjointventures.In(B),eachbarrepresentsthepercentageof??rmsthatregardsa??nancingsourceasveryimportant(25–50%)orextremelyimportant(450%)duringtheirstart-upandgrowthperiods.Notes:PCA??privatecreditagencies;Budget??state/localbudget,andVC??venturecapital.Fig.4(<>C)presentsresultsonselectedgovernancemechanismsamongsurvey??rms.ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11694StartupfinancingBanks20% 30%40%50%Financingingrowthperiod(year3-8)FriendsPCATradeEthnicChinese20%30%agreementshaveworkedeffectively.Second,internal??nancing,intheformofretainedearnings,isalsoimportant(notreportedinFig.4B):Survey??rmsretainedanaverageof55%to65%oftheirnetincomeforreinvestmentduringtheinitialtwo tothreeyearsofexistence.Third,fundingfrom??nancialintermediariesisoneofthemostimportantsourcesforthesurveyed??rms.Intermsofstart-up??nancing,over40%of??rmssurveyedLeastcostlyfinancingsourcesStatebudgetFDIIPO,bondTradecreditsEthnicChinesePrivateequity/debtLTbankLoan STbankloanFamily&friends0%10%20%30%40%50%60%70%80%90%Percentageoff irmsOthersFDIFriendsFounderBudgetPCATradeCreditsVCEthnicChinese0%10%VC FDIOthersBudgetFounderBanksCredits0%10%(B)Relationshipwithlocalgovernment0%10%20% 30%40%50%Governmentinvests/ownspartiallySupports,returnisnotdemandedDemandsprofitsharingNorelationshipPossibilityofassetsbeingacquiredVerylikely NotpossiblePossible0%10%20%30%40%50%60%70%ConcernsfollowingthefailureoffirmLegal/SafetyconsequenceEconomiclossinvarioussenseReputationlossinvarious sense0%10%20%30%40%50%60%70%80%90%100%(<>C)Fig.4.(Continued)ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–116 95regard‘‘banks’’aseithera‘‘veryimportant’’(25–50%oftotalfundingneeds)oran‘‘extremelyimportant’’(morethan50%oftotalfundingneeds)??nancingsource.Thefourlargeststate-ownedbanksarerankedthehighestintermsofprovidingfunds,whileotherstate-ownedbanksarerankedsecond.However,itisnotclearthatstate-ownedbanksprovidethecheapeststart-up??nancingchannelforallPrivateSector??rms.Thecaveatisthatalmostallthesurveyed??rmsthatreceivedstart-up??nancingfromstate-ownedbankshadalreadyestablishedcloserelationshipswiththosebanksbeforetheirinceptionasshowninFig.4A.Infact,notasingle??rmratesbanksasveryimportantorextremelyimportantduringtheirgrowthperiod.Financingfromprivatecreditagencies(PCAs),insteadofbanks,isthemostimportantchannelduringa??rm’sgrowthperiod.Thesenonstatelendersusuallychargeveryhighinterestratesand/orrequirealargeamountofcollateralonloans, andcanforceliquidationshouldtheentrepreneursdefault;theassociatedloancontractsresemblejunkbondstoacertaindegree.Onaverage,eachsurveyed??rmcurrentlyhasaloanrelationshipwith4.3banksorother??nancialintermediaries,withthemaximum(minimum)being12(1).Collateralvaluecountsfor82.6%oftheloanvalueonaveragewithamaximum(minimum)of120%(20%).Fixedassetsarethemostpopularformofcollateral,withthird-partyguaranteesbeingthesecond-mostpopularform.Thesefactsimplythat??nancialinstitutions,stateorprivate,seemtounderstandtheriskofstart-up??rmsandtryto‘‘price’’thisriskintheirloancontracts.Inafewcasesthelocalgovernmentprovidesthethird-partyguarantee,indicatinganactiveroleplayedbygovernmentof??cialsinsupportingthegrowthof??rms.Duringa??rm’sgrowthperiod(Fig.4B),thereareafewotherchannelsthatareimportantsourcesof??nancing,inparticular,investmentfrom ‘‘ethnicChinese’’(investorsfromHongKong,Taiwan,andoverseasChinese),mostlyintheformofprivateloansandequity.This??nancingsource,ascomparedtoinvestmentfromnon-Chineseforeigndirectinvestment(FDI),reliesontherelationshipbetweentheinvestorsandtheentrepreneurs.Othersourcesincludetradecreditsamongbusinesspartners,stateandlocalbudgets,andFDIs,whileinvestmentfromventurecapitalists(VC)isnotwidelyusedduringeitherthestart-upstageorthegrowthperiod.Whenaskedaboutwhich??nancingchannelsareleastcostly(bottomhistograminFig.4B),whilemostofthesurveyed??rmspointtoshort-andlong-termbankloans,almost60%of??rmsindicatetradecreditsamongbusinesspartners(<>c.f.,McMillanandWoodruff,1999a).Forstart-up??rms,securinglandandother??xedassetsisimportantfortheirsurvival.Whilenotreportedinthe??gures,morethanhalfofthesurveyed??rms purchasethelong-term‘‘operation-rights’’oftheland(20–50years)fromthegovernment,whichhastheultimatecontrol.Withoperationrights,a??rmhasmorecontroloverthelandthanundera‘‘landrental’’contract.Forexample,??rmscanrentthelandtoanotherpartyonceobtainingtheoperationrightsfromthegovernment.Landrentalcontractshaveshortertermsonaverage(5–10years).Intermsof??xedassets,16outofthe17??rmspurchasedandownalloftheir??xedassets.Amongthem,nine??rmspurchasedtheir??xedassetsfromtheStateSector,andsevenoutofthenine??rmsconsideredthepricetheypaidtobethesameasthewerenotoperatingef??ciently,withinthreetosixmonths20%ofitsmarketsharewouldbetakenaway,while80%of??rms’founders/executivesbelievetheentire ARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11696marketshareofthe??rmwouldbetakenawayintwoyears.Whenaskedaboutwhattypeoflossesconcernthemthemostifthe??rmweretofail(top-righthistograminFig.4<>C),every??rm’sfounders/executives(100%)saidreputationlossisamajorconcern,whileonly60%ofthemsaideconomiclossesareofmajorconcern.Thesuccessofa??rminthePrivateSectordependscruciallyonthesupportfromlocalgovernment.Over40%ofsurvey??rmsstatethelocalgovernment‘‘supports’’thegrowthofthe??rmwithoutdemandingpro??tsharing,whileforsomeother??rms,thegovernmentiseitherapartialownerordemandspro??tsharingwithoutinvestinginthe??rm(bottomhistograminFig.4<>C).Thesupportiveattitudeofthelocalgovernmenttoward??rmsinthePrivateSectorisremarkableconsideringthefactthattheChinesegovernmentiswidelyregardedascorrupt(e.g.,Table 2<>C)anddisrespectfulofpropertyrights(e.g.,LLPS,2004).5.3.DiscussionInthissectionwediscussmechanismssupportingthegrowthofthePrivateSector.Webelievethemostimportantreasonforthegrowthistheworkofalternative??nancingandgovernancemechanisms.Perhapsthemostimportantmechanismisreputationandrelationships.Greif(1989,1993)arguesthatcertaintraders’organizationsintheeleventhcenturywereabletoovercomeproblemsofasymmetricinformationandthelackoflegalandcontractenforcementmechanisms,becausetheyhaddevelopedinstitutionsbasedonreputation,implicitcontractualrelations,andcoalitions.CertainaspectsofthegrowthoftheseinstitutionsresemblewhatworksinChina’sPrivateSectortoday,intermsofhow??rmsraisefundsandcontractwithinvestorsandbusinesspartners.Inaddition,Greif(1994)andStulzand Williamson(2003)pointouttheimportanceofculturalandreligiousbeliefsonthedevelopmentofinstitutions,legalorigin,andinvestorprotection.Gomes(2000)marketvalueoftheassets.One??rm’sexecutiveindicatedthatfortherentalportionof??xedassetsfromSOE,therearenoformalcontractsbetweenthe??rmandtheSOE.Finally,whenaskedabouttheprospectsofgoingpublic,foundersandexecutiveslist‘‘accesstolargescalefunding’’and‘‘reputationincrease’’asthemostimportantbene??ts,andthe‘‘disclosureofvaluableinformationtocompetitorsandoutsiders’’and‘‘largeamountoffeespaid’’tothegovernment,investmentbanks,andconsulting??rmsasthemostcriticaldisadvantagesofgoingpublic.Corporategovernance.Fig.4<>Cprovidessomeinformationongovernancemechanisms.First,over60%(30%)of??rmsbelievethatiftheirown??rmwerenotrunef??cientlyandwereto??nditselfin??nancialdistress, itis‘‘possible’’(‘‘verylikely’’)itsassetswouldbepurchasedbyanother??rmorinvestor;nooneanswereditis‘‘notpossible’’forthistooccur.Notreportedinthe??gure,wealsoasked??rmsaboutproductmarketcompetition:40%ofsurveyed??rmsbelievethatiftheir??rmdemonstratesthatamanagerialreputationeffectcanreplaceformalgovernanceinARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11697anIPO??rm,consistentwiththeevidencefromtheChineseventurecapitalindustry(e.g.,BrutonandAhlstrom,2002).TheabovefactorsareofparticularrelevanceandimportancetoChina’sinstitutionaldevelopment.Withoutadominantreligion,onecanarguethatthemostimportantforceshapingChina’ssocialvaluesandinstitutions isthesetofbeliefs??rstdevelopedandformalizedbyKongZi(Confucius).Thissetofbeliefsclearlyde??nesfamilyandsocialorders,andareverydifferentfromwesternbeliefsonhowlegalcodesshouldbeformulatedandhowindividualsandbusinessesnegotiate(e.g.,Pye,1982;Chow,2002).UsingtheWorldValuesSurveyconductedintheearly1990s,LLSV(1997b)??ndthatChinahasoneofthehighestlevelsofsocialtrustamongagroupof40developedanddevelopingcountries.WeinterprethighsocialtrustinChinaasbeingin??uencedbyConfucianbeliefs.Interestingly,thesamesurvey,usedinLLSV(1997b),??ndsthatChinesecitizenshavealowtendencytoparticipateincivilactivities.However,ourevidenceshowsthatwitheffectivealternativemechanismsinplace,citizensinthedevelopedregionsofChinahaveastrongincentivetoparticipateinbusiness/economicactivities.Thesecondmostimportantmechanismiscompetitioninproductandinput markets,whichhasworkedwellinbothdevelopedanddevelopingcountries(e.g.,McMillan,1995,1997;AllenandGale,2000b).WhatweseefromthesuccessofPrivateSector??rmsinWenZhouandothersurveyed??rmssuggeststhatitisonlythose??rmsthathavethestrongestcomparativeadvantageinanindustry(ofthearea)thatsurviveandthrive.DLLS(2002)examineentrybarriersacross85countriesincludingChina.EntrybarriersarearelevantfactorforthegrowthofChina’sPrivateSector,aslowerentrybarriersfostercompetition.DLLS??ndthatcountrieswithheavier(lighter)regulationofentryhavehighergovernmentcorruption(moredemocraticandlimitedgovernments)andlargerunof??cialeconomies.Withmuchlowerbarrierstoentrycomparedtoothercountrieswithsimilar(low)percapitaGDP,Chinaisonceagainan‘‘outlier’’intheDLLSsample.TheoutlierstatusisevenstrongerconsideringthatChinaisoneoftheleast democraticcountries,andsuchcountriestendtohavehighbarrierstoentry.Basedonoursurveyevidence,weconcludethatthereexistnon-standardmethodstoremoveentrybarriersinChina:First,16outofthe17??rmsappliedforalicense(required)beforethebusinessstarted,with50%ofthemindicatingthatittakestwoweekstoonemonthtogothroughtheprocedureand37.5%sayittakesonetotwomonths.Themainproblemfortheapplicationforalicenseseemstobedealingwithgovernmentbureaucracy.Toeasethisproblem,mostofthe??rms’founders/executivesaskthefriendsofgovernmentof??cialstonegotiateontheirbehalf,orthe??rmscanofferpro??tsharingtogovernmentof??cials.ButthesemethodsareconsistentwithourresultsthatalternativemechanismsbasedonreputationandrelationshipsprovidethemostimportantsupportforthegrowthofthePrivateSector.Thereareothereffectivecorporategovernancemechanisms.First,Burkartetal. (2003)linkthedegreeofseparationofownershipandcontroltodifferentlegalenvironments,andshowthatfamily-run??rmswillemergeasthedominantformofownershipstructureincountrieswithweakminorityshareholderprotection,whereasARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11698professionallymanaged??rmsmustbetheoptimalformincountrieswithstronginvestorprotection.OursurveyevidenceonthePrivateSectorandempiricalresultsontheListedSector,alongwithevidenceinClaessensetal.(2000,2002),suggeststhatfamily??rmsareanorminChinaandotherAsiancountries,andthese??rmshaveperformedwell.Second,AllenandGale(2000a)showthatifcooperationamongdifferentsuppliersofinputsisnecessaryandallsuppliersbene??tfromthe ??rmdoingwell,thenagoodequilibriumwithnoexternalgovernanceispossible,asinternal,mutualmonitoringcanensuretheoptimaloutcome.Wehaveshowntradecreditsareanimportantformof??nancingfor??rmsduringtheirgrowthperiod.Third,thecommongoalofsharinghighprospectivepro??tscanaligninterestsoflocalandforeigninvestorswithentrepreneursandmanagerstoovercomenumerousobstaclesandachievetheircommongoal.Underthiscommongoalinamultiperiodsetting,implicitcontractualagreementsandreputationcanactasenforcementmechanismstoensurethatallpartiesful??lltheirrolestomakethe??rmsuccessful.Pro??tsharingalsomakesitincentivecompatibleforof??cialsatvariouslevelstosupportthegrowthofthe??rm.Finally,thereisastrandofliteraturestudyingtransitionaleconomies,suchasRussia,China,Vietnam,andEasternEuropeancountries,fromSocialistsystemstomarketsystems.ItisimportanttopointoutwhyChinadiffersfrom othertransitionaleconomies.First,withtheexceptionofRussia,China’seconomyismuchlargerandmorediversi??edthanothertransitionaleconomies.Withasmallandhomogenouseconomy,acountrycanadjustitslegaland??nancialsystemstothestrengthsofitseconomymucheasierthanalargecountrycan.TherecenteconomicstruggleinRussiaillustratesthispoint(e.g.,ShleiferandTreisman,2000).ThesuccessofChina’sPrivateSectordemonstratesthatalternativemechanismscanworkwonderseveninlargeanddiversi??edeconomies.Second,itisprobablyeasierforothercountriestoadoptdrasticreformmeasuresintheshortrun.China,underthein??uenceofConfucius’views,isdifferentinthatpeopleholdthebeliefthatfundamentalchangesinsocietyshouldbegradualandshouldbefullyimplementedonlyaftertheyareprovencorrect.This<#00aa00'>view,however,doesnotpreventregionalexperimentsconductedata smallerscale.Accordingly,Chinaadoptedagradual,‘‘dualtrack’’pathinitseconomicreform,wherethecontinuedenforcementoftheexistingplanningsystemgoesalongsidewiththefast-paceddevelopmentof??nancialmarkets,ascomparedtothe‘‘bigbang’’approachtakenbysomeothercountries(e.g.,Lauetal.,2000).Third,theroleplayedbythegovernmentduringthereformprocessisverydifferentinChinathaninmostothertransitioneconomies,andinparticular,Russia(e.g.,BlanchardandShleifer,2001).Inabroadercontext,LLSV(1999)??ndthatgovernmentsincountrieswithFrenchorsocialistoriginshavelowerquality(intermsofsupportingeconomicgrowth)thanthosewithEnglishcommonlawsandrichercountries.However,ChinaisacounterexampletoLLSV’sargumentongovernment:WhiletheChineseCommunistPartylargelyremainsautocratic,governmentof??cials,especiallythoseinthemostdevelopedareas (e.g.,JiangsuandZhejiangprovinces),playedanactivesupportingroleinpromotingthegrowthofthethanthatoftheothersectors.ThesystemofalternativemechanismsandinstitutionsplaysanimportantroleinsupportingthegrowthinthePrivateSector,andtheyareARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–11699goodsubstitutesforstandardcorporategovernancemechanismsand??nancingchannels.Goingforward,ourresultsposeanimportantquestionforbothresearchersandpolicymakers:ShouldChinaalsotransformthePrivateSectortowardthe‘‘standardform’’likeithasbeendoingfortheStateSector?Giventhesuccess ofthePrivateSectorandthede??ciencyintheStateandListedsectorsinChina,muchmoreresearchisrequiredinordertobetterunderstandhowalternativemechanismsworkwherestandardmechanismsarenotavailableornotsuitable.TheseeffectivesubstitutesworkedwellinChina,andsimilarsubstitutesbasedonrelationshipandreputationmayhavealsoworkedwellinothereconomiesincludingdevelopedeconomies.Ourresultsthushavegeneralimplications:thereareimportantfactorsconnectinglaw,institutions,??nance,andgrowththatarenotwellunderstood.AbetterunderstandingofhowthesenonstandardmechanismsworktopromotegrowthcanshedlightonoptimaldevelopmentpathsnotonlyforChina,butalsoPrivateSector.Thisisdifferentfromthe‘‘grabbinghand’’roleplayedbygovernmentof??cialsinothercountries(FryeandShleifer,1997).Thereasonforthissupportingroleisthreefold.First,asLi(1998)pointsout,startingintheearly 1980s,thecentralgovernmentofChinaimplementedamandatoryretirementageforalmostallbureaucratsatvariouslevels,whichmadetheof??cialsyoungerandmorefamiliarwithcapitalistideas.InRussia,of??cialsfromtheoldregimewereentrenchedandabletoextractrentsfromtheneweconomywithoutanycontribution.Next,duringearlystagesofChina’sreform,TVEs,inwhichlocalgovernmentsarepartialowners,providedthemostimportantsourceofgrowthinthePrivateSector.TheenormoussuccessofTVEsandthepromotionoftheassociatedof??cialsprovidedexamplesandincentivestootherof??cialstofollowsuit.Finally,asdiscussedabove,pro??tsharingwith??rmsinamultiperiodsettingalsomakesitincentivecompatibleforof??cialsatvariouslevelstosupportthegrowthofthe??rm.6.ConcludingremarksInthispaperweexamineandcompareChina’sformalsystemsoflawand ??nanceandthealternativeinstitutionalarrangementsandgoverningmecha-nisms,andtherelationbetweenthedevelopmentofthesesystemsandChina’seconomicgrowth.Withoneofthelargestandfastestgrowingeconomiesintheworld,Chinadiffersfrommostofthecountriesstudiedinthelaw,institutions,??nance,andgrowthliterature,andisanimportantcounterexampletotheexisting??ndings:Itslegaland??nancialsystemsaswellasinstitutionsareallunderdeveloped,butitseconomyhasbeengrowingataveryfastrate.Moreimportantly,thegrowthinthePrivateSector,whereapplicablelegaland??nancialmechanismsarearguablypoorerthanthoseintheStateandListedsectors,ismuchfastermanyothercountries.AppendixA.BriefdescriptiourcesA.1.Creditor/ShareholderRD) LegaloriginReynoldsandFlores(1989),LLSV(1997a)ProxybymailallowedCompanylaworcommercialcodeproportionalcancastalltheirvotesforcommercialcodeARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–116100representationonecandidatetotheboardofdirectors(cumulativevoting)oramechanismofproportionalrepresentationintheboardbywhichminorityinterestsmayCumulativevotingorEqualsoneifshareholdersCompanylawordeposittheirsharespriortoageneralshareholders’meeting(topreventselling shares),andzerootherwise.Sharesnotblockedbeforemeetingnonvotingordinarysharesanddoesnotallow??rmstosetamaximumnumberofvotespershareholderirrespectiveofthenumberofsharesowned,andzerootherwise.Equalsoneifshareholderscanmailtheirproxyvotetothe??rm,andzerootherwise.Equalsonesif??rmscannotrequireshareholderstoCompanylaworcommercialcodeshare,andzerootherwise;(2)equalsone,whenthelawprohibitstheexistenceofbothmultiple-votingandOneshare–onevote Identi??esthelegaloriginofthecompanylaworcommercialcodeofeachcountry.(1)EqualsoneifordinarysharescarryonevoteperCompanylaworcommercialcodeVariablesonofourvariablesandtheirsightsVariables(Tables<2A–DescriptionSourcesOppressedminoritiesmechanismCompanylaworcommercialcode commercialcodePercentageofsharecapitaltocallanextraordinaryshareholders’meetingCompanylaworcommercialcodeAntidirectorrightsCompanylaworcommercialcodeARTICLEINPRESSF.Allenetal./JournalofFinancialEconomics77(2005)57–116101the??rm;(2)shareholdersarenotrequiredtodeposittheirsharespriortothegeneralshareholders’Equalsonewhengrantsshareholdersthe??rstopportunitytobuynewissuesofstock,andthis rightcanbewaivedonlybyashareholders’vote;equalszerootherwise.Theminimumpercentageofownershipofsharecapitalthatentitlesashareholdertocallforanextraordinaryshareholders’meeting;rangesfrom1%to33%.Theindexisformedbyaddingonewhen:(1)thecountryallowsshareholderstomailtheirproxyvotetomeeting;(3)cumulativePreemptiverightsmergersandassetdispositions);equalszerootherwise.Minorityshareholdersarede??nedasthoseshareholderswhoown10%ofsharesorless.Companylawormanagementorthe assemblyortherightto