《亚开行-中国的长期增长前景（英）-2022.12-67页 2023-1-7 154624 15.pdf》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。
TheLong-TermGrowthProspectsofthePeople’sRepublicofChinaEconomicgrowthinthePeople’sRepublicofChinahasmoderatedoverthepastdecade,evenbeforethecoronavirusdisease(COVID-19)pandemic.AcriticalquestionistowhichlevelofeconomicgrowththecountrywillreturntoafterCOVID-19and,morebroadly,howitslong-termgrowthprospectsare,givenarapidlyagingsociety,theongoinghighdependenceofgrowthoninvestment,andachangedinternationalenvironment.Thispaperestimatesthecountry’spotentialgrossdomesticproductgrowthupto2040basedonaCobb-Douglasproductionfunctionandprovidespolicyrecommendationsonstructuralreformstoboostpotentialgrowth.AbouttheAsianDevelopmentBankTHELONG-TERMGROWTHADBiscommittedtoachievingaprosperous,inclusive,resilient,andsustainableAsiaandthePacific,whilesustainingitseffortstoeradicateextremepoverty.Establishedin1966,itisownedby68members—49fromtheregion.Itsmaininstrumentsforhelpingitsdevelopingmembercountriesarepolicydialogue,PROSPECTSOFTHEPEOPLE’Sloans,equityinvestments,guarantees,grants,andtechnicalassistance.REPUBLICOFCHINADominikPeschelandWenyuLiuNO.54ADBEASTASIADecember2022WORKINGPAPERSERIESASIANDEVELOPMENTBANK6ADBAvenue,MandaluyongCity1550MetroManila,PhilippinesASIANDEVELOPMENTBANKwww.adb.org
每日免费获取报告1、每日微信群内分享7+最新重磅报告；2、每日分享当日华尔街日报、金融时报；3、每周分享经济学人4、行研报告均为公开版，权利归原作者所有，起点财经仅分发做内部学习。扫一扫二维码关注公号回复：研究报告加入“起点财经”微信群。。
ADBEastAsiaWorkingPaperSeriesTheLong-TermGrowthProspectsofthePeople’sRepublicofChinaDominikPeschelandWenyuLiuDominikPeschelisheadoftheEconomicsUnitattheAsianDevelopmentBank’sResidentMissionintheNo.54|December2022People’sRepublicofChina.WenyuLiuisaconsultantattheAsianDevelopmentBank’sResidentMissioninthePeople’sRepublicofChina.ASIANDEVELOPMENTBANK
CreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)©2022AsianDevelopmentBank6ADBAvenue,MandaluyongCity,1550MetroManila,PhilippinesTel+63286324444;Fax+63286362444www.adb.orgSomerightsreserved.Publishedin2022.PublicationStockNo.WPS220567-3DOI:http://dx.doi.org/10.22617/WPS220567-3Theviewsexpressedinthispublicationarethoseoftheauthorsanddonotnecessarilyreflecttheviewsandpoliciesof theAsianDevelopmentBank(ADB)oritsBoardofGovernorsorthegovernmentstheyrepresent.ADBdoesnotguaranteetheaccuracyofthedataincludedinthispublicationandacceptsnoresponsibilityforanyconsequenceoftheiruse.ThementionofspecificcompaniesorproductsofmanufacturersdoesnotimplythattheyareendorsedorrecommendedbyADBinpreferencetoothersofasimilarnaturethatarenotmentioned.Bymakinganydesignationoforreferencetoaparticularterritoryorgeographicarea,orbyusingtheterm“country”in thisdocument,ADBdoesnotintendtomakeanyjudgmentsastothelegalorotherstatusofanyterritoryorarea.ThisworkisavailableundertheCreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)https://creativecommons.org/licenses/by/3.0/igo/.Byusingthecontentofthispublication,youagreetobeboundby the termsofthislicense.Forattribution,translations,adaptations,andpermissions,pleasereadtheprovisionsand termsofuseathttps://www.adb.org/terms-use#openaccess.ThisCClicensedoesnotapplytonon-ADBcopyrightmaterialsinthispublication.Ifthematerialisattributedto anothersource,pleasecontactthecopyrightownerorpublisherofthatsourceforpermissiontoreproduceit.ADBcannotbeheldliableforanyclaimsthatariseasaresultofyouruseofthematerial.Pleasecontactpubsmarketing@adb.orgifyouhavequestionsorcommentswithrespecttocontent,orifyouwishto obtaincopyrightpermissionforyourintendedusethatdoesnotfallwithintheseterms,orforpermissiontousethe ADBlogo.CorrigendatoADBpublicationsmaybefoundathttp://www.adb.org/publications/corrigenda.Notes:Inthisreport,“$”referstoUnitedStatesdollars,and“CNY”referstoChineseyuan.ADBrecognizes“China”asthePeople’sRepublicofChina.TheADBEastAsiaWorkingPaperSeriesisaforumforstimulatingdiscussionandelicitingfeedbackonongoingandrecentlycompletedresearchandpolicystudiesundertakenbytheEastAsiaDepartmentoftheAsianDevelopmentBank(ADB)staff,consultants,orresourcepersons.Theseriesdealswithkeyeconomicanddevelopmentproblems,aswellasconceptual,analytical,ormethodologicalissuesrelatingtoproject/programeconomicanalysis,andstatisticaldataandmeasurement.TheseriesaimstoenhancetheknowledgeonAsia’sdevelopmentandpolicychallenges;strengthenanalyticalrigorandqualityofADB’scountrypartnershipstrategies,anditssubregionalandcountryoperations;andimprovethequalityandavailabilityofstatisticaldataanddevelopmentindicatorsformonitoringdevelopmenteffectiveness.TheADBEastAsiaWorkingPaperSeriesisaquick-disseminating,informalpublicationwhosetitlescouldsubsequentlyberevisedforpublicationasarticlesinprofessionaljournalsorchaptersinbooks.TheseriesismaintainedbytheEastAsiaDepartment.
CONTENTSTABLES,FIGURES,ANDBOXESivACKNOWLEDGMENTSviCURRENCYEQUIVALENTSviiABBREVIATIONSviiEXECUTIVESUMMARYviiiI. Introduction1II. DivergingViewsonthePRC’sLong-TermGrowth3A.ConditionalConvergence3B.ApproachestoLong-TermEconomicGrowthForecasts4III. ModelSpecificationandResults7A.FunctionalForm7B.Labor8C.HumanCapital9D.Capital10E.TotalFactorProductivity13F.Results18G.PotentialImpactofRecentDevelopments19IV. ReformstoBoostLong-TermGrowth23A.ReformAreas23B.Labor23C.HumanCapital24D.Capital26E.TotalFactorProductivity36V. OutlookonPolicyChallengesandReforms39Appendixes41Appendix1:Barro’sEstimateofConvergence41Appendix2:HumanCapitalIndex43Appendix3:ComparisonofResults45Appendix4:EstimatingCapitalStockGrowth46Appendix5:ImprovingtheQualityofEducation47Appendix6:SummaryofPolicyRecommendations48References50
TABLES,FIGURES,ANDBOXESTables1SelectedLong-TermGrowthForecastsforthePRC42ComparisonofSelectedLong-TermGrowthEstimatesforthePRC19A1Barro’sExplanationofthePRC’sPerCapitaGrowthPattern42A3.1ComparisonofPotentialGDPGrowthEstimateswithBailliuetal.(2016)45A3.2ContributionstoPotentialGDPGrowth,2020–204045A6SummaryofPolicyRecommendations48Figures1RealGDPGrowth,2000–202112RealGrowthinIncomeandConsumption,2000–202123GrowthinInvestmentbyCategory,2004–202124Japan’sandtheRepublicofKorea’sCatch-UpProcesstotheUnitedStates,1960s–201935PRC’sPerCapitaGDPinComparison46Long-TermGrowthForecasts,2013–203867InputFactorsoftheCobb-DouglasFunction78Working-AgePopulation,2000–205089ImpactofFertilityAssumptionsonWorking-AgePopulationProjectionsupto2050810TertiarySchoolEnrollment,2010–20401011HumanCapitalIndex,2000–20401012CapitalStockEstimateswithDifferentDeflators,2000–20191013EstimatedCapitalStockwithandwithoutHousing,2000–20161014GrossFixedCapitalFormation,1960–20211215CapitalFormationandLaborSupply,1960–20401216RealGrowthinFDIStock,2000–20401417RealImportGrowth,2001–20401418R&DExpenditure,2000–20401519EmploymentbySector,2000–20401520ShareofAgricultureinTotalEmployment1621TotalFactorProductivityGrowth,2001–20191722DecompositionofTFPGrowth,2001–204017
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina v23DecompositionofContributionstoPotentialGDPGrowth,2001–20401824BirthRate,1960–20212325FertilityRateforSelectedCountries,1960–20202326ComparisonofAverageYearsofSchooling,1950–20152527ComparisonofEducationAttainment,20152528YearsofSchooling,1985–20192629UrbanandRuralSchoolCompletionRates,2005,2015,and20192630IncrementalCapitalOutputRatio,2000–20192731FixedAssetInvestmentbyOwnershipType,2004–20172732EmploymentbyCategoryofEmployer,1980–20192833SOEAssets,andReturnonAssetsbySector,2003–20202834SelectedFinancingSources,2002–20213135ShadowBankFinancing,2002–20213136LoansOutstandingbyCategory,2009–20213237MaturityofFinancing,2013–20203238MarketSharebyTypeofBank,2003–20213239KeyInterestRates,2008–20223240HouseholdLoansbyType,2007–20223341CorporateDebt,2007–20213342TradeOpenness,1980–20213643ProcessingTrade,1993–20213644FDINetInflowintothePRC,2000–20203745FDIRestrictivenessIndex,2003,2006,2010,2015,and202037B1DebtStructure,2007–202221B2CreditGapandChangeinDebt,2008–202121B3PublicDebtandGrossRegionalProductPerCapitabyProvince,202122B4IndustrialStateHoldingEnterprises30B5SectorShareinGDP,1980–202135B6HouseholdIncomeandConsumption,andInvestment,1980–202135A1RealPerCapitaGDPGrowthinthePRC:Barro’sEstimateofConvergence,1960–202041Boxes1GDPGrowth,Investment,andDebt212IndustrialStateHoldingEnterprises303StructuralChange35
ACKNOWLEDGMENTSInsightfulcommentswereprovidedbyYolandaFernandezLommen,advisor,andPeterRosenkranz,financialsectorspecialist,EastAsiaRegionalDepartment,AsianDevelopmentBank(ADB);andMarcelSchroder,economist,EconomicResearchandRegionalCooperationDepartment,ADB.ValuableresearchassistancewasprovidedbyWenQi,associateeconomicsofficer,People’sRepublicofChinaResidentMission,ADB.
CURRENCYEQUIVALENTS(Asof1December2022)Currencyunit–CNYCNY1.00=$0.1410$1.00=CNY7.0924ABBREVIATIONSADB–AsianDevelopmentBankBIS–BankforInternationalSettlementsCOVID-19–coronavirusdiseaseFAI–fixedassetinvestmentFDI–foreigndirectinvestmentGDP–grossdomesticproductGFCF–grossfixedcapitalformationIMF–InternationalMonetaryFundICOR–incrementalcapitaloutputratioIP–intellectualpropertyNBS–NationalBureauofStatisticsOECD–OrganisationforEconomicCo-operationandDevelopmentPRC–People’sRepublicofChinaPPP–purchasingpowerparityR&D–researchanddevelopmentSOE–state-ownedenterpriseSME–smallandmedium-sizedenterpriseTFP–totalfactorproductivityUN–UnitedNationsUS–UnitedStates
EXECUTIVESUMMARYEconomicgrowthinthePeople’sRepublicofChina(PRC)hasmoderatedoverthepastdecade,evenbeforethecoronavirusdisease(COVID-19)pandemic.Acriticalquestionistowhichlevelofgrossdomesticproduct(GDP)growththecountrywillreturntoafterCOVID-19andhowitslong-termgrowthprospectsare,givenitsrapidlyagingsociety,acontinueddependenceofgrowthoninvestment,andachangedinternationalenvironment.Thispaperestimatesthecountry’spotentialGDPgrowthupto2040basedonaCobb-Douglasproductionfunction,estimatingallinputfactorsseparatelywithnewestavailabledata.Manyassumptionsinthepast,e.g.,onthedemographicdevelopment,haveoftenbeentoooptimistic,resultingininflatedgrowthestimates.Accordingtoourestimate,potentialGDPgrowthaverages5.3%in2020–2025,declininggraduallyto2.0%in2036–2040.Comparedtothreeearlierpapersthatalsousetheproductionfunctionapproach(eachusingdifferentspecifications)—Bailliuetal.(2016),RobertsandRussell(2019),andHiggins(2020)—ourestimateoffuturepotentialgrowthislowerthanBailliuetal.(2016)andclosertotheothertwo.Ourestimatesofthedifferentinputfactorsallowustoanalyzethedriversofthegrowthmoderationandtousetheseinsightsforpolicyrecommendations.Capitalandtotalfactorproductivity(TFP)areidentifiedasthemajorcontributorstogrowthinthefuture.Meanwhile,ashrinkingworking-agepopulationwillincreasinglyweighongrowth,whilethecontributionofhumancapitaltogrowthiscomparativelysmall.Toincreasepotentialgrowth,thispapersuggestsreformsinfourareas—labor,humancapital,capital,andTFP:Labor.Givenrapiddemographicaging,mitigatingtheimpactofthedeclininglaborforceiskey.Onemeasureistoraisetheretirementage,inparticulartheactualaverageretirementagethatisinthemid-50s.Othermeasurestocopewithanagingworkingforceincludeimprovinghealthcare,raisingfemaleworkforceparticipationrates,andincreasinglabormobility.Humancapital.ThePRChasroomtocatchupwithadvancedeconomiesinbothyearsofandqualityofeducation,especiallyinruralareas.Furthermore,educationcanbeimprovedbyexpandingearlychilddevelopment,enhancingthequalityoftertiaryeducation,strengtheningtechnicalandvocationaleducationandtraining,andincreasingon-the-jobtraining.Capital.Changestotheallocationofcapitalandcreditareconsideredtobeoneofthebiggestfactorstoraisepotentialgrowth.State-ownedenterprise(SOE)reformsareneeded.ThisrequiresclarifyingSOEs’scopeandfunction,levelingtheplayingfieldfortheprivatesector,separatingsocialfunctionsfromSOEs,andimprovingSOEmanagement.Moreover,creditallocationneedstobeshiftedinfavoroftheprivatesector,inparticularmicroandsmallbusinesses.Tothisend,theimplicitguaranteesofSOEsneedtoberemoved,whilebanksmuststrengthentheircreditriskassessmentcapabilitiesandincreasetheiroperationalefficiency.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina ixTotalfactorproductivity.First,tofacilitatethesectorallaborshift,residencepermitrestrictionsshouldbeloosenedtofacilitatelabormigrationtourbanareasandtheaccesstobasicsocialservicesshouldbemadeeasierforlabormigrants.Second,toincreasethecountry’sattractivenessforforeigndirectinvestment(FDI)inflows,thenegativelistforFDIshouldbeshortenedandjointventurerequirementsreduced.Third,advancingtradeandinvestmentagreementscouldhelpincreaseopennessandfacilitatedomesticreforms.Finally,toimprovethecountry’sresearchanddevelopment(R&D)effectiveness,theprevailingmisuseofR&Dfundsmustbeaddressed,theshareofbasicresearchincreased,thequalityofpatentsissuedraised,andintellectualproperty(IP)rightsprotectionstrengthened.
I. INTRODUCTIONThelong-termeconomicgrowthprospectsofthePRCmatter,notonlyforthecountrybutalsofortheworld.WhilesomeworryaboutthePRC“growingoldbeforebecomingrich,”othersseethePRCemergingastheworld’sbiggesteconomy.1AthirdgrouphighlightsadverseeffectsonglobalgrowthincasethePRC’seconomicgrowthslows(IMF2016;Ahmedetal.2019).However,thereisnoconsensusonthePRC’slong-termgrowthprospects.2Thispaperdiscussestheassumptionsandfindingsofdifferentapproachestolong-termgrowthforecastsforthePRC,estimatesthecountry’spotentialgrowthuntil2040,andprovidespolicyrecommendationstoraiseit.ThePRChashadalongstretchofhighgrowthfromtheendofthe1970suntil2010,whichisexceptionalininternationalperspective.3Atthesametime,domesticdebthasincreasedsharplyoverthepastdecade,andeconomicgrowthhasbeenonamoderatingtrend,evenbeforetheCOVID-19pandemichittheeconomy.4ThebigquestioniswhichgrowthratethePRCcansustainoverthenext2decadesandhowthedeclineinpotentialgrowthevolves.Answerstothesequestionswouldbehelpfulforeconomicpolicymakingandsocialplanning.Figure1:RealGDPGrowth,2000–2021 GDP GDPpercapita changeyearonyear GDP=growthdomesticproduct.Note:Inrecentyears,thepopulationsizevirtuallystagnatedsothatGDPgrowthandpercapitaGDPgrowthaligned.Source:NationalBureauofStatistics(NBS)viaCEICDataCompany(accessed10June2022).Economicgrowthmoderatedinthepastdecade,wellbeforetheCOVID-19pandemicin2020(Figure1).Afteryearsoftwo-digitgrowthrates,GDPgrowthslowedfrom9.0%in2011to6.0%in2019inlinewithpotentialgrowthmoderating(asourestimatesinpartIIIshow).Onthedemandside,growthinbothhouseholdincomeandconsumptionhassoftenedinthedecadebeforeCOVID-19,whileinvestmentgrowthhasbeenonadecliningtrend(Figures2and3).1SeeforinstanceEconomist(2019)asanexampleforthediscussiononaging,andZhuandOrlik(2022)onthePRC’schangestoemergeastheworld’sbiggesteconomy.2Foreconomists,usuallynottheabsolutesizeofaneconomymatters,butGDPpercapita.Thus,theimportantquestionisnotifthePRCtakesoverastheworld’sbiggesteconomyinthecomingdecades,buthowfastGDPpercapita—ameasureforthestageofdevelopmentofaneconomywhenadjustedforpurchasingpowerparity(PPP)—evolves.Dollar,Huang,andYao(2020)forecastthePRC’spercapitaGDP(PPPadjusted)relativetotheUStorisefromaboutone-quarterin2018toaroundtwo-thirdsin2049.3SeeforinstanceSummersandPritchett(2014).JohnsonandPapageorgiou(2020)provideatableofthefastest-growingeconomiesperdecade.ThePRCcomesfirstforthe1980sand1990s,andthirdforthe2000s(followingtwocommodityexporters).4Inthispaper,projectionsstartin2020toavoiddistortionsfromtheCOVID-19pandemic.
2 ADBEastAsiaWorkingPaperSeriesNo.54Figure2:RealGrowthinIncomeandFigure3:GrowthinInvestmentbyConsumption,2000–2021Category,2004–2021IncomepercapitaGFCFConsumptionpercapita FAI Manufacturing
Infrastructure Realestate changeyearonyear changeyearonyear - - -
FAI=fixedassetinvestment,GFCF=grossfixedcapitalformation.Note:GrowthratesinGFCFandFAIdiffer,amongothers,duetodifferentscopeofdata.AbreakdownbysectorisonlyavailableforFAI.Sources:NBSviaCEICDataCompany(accessed27MaySources:NBSviaCEICDataCompany(accessed26May2022);andADBcalculations.2022);andADBcalculations.Understandingthecountry’sfuturepotentialgrowthiscriticalfordesigningeconomicpolicies.Potentialgrowthisatheoreticalconstruct,anestimateoftheoutputthattheeconomywouldhaveproducediflaborandcapitalhadbeenemployedattheirmaximumsustainablerates,i.e.,ratesthatareconsistentwithsteadygrowthandstableinflation(Powell,Sheiner,andWessel2021).5PartIIwilldiscussdifferentapproachestoestimatinglong-termgrowth.Theproductionfunctionapproach,themethodwechoseforourmodelinpartIII,firstestimatesinputfactors—labor,humancapital,capitalstock,andtotalfactorproductivity(TFP)—andthencombinesthemintooneproductionfunction.Whilethisapproachyieldsabetterunderstandingofthedriversofpotentialgrowth,itrequiresestimatingthetrendsinmanyvariables,whichcanbeachallenge(ArsovandWatson2019).6Lardy(2019)pointsoutthechallengesasfollows:(i)itisextremelydataintensiveandrequiresinnumerablejudgementandassumptions;(ii)itrequiresreasonableaccuratehistoricaldata(onthecapitalstock,laborforce,andhumancapital);and(iii)projectionsrequiremanyjudgements,withcapitalstockgrowthandimprovementsinhumancapitalbeingdifficulttoestimate,andfutureTFPgrowthbeinganevengreaterchallenge.Ourestimateofpotentialgrowthindicatesthat,GDPgrowthwilldeclinenotablywithinthisdecadewithoutmajorreforms,whichareneededinparticulartothecountry’scapitalandcreditallocation,asaremeasurestoaddresstheshrinkinglaborforce.Furthermore,TFPgrowthwillbemuchlowerinfuture,callingforimprovementsinR&Defficiency.Theremainderofthispaperisorganizedasfollows:PartIIdiscussesthedivergingviewsonthePRC’slong-termeconomicgrowthpotential,includingdifferentapproachestoforecastinglong-termgrowth.PartIIIspecifiesaproductionfunctionandestimatespotentialgrowthfor2020–2040.PartIVdevelopspolicyrecommendationsbasedonthefindingsinthepreviouspart.PartVconcludeswithanoutlookonpolicychallengesandreformstoboostpotentialgrowth.5ForthePRC,suchapproachfaceslimitationsgiventhedriversofinflationinthePRC—adetaileddiscussioncanbefoundinADB(2021a)—andthefactthatlabormarketstatisticshaveshortcomings(ADB2020a).6Fordifferentwaystoestimatepotentialgrowth,seeGuisinger,Owyang,andShell(2018)fortheUS.
II. DIVERGINGVIEWSONTHEPRC’SLONG-TERMGROWTHA.ConditionalConvergenceTheneoclassicalgrowthmodelpredictspoorcountriestogrowfasterthanrichones,leadingtoaconvergencetowardrichcountries’levelsofpercapitaGDPovertime.Inmodernformsofthemodel,theconvergencearisesduetodiminishingreturnstoinvestmentsinphysicalandhumancapital,andpoorcountries’abilitytoimitateadvancedcountries’technologicalinnovations.However,thisconvergenceholdsempiricallyonlyconditionally,i.e.,poorcountriestendtogrowrapidlyonlyiftheyfirstcreatereasonablywell-functioninglegal,political,andmarketinstitutions(Barro2017).7Barro(2016)predictsconvergencetoprevailinthelongrunandagrowthrateof3%–4%overthenext2–3decades.ForhismodelofthePRC’spastgrowthratesandhisexplanationofthecountry’shistoricalgrowthpattern,seeAppendix1.InEastAsia,JapanandtheRepublicofKoreainitiallycaughtuprapidlytotheUS,beforetheprocessslowedaftertheirgrowthpeakedin1969and1995,respectively(Figure4).Figure4:Japan’sandtheRepublicofKorea’sCatch-UpProcesstotheUnitedStates,1960s–2019Japan’sPerCapitaGDPasShareoftheUnitedRepublicofKorea’sPerCapitaGDPasShareoftheStates,1960–2019UnitedStates,1963–2019 GDPpercapita(changeyearonyear) GDPpercapita(changeyearonyear) - GDPpercapitaasshareoftheUS()GDPpercapitaasshareoftheUS()GDP=grossdomesticproduct,US=UnitedStates.Note:GrowthinpercapitaGDP(at2017$,purchasingpowerparityadjusted)isa10-yearmovingaverage.Reddotsindicatetherisinggrowthratesandbluedotsdecliningones.Sources:JiangandYi(2015);PennWorldTable10.0;andADBcalculations.ThankstofasterGDPgrowth,thePRCcaughtupandnarrowedthedistanceintermsofGDPpercapita(adjustedforpurchasingpowerparity)toJapanandtheRepublicofKoreaovertime,especiallyinthepast2decades,aswellastotheUS(Figure5).7Forarecentdiscussion,seeJohnsonandPapageorgiou(2020),andspecificallyonthePRC,Barro(2016).
4 ADBEastAsiaWorkingPaperSeriesNo.54Figure5:PRC’sPerCapitaGDPinComparisonComparedtoJapanandRepublicofKorea,AsShareofUSPerCapitaGDP,1980–20211962–2019PRC Japan RepublicofKorea PPP( )GDPpercapita( changeyearonyear) - GDPpercapitaasshareoftheUS( )PRC=People’sRepublicofChina,PPP=purchasingpowerGDP=grossdomesticproduct,US=UnitedStates.parity.Notes:GrowthinpercapitaGDP(at2017$PPP)isa10-yearEstimatesstartafter2020forthePRCandtheRepublicofmovingaverage.ReddotsindicaterisinggrowthratesandKorea,and2015forJapan.bluedotsdecliningones.GDPgrowthratesinPennWorldTabledifferfromofficialones.Source:IMF,WorldEconomicOutlookDatabase2022.Sources:JiangandYi(2015);PennWorldTable10.0;andADBcalculations.B.ApproachestoLong-TermEconomicGrowthForecastsThereisarangeofopinionsonthePRC’slong-termgrowth.Table1summarizesaselectionofforecastsfromdifferentyears,usingdifferentapproachesandtimeframes.Table1:SelectedLong-TermGrowthForecastsforthePRCStudyApproachYearsResultSummersandPresentsbasicevidenceonregression2013–2033Usingregressionsanalysis,theauthorspredictPritchett(2014)tothemeanincountrygrowthrates.GDPgrowthpercapitatobe3.9%onaverageThenusesregressionsthatpredictovera20-yearhorizon.Fora10-yearhorizoncountries’growthratesbasedontheir(differentregression),theypredict5.0%inpastgrowth2013–2023and3.3%in2023–2033.Barro(2016)Conductsconditionalconvergenceand2015–2035GDPgrowthpercapitaisestimatedat3%–4%cross-countrygrowthregressionstooverthenext2to3decadesdetermineeconomicgrowthBailliuetal.UsesaCobb-Douglasproduction2015–2030GDPgrowthisforecasttodeclinegradually(2016)functiontoforecastthePRC’spotentialover15yearstoaround5%growthuntil2030Zhuetal.(2019)Predictspotentialgrowthbyexamining2018–2030Potentialgrowthinthebaselinescenariowillpastsectoraltransitionsslow,butremainrobustataround4%by2030RobertsandRus-Extrapolatesgrowthtrendsininput2020–2030ThePRC’sGDPgrowthisestimatedtoaroundsell(2019)factors(labor,capital,andTFP)4%–5%in2020-2030Higgins(2020)Growthaccountingofcontributions2018–2038Inthe“prettygood”scenario,thePRC’srealfromlaborforcegrowth,capitalpercapitaincomegrowthaverages3.8%inaccumulation,andTFP2018–2028,slowingto2.1%in2028–2038
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 5StudyApproachYearsResultSasakietal.FollowsapproachinZhu,Zhang,and2020–2035ThesizeofthePRC’seconomycanpotentially(2021)Peng(2019),butadoptsdifferentdoubleby2035(implyinganaveragegrowthindustry-levelconvergenceratesratherrateofabout4.8%until2035)—aslongasthethanafixedratecountryfollowsthecatch-upprocessachievedbyotherEastAsianeconomies.GDP=grossdomesticproduct,PRC=People’sRepublicofChina,TFP=totalfactorproductivity.Note:Onlystudiesthatarefurtherdiscussedinthispaper.Source:Authors’compilation.TherearethreemainapproachesintheliteraturecitedinTable1:8(i)Conditionalconvergence:Poorcountries—oncetheyhavecreatedreasonablywell-functioninglegal,political,andmarketinstitutions—tendtogrowfasterthanrichones,leadingtoaconvergencetowardrichcountries’levelsofpercapitaGDPovertime(Barro2017).Forinstance,Barro(2016)usesacross-countrydatasetonconvergence,takingintoaccountseveralfactors,toforecastthePRC’sgrowthrate,andconcludes3%–4%growthinthenext2to3decades.SummersandPritchett(2014)forecast3.9%growthoverthenext2decades.Atthesametime,theconvergenceapproachcanonlyprovidearoughestimateofpotentialgrowth(MaliszewskiandZhang2015).MaliszewskiandZhang(2015)furtherclaimthattheconvergencepathofthePRCmightbedifferentduetothesizeoftheeconomy.Zhu,Zhang,andPeng(2019)arguethatitisunlikelythatthePRCwillfollowtheEastAsianeconomies’growthpathsduetothechangingglobalenvironmentandthesizeofitseconomy.However,itseemsunlikelythatinthelongrunacountrycansustainaneconomicgrowthrateexceedingallothercountries’experiencesandkeepconstantlygrowingatahighrate(SummerandPritchett2014;Barro2016).Hence,theconvergenceapproachshouldbeabletoprovideatleastaroughideaofhowthePRC’slong-rungrowthwillevolve.Also,otherEastAsianeconomies’growthpathmightbethebestcomparatorgiventheirgeographicalproximityandtheirsimilardevelopmentpatternviaindustrializationandexport-ledgrowthinearlierstages.Inthiscontext,Higgins(2020)pointsouttothesimilaritiesbetweenthePRCandJapanandtheAsianTigerswithrespecttothedevelopmentoftheshareinglobalmanufacturingexportsaswellassimilaritieswithrespecttotheirhighinvestmentshareinGDP.9(ii)Sector-basedapproach:Anotherapproach,usingsectoralcatch-up,deploysasimilarideaasconvergence:dividingtheeconomyintodifferentsectors,productivityofthosethatarefurtherfromtheinternationalfrontier(i.e.,frequentlythoseinadvancedeconomies)willgraduallycatchup.Zhu,Zhang,andPeng(2019)highlightthatsectoraltransitionisakeydriverofthePRC’spotentialgrowth.Sasakietal.(2021)followthisapproach,butusedifferentiatedindustry-levelconvergencerates.Bothpapersreachsimilarconclusions,thoughtheestimatebySasakietal.(2021)comesinslightlyhigher.Ascatch-upprocessesandtheirspeeddependonmanyvariables,projectionsbasedontheseprocessestendtobehighlyuncertain(Higgins2020).Also,asseenabove,resultsaresensitivetotheselectionoftheconvergenceratesindifferentsectors.8Anotherapproachtoestimatepotentialoutputusesstatisticaltechniquestodecomposeoutputintotrendandcyclicalcomponents,withthetrendcomponentassumedtobepotentialoutput(ArsovandWatson2019)—seeWorldBank(2019and2021)applyingthisapproachtothePRC’scase.GrowthaccountingusingaCobb-Douglasfunctionistheapproachmostlyusedtoestimatepotentialgrowthforadvancedeconomies(ArsovandWatson2019).9WewilldiscussthisinthepartIIIwhenlookingatcapitalformationandtheexpansionpathofthecapitalstock.
6 ADBEastAsiaWorkingPaperSeriesNo.54(iii)ExtendedSolowmodel10:Bailliuetal.(2016),RobertsandRussell(2019),andHiggins(2020)adoptaCobb-Douglasproductionfunctionwithlabor(qualityandquantity),capital,andTFPasinputfactorstoforecastlong-termgrowth.Thechallengeofthisapproachisthatitneedsseveralassumptionsonthetrendoftheinputvariables.Here,thethreepaperstakedifferentroutes:RobertsandRussell(2019)extrapolatetrendsbasedonthepast10years,Higginslooksatcomparatorgroupsanddevelopsdifferentscenarios,andBailliuetal.(2016)estimatethetrendofeachinputfactorseparately.InFigure6,weclusterthesevenselectedstudiesaccordingtotheapproachused.Figure6:Long-TermGrowthForecasts,2013–2038SummersandPritchett( )Barro( )Sasakietal( )Zhuetal( )Bailliuetal( )Higgins( )RobertsandRussell( )tPercen Note:Convergenceapproachinorange,sector-basedapproachinblue,andproductionfunctionapproachingreen.Barro(2016)providesarangeof3%–4%fortheentireperiod(splitintwoperiodsinthechartforbettercomparison).Sources:Authors’compilationfromSummersandPritchett(2014);Barro(2016);Bailliuetal.(2016);Zhu,Zhang,andPeng(2019);RobertsandRussell(2019);Higgins(2020);Sasakietal.(2021).Theresultsusingtheconvergenceapproach(SummersandPritchett2014;Barro2016)varyonlyslightlyforthefutureandarelowerthantheforecastsfromthesector-basedapproach(Zhu,Zhang,andPeng2019;Sasakietal.2021)thatarewithinarangeof4%–5%.However,estimatesbasedonaproductionfunctiondiffermorewidely,givendifferentassumptionsanddataavailabilityatdifferentpointsintime,allofwhichcaninfluencetheestimatesofinputfactors.TheCobb-Douglasproductionfunctionwarrantsfurtherresearchgiventhattheestimatesofthestudiescitedabovediffermostwidelyandgivenrecentdataupdates,suchasthe2020censusresults,updatedcapitalstockestimates(IMF2021a),andanewdatasetoneducationalattainment(BarroandLee2021).GiventhemoderationinactualGDPgrowthinrecentyears,evenbeforetheCOVID-19shockontheeconomy,itseemswarrantedtore-estimatingpotentialgrowthusingthesedataupdates(andothers).UsingtheapproachbyBailliuetal.(2016),wemodifyitasneededtointegratedataupdatesandnewerfindingsanddevelopments.10SeeMankiw,RomerandWeil(1992)foraversionwithhumancapital.
III. MODELSPECIFICATIONANDRESULTSA.FunctionalFormWeadoptaCobb-DouglasproductionfunctiontoestimateGDPgrowth:11(1),whereYisGDP,Aisproductivity,Kis(physical)capitalstock,Lislabor,hishumancapital,αistheshareofcapitalincomeinGDP,and(1-α)istheshareoflaborincomeinGDP.Bylog-linearization,weget(2),whereTFPistotalfactorproductivity.Then,takingthefirstdifferenceofequation(2),weobtaingrowthrates:(3).Equation(3)derivestheGDPgrowthrate.Toabstractfromthebusinesscycle,allfactorinputsareassessedattheirtrendlevel:(4),whereX*denotesthetrendlevelofvariableX.Figure7illustratestheinputfactorsofGDPgrowthinthemodel.Figure7:InputFactorsoftheCobb-DouglasFunctionGDPGrowthHumanInputFactorLaborCapitalTotalFactorProductivityCapitalΔWorking-AgeΔHumanCapitalΔCapitalStockSectoralΔFDIΔImportsΔR&DEstimatePopulationIndexLaborStockExpenditureShiftFDI=foreigndirectinvestment,GDP=grossdomesticproduct,R&D=researchanddevelopment.Source:Authors.11SeeforinstanceBailliuetal.(2016).
8 ADBEastAsiaWorkingPaperSeriesNo.54Inthefollowing,wediscusstheinputfactorsandtheirestimatesandforecasts.Challengesrelatetoestimatingthecapitalstockand(future)TFP.ApossibleapproachtoestimateTFP,butonlylookingbackward,istocalculatetheresidual,12i.e.,thedifferencebetweenactualeconomicgrowthandthepartofgrowthexplainedbycapital,labor,andhumancapitalinput.Conversely,weneedtouseacombinationofvariablesthatcanbeforecasttoobtainanestimateoffutureTFPgrowth.13B.LaborAsforthegrowthofworking-agepopulation,weassumeinlinewithBailliuetal.(2016)that—atfullemploymentandanunchangedlaborparticipationrate—thetrendlaborgrowthismeasuredbythegrowthrateoftheworking-agepopulation,asshownintheequation:(5).Weconsidertheworking-agepopulationasthegroupofpeoplebetweentheagesof15and59.Whilethereareplansinthe14thFive-YearPlantograduallyraisethelegalretirementage,theactualretirementageinthePRCisfarbelow60(Economist2021)sothatinourmodelanextensionofworklifebeyond60doesnotseemwarranted.ThePRC’sworking-agepopulationstartedtofallin2012.Fromitspeakin2011until2021,theworking-agepopulationdeclinedby42million.In2021,itaccountedfor64%oftotalpopulation,downfromanaverageof69%in2000–2010.Theratioisprojectedtofallto61%in2030andfurtherto52%in2050(Figure8).Figure8:Working-AgePopulation,Figure9:ImpactofFertilityAssumptions2000–2050onWorking-AgePopulationProjectionsupto2050Age to Age to Age andabove Actualworking-age population MillionMillionLowfertilityvariant Mediumfertilityvariant Highfertilityvariant f f f f f f f f f f f fNote:Nodataavailablebetween2000and2005.UntilNote:ActualfiguresinbluefromNBSuntil2020,forecast2021actualfiguresfromNBS,from2022basedonfromUN(2019),withvariantsoverlappinguntil2035.forecastfromUnitedNations(UN)(2019),lowfertilityNBSfigurestakeintoaccountthe2020Census(withitsvariant(usinglinearextrapolation).downwardrevisions).Sources:NBSviaCEICDataCompany(accessedSources:NBSviaCEICDataCompany(accessed22December2022);UN(2019),WorldPopulationDecember2022);andUN(2019),WorldPopulationProspects;andADBcalculations.Prospects.12So-calledSolowresidual,namedaftertheeconomistRobertSolow.13Here,weneedtobrieflyaddressconcernsaboutGDPgrowthdataitself.ThelatestWorldPennTablesdatasetshowslowerGDPgrowth(especiallyinrecentyears)thanofficialstatistics,whichweuse.Chenetal.(2019)findthatGDPgrowthrateswereoverstatedfor2008–2016.Ontheotherhand,Clark,Pinkovskiy,andSala-i-Martin(2020)findforupto2015thatofficialGDPgrowthfiguresmaybeunderstated.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 9ProjectionsfromUN(2019)showthatthePRC’sworking-agepopulationwilldeclineinall(i.e.,low,medium,andhigh)fertilityscenarios(Figure9).Inthelow-fertilityscenario,theworking-agepopulationwillcontinuetodeclinefastafter2035,whileintheothertwoscenariosthedeclinedecelerates.AccordingtodatafromNBS,thePRC’sbirthrate—thenumberoflivebirthsper1,000ofpopulationperyear—felltoahistoricallowof7.5in2021,downfrom14.6in2012,whichwasthehighestratereportedwithinthepast2decades.ThissharpdeclinehappeneddespiteattemptstoincreaseitbyadoptingtheTwo-ChildPolicyin2015andevenaThree-ChildPolicyin2021.Thus,weusethelowfertilityvariantfromtheUNforourestimates.C.HumanCapitalThequalitativecomponentofhumancapitalisdeterminedbyyearsofschoolingandreturnstoeducation.Foreducationalattainment,weusetheBarro-Leedataset(2021)endingin2015.14Weconstructfiveagegroups:15–24,25–34,35–44,45–54,and54–59.BasedonthePRC’seducationsystem,weassume6,12,and16yearsofeducationforthecompletionofprimary,secondary,andtertiaryeducation.InlinewithBailliuetal.(2016),weassumethosewhohaveattendedbutnotcompletedprimary,secondaryandtertiaryeducationreceived3,9,and14yearsofeducation,respectively(andtheassumedreturnfornoeducationiszero).Asforreturnstoeducation,weassumedifferentannualreturnstoeducationforprimary,secondaryandtertiaryeducationasfollows:7.69%,8.92%,and13.38%(Bailliuetal.2016),whichaverages10%.15Whilethereturns-to-educationapproachprovidesadetailedpicture,severalstepsareneededtoobtainthehumancapitalindexanditsforecast(Appendix2).16Onecriticalassumptionishighereducationattainmentratesandtheirdevelopment.Bailliuetal.(2016)forecastsecondaryandtertiaryschoolattainmentontherelationshipbetweenschoolenrolmentratesandpercapitaincomeofhigh-incomecountries.However,thePRC’stertiaryschoolenrolmentratehasgrownrapidlyfrom24.2%in2010to58.4%in2020(Figure10).High-incomecountriesachievedatertiaryschoolenrolmentrateofabout58%in2001,withGDPpercapitaclosedto$40,000(in2017$,PPP-adjusted),whilethePRC’spurchasing-power-adjustedGDPpercapitawasonly$16,000in2020—whenthePRCalreadyhadreached58%.Weaccountforthatinourestimate(seeAppendix2fordetails).ThehumancapitalindexisshowninFigure11.Thetrendoftheindexisrisingfrom2000to2040.Accordingtoourestimate,a11%averagereturntoeducation—insteadoftheassumed10%return—wouldincreasethehumancapitalindexby0.27onaverage.14Inthedataset,dataareavailableevery5years(i.e.,for2000,2005,2010,and2015)forfiveagegroups:15–24,25–34,35–44,45–54,and54–64.Gapsforyearsarefilledbylinearextrapolation.Furthermore,weapproximatevaluesfortheagegroup54–59asBarro-Lee(2021)extendedtheagegroupto64.15MaandIwasaki(2021)conductedameta-analysisof213studiesonreturnstoeducation,confirmingthatamajorityofpreviousstudiesreportedabout10%(whilesomemorerecentestimatesvarygreatly).16RobertsandRussell(2019)estimatetrendgrowthoflaborqualitybyadjustingforaverageyearsofschoolingandusethatestimateasthecombinedcontributionfromlaborandhumancapital.Higgins(2020)includesaveragehoursworkedandtotalemploymentaslaborinputsbesideshumancapital.
10 ADBEastAsiaWorkingPaperSeriesNo.54Figure10:TertiarySchoolEnrollment,Figure11:HumanCapitalIndex,2010–20402000–2040 HumanCapitalIndex PRC High-incomecountries Uppermiddle-incomecountries PRC=People’sRepublicofChinaSources:WorldBank,WorldDevelopmentIndicatorsSource:Authors’estimates.Database2022;andauthors’estimates.D.CapitalThecapitalstockanditsgrowthrateneedtobeestimatedfortheproductionfunction.Theso-calledperpetualinventorymethodisfrequentlyusedtodeterminethecapitalstockeachyear:capitalstockatastartingpointplusinvestment(eachyear)minusdepreciationofthecapitalstock.EstimatingthePRC’scapitalstockischallengingforthefollowingreasons:first,thecapitalstockmustbeestimatedforthestartingdate(norelatedstatisticsexist);second,estimatesofthedepreciationratedifferwidely—Herd(2020)providesanoverview.Finally,fixedassetinvestment(FAI)datafrequentlydiffersfromthatongrossfixedcapitalformation,whichcanbeonlypartlyexplainedbytheirdifferentmethodology(BrooksandBarnett2006).WeusethecapitalstockestimatebytheInternationalMonetaryFund(IMF),whichcoverstheperiod1960–2019.SincetheseestimatesareincurrentCNY,wehavetoaddresspriceeffects.Todoso,wedeflatethecapitalstockgrowthwiththeGDPdeflator(Figure12),whichresultsinanestimateclosetothatinBailliuetal.(2016)for2000–2014.UsingtheFAIindexasdeflatorwouldyieldahigherestimate—wefurtherdiscussdeflatorchoicesandcross-checksinAppendix4.Figure12:CapitalStockEstimateswithFigure13:EstimatedCapitalStockwithandDifferentDeflators,2000–2019withoutHousing,2000–2016UsingGDPdeflatorShareofhousingintotal(RHS) UsingFAIindex IncludinghousingExcludinghousing Percent Trillion CNY Trillion CNY FAI=fixedassetinvestment,GDP=grossdomesticNote:Inconstant2010prices;differentfromFigure12.product.Sources:IMFInvestmentandCapitalStockDatasetSources:Herd(2020)viaWorldBank’sDevelopment(2021);NBSviaCEICDataCompany(accessed26MayDataHub(https://datacatalog.worldbank.org/);and2022);andauthors’calculations.authors’calculations.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 11Anotherpotentialchallengecomesintobearingwhenseparatingthecapitalstockintohousingandnonhousing.Theargumentbehindthisisthathousingisnotproductivetoproducegoodsandservices,andhenceshouldbeexcludedfromthecapitalstock(Bailliuetal.2016).Forthatreason,Higgins(2020)usesasubsetfromPennWorldTablethatestimatestheflowofservicesfromthecapitalstock,butconcedesthattheresultswouldbebasicallythesameusingthecapitalstockseries.Weusethegrowthrateoftheentirecapitalstock(includinghousing)for2001–2019.Figure13showsHerd’s(2020)estimatesofcapitalstockwithandwithouthousingaswellastheshareofhousinginthecapitalstock,whichvariesovertime.InAppendix4,weprovideadiscussiononthatissue.17Furthermore,Box1willdiscussoverinvestment.Asmentioned,thefuturegrowthrateofthecapitalstockiskeyforourestimate—andthisrateisexpectedtodecline.In2001–2019,theGDP-deflatedcapitalstockgrewannuallyby10.4%onaverage.Inthefuture,thegrowthrateofthecapitalstockshoulddeclineasthecapitalstockagesanddepreciationincreases,whileinvestmentasashareinGDPisforecasttofall(Figure14).Inordertoestimatethefuturegrowthrateofthecapitalstock,weapplytwomethods.First,welookatthehistoricalinvestmentpatternofJapanandtheRepublicofKorea(Figures14and15)andestimatethePRC’sshareofinvestmentinGDPbasedonthedevelopmentpatternofthesetwocountries.Second,weforecastthePRC’sfuturecapital-laborratioinlinewiththepreviousdevelopmentofthecomparatorsandcomparetheresultswiththefirstmethodasacross-check.ShareofinvestmentinGDP.Thoughtheshareofgrossfixedcapitalformation(GFCF)inGDPhasbeenhigherinthePRCthanintheRepublicofKoreaandJapan,itstartedtodeclineafter2010,broadlyfollowingtheexperienceofthetwocomparators(Figure14).18TwentyyearsaftertheirrespectivepeakinJapanandtheRepublicofKorea,theshareofGFCFinGDPwasabout30%onaverageandonlyslightlylower—28%onaverage—30yearsafter.AccordingtoourprojectionforthePRC,theshareofGFCFinGDPwillreach32%by2030(i.e.,20yearsafterthepeak)and29%by2040(or30yearsafter),takingintoaccountthatitpeakedatahigherrateinthePRC(Figure14).19Thisimpliesanannualgrowthrateofthecapitalstockof4.2%onaveragein2020–2040.17Thecross-checksinAppendix4showthatthedeflatorchoicemattersforthecapitalstockgrowthrate,whiletheexclusionofhousinghardlyaffectstheestimateofthegrowthrate.18DollarandWei(2007)arguethatthePRC’scapitalallocationhasbeeninefficientandcouldbeimproved—wewilldiscussbothissuesinpartIV.19Backingoutthe(growthin)capitalstockfromaprojectionofthecapital-laborratio,Bailliuetal.(2016)forecastadeclineintheinvestmentratetoaround25%ofGDPby2030.Accordingtoourestimate,thedeclineuntil2030willbelesspronounced,asGFCFstillstoodat42%ofGDPin2020and2021.
12 ADBEastAsiaWorkingPaperSeriesNo.54Figure14:GrossFixedCapitalFormation,Figure15:CapitalFormationandLabor1960–2021Supply,1960–2040PRCGrossdomesticsavings(ofGDP)JapanGrossfixedcapitalformation(ofGDP)RepublicofKoreaWorking-agepopulation(oftotalpopulation) PRC Percent t Percen Japan t- t- t- t- t- t- t- t-t- t-Peaktt tt t t t t t t Percent RepublicofKorea Percent GDP=grossdomesticproduct,PRC=People’sRepublicGDP=grossdomesticproduct,PRC=People’sRepublicofChina.ofChina.Note:GrossfixedcapitalformationasashareinGDPNote:Working-agepopulationreferstoage15–59forallpeakedinJapanin1973,intheRepublicofKoreain1991,threecountries.GrossdomesticsavingsforJapanisonlyandinthePRCin2010.DataforthePRCstartsin1962availablefrom1970onward.(toexcludedistortionsfromtheGreatLeapForward).Sources:NBS,OECDandWorldBankviaCEICDataSources:NBSandOECDviaCEICDataCompanyCompany(accessed31August2022);UN(2019),World(accessed31August2022);andauthors’calculations.PopulationProspects(lowfertilityvariant);andauthors’calculations.Inthetwocomparatorcountries,thedeclineinGFCFasshareinGDPhappenedquickly.SuchadropinGFCF(asshareinGDP)hasnothappenedinthePRC.However,suchadjustmentmighthappeninthePRCinthefutureasaccumulateddebthasrisensharplyafter2008,forcingadjustmenttotheinvestment-ledgrowthmodelofthe2010s(Box1attheendofthispart).
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 13Capital-laborratio.Anotherwaytoestimatethefuturegrowthrateofthecapitalstockgoesviathecapital-laborratio.ThePRC’scurrentcapital-laborratioisclosetothatofJapanandhigherthanthatoftheRepublicofKoreaatthecomparableGDPpercapitalevel.Inordertoforecastthefuturecapitalstock,weuseJapan’sandtheRepublicofKorea’sexpansionpathsoftheircapital-laborratiosatsimilarGDPpercapitalevels,butassumeslightlyhigherratiosforthePRCgiventhehistoricalprecedent.Thisyieldsanannualcapitalstockgrowthof4.2%onaveragein2020–2040.Result.TheforecastofcapitalstockderivedfromtheshareofGFCFinGDPapproachandfromthecapital-laborratiobothyieldvirtuallythesameaveragegrowthrateofthecapitalstockfor2020–2040,onlythedynamicsovertimedifferslightly.ThecontributionstoGDPgrowthareinitiallyhigherwiththefirstmethod(GFCF)—whichweuse—buttheyalsodeclinefaster.E.TotalFactorProductivityToexplainthepasttrendofthecontributionoftotalfactorproductivity(TFP)togrowthandtoforecastTFPgrowthoverthenext2decades,wefollowBailliuetal.(2016)anddecomposeTFPimprovementintofourfactors:(i)foreigndirectinvestment(FDI),(ii)importgrowth,(iii)researchanddevelopment(R&D)expenditure,and(iv)sectorallaborshift.ForeigndirectinvestmentWelookatFDIasadriverofTFPgrowth.Lee,Lee,andKim(2011)foundapositivecorrelationbetweengrowthinFDIstockandTFPof0.0218,i.e.,a10%increaseintheFDIstockimprovesTFPbyabout0.2%.FollowingBailliuetal.(2016),wedeflatethenominalFDIstock,denominatedinUSdollar,bytheconsumerpriceindex(CPI)fortheUStoobtainanestimateofrealFDIstockgrowth.WeassumethesamecoefficientfromLee,Lee,andKim(2011)fortheforecastandassumethatgrowthinthePRC’sFDIstockwillslowgraduallytothesimpleaverageofthecurrentrealgrowthratesinJapan’sandtheRepublicofKorea’sFDIstock,whichaveragedcloseto5%in2015–2020.WeassumethePRC’srealgrowthinFDIstockwillslowfrom8.4%in2015–2020to5.0%by2030andthenmarginallyeasefurtherto4.0%in2040(Figure16).ImportgrowthWeusetheimportindexfromtheCPENetherlandsBureauforEconomicPolicyAnalysistogetthegrowthrateofimportsinrealterms.In2001–2019,annualrealimportgrowthaveraged3.8%(Figure17).20Itisforecasttodeclineto2.5%by2030andfurtherto2.0%by2040.Wefurtherassume,inlinewithLee,Lee,andKim(2011),thata1%increaseinrealimportgrowthraisesTFPby0.068percentagepoints.2120Theindexisbasedon2010prices.UsingtheGDPdeflatortocalculaterealimportgrowthin2003–2013,Bailliuetal.(2016)forecastaslightlyhighercontributionofrealimportgrowthtofutureTFPgrowththanwedo.21Weusetheregressionresultforthebiggestsample,i.e.,allcountriescoveredinLee,Lee,andKim(2011).
14 ADBEastAsiaWorkingPaperSeriesNo.54Figure16:RealGrowthinFDIStock,Figure17:RealImportGrowth,2000–20402001–2040 FDFDIstIstocockk ImportsImports AveragAverage – e – earonyearearonyear yy %change%change%changeyearonyear%changeyearonyear - - - - - - FDI=foreigndirectinvestment. Note:In2000$.Forecaststartsin2021.PossibleNote:Indexisbased2010=100.Forecaststartsin2020distortionsfromCOVID-19for2020donotaffectthetoavoiddistortionsfromCOVID-19,whichwouldbeassumedtrend.sizable.Sources:UnitedNationsConferenceonTradeandSources:CPENetherlandsBureauforEconomicPolicyDevelopmentviaCEICDataCompany(accessed26MayAnalysis,WorldTradeMonitor;andauthors’estimates.2022);andauthors’estimates.ResearchanddevelopmentR&DisthemajorcontributortoTFPimprovement.InthePRC,annualR&Dexpenditureincreasedrapidlyfromlessthan1%ofGDPintheearly2000sto2.2%ofGDPin2019.WeassumeannualR&Dtoreach2.5%ofGDPin2030,and,by2040,thecurrent2.6%-of-GDPlevelofhigh-incomecountries(Figure18).Weassumethata1%increaseinR&DspendingraisesTFPby0.4percentagepointsfor2000–2019and0.23pointsin2020–2040.TheelasticitychosenhasasubstantialeffectontheTFP’scontributiontogrowthbecauseR&Disitsbiggestcomponent.However,estimatesoftheelasticityofR&DexpenditureonGDPgrowthvarywidely.22For2000–2019,weuseacoefficientof0.4,whichliesinbetweenGriffithetal.(2004)andotherestimatesthataremuchlower.23For2020–2040,weassumeacoefficientof0.23,which—exceptforoneoutlier—isthehighestelasticityreportedinthecompilationbyMcMorrowandRoeger(2009).24Asaccesstotechnology,industry-specificknowledge,andsomemarketshasbecomemorerestrictiveinrecentyears,thereturntoR&Dshoulddeclinegivencomplementaritiesofinputsfortechnological22Griffithetal.(2004)foundacoefficient0.669onlaggedlevelofR&Dintensity,whichBailliuetal.(2016)uses.Haider,Kunst,andWirl(2020)refutethefindinginGriffithetal.(2004)thattheelasticityishigherforcountriesthatarefurtherawayfromtheinternationalfrontier.Also,thementionedelasticityofnearly0.67seemstobeveryhighandpertainsonlytothemanufacturingsector(McMorrowandRoeger2009;CBO2005).Otherestimatesaremuchlower,frequentlyroughlyranging0.1–0.2forstudiesusingeconomy-widedata(McMorrowandRoeger2009).23ThechoiceofelasticitydoesnotaffectpastTFPgrowth,whichisestimatedasasmoothedresidual.UsingalowercoefficientwouldonlyincreasetheunexplainedpartoftheTFPestimatefor2001–2019.24CBO(2005)statesthattheevidenceismixedonthequestionofwhetherthereturntoR&Dhaschangedovertime.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 15products.25Inaddition,top-downresearchdecisionsimposecostsonresearchquality(Acemoglu,Yang,andZhou2022).Figure18:R&DExpenditure,Figure19:EmploymentbySector,2000–20402000–2040PRCHigh-incomecountriesAgricultureIndustryServices GDP of otalemployment t of GDP=grossdomesticproduct,PRC=People’sRepublicofChina,R&D=researchanddevelopment.Sources:NationalBureauofStatisticsviaCEICDataSources:MinistryofHumanResourcesandSocialCompany;WorldBank,WorldDevelopmentIndicatorsSecurityviaCEICDataCompany(accessed28May2022;(bothaccessed28May2022);andauthors’estimates.andauthors’estimates.SectorallaborshiftTheshareofemploymentinprimaryindustryhasbeenfallingfrom50.0%in2000to22.9%in2021.Meanwhile,theshareofemploymentintertiaryindustryrosefrom27.5%to48.0%duringthesameperiod(Figure19).WeestimatethehistoricalcontributionofsectorallaborshiftsbyconductingacounterfactualscenariointwostepsfollowingBailliuetal.(2016).First,wederivetheproductivityforprimary,secondary,andtertiaryindustriesfromGDPandemploymentofeachsector.Second,weconductacounterfactualscenariobykeepingtheemploymentsharesforallthreesectorsunchangedat2000levelswhileapplyingtheproductivityderivedinthefirststep.WethencalculateGDPgrowthintheactualandthecounterfactualscenario.Thedifference(actualgrowthminuscounterfactual)istheestimatedcontributionofsectorallaborshifttoTFPgrowth.Toestimatethefuturecontributionofthesectorallaborshifttogrowth,weneedtoforecastthedevelopmentofeachsector’semploymentshare.Todoso,wedrawontheexperienceofselectedcountries,includingJapan,theRepublicofKorea,andBraziltoforecasttheshareofagricultureinemployment(Figure20).26WithrisingGDPpercapita,theagriculturalshareofemploymenttendstofall.DatafromselectedcountriessuggestthatwhenGDPpercapitareaches$30,000(in2017$PPP),theagriculturalshareofemploymentisaround7%.However,asthePRCstillhasalargershareofworkforceintheprimarysectorthanthesecountries,thedeclineintheshareofemploymentinagriculturalwilllikelybemoregradually,fallingto15%in2030andfurtherto11%in2040accordingtoourestimates.25PRCcompaniesprefertofocusondevelopingnewtechnologiesandproductswithclearcommercialapplication(Zheng,Zhuang,andWang2020).26BrazilischosentocompareadjustmentintheprimarysectoratapercapitaGDPintherangeof$10,000–$15,000(PPP).
16 ADBEastAsiaWorkingPaperSeriesNo.54Further,weassume—unlikeBalliuetal.(2016)butinlinewithnewerresearchbyRozelleetal.(2020)—thattheemploymentshareofthesecondarysector,havingpeakedin2012,willcontinuetodeclineattheaverageratein2012–2021.Finally,theshareoftertiaryemployment,calculatedastheresidual,isexpectedtorisefrom48%in2021to57%in2030and63%by2040.Figure20:ShareofAgricultureinTotalEmploymentPRCRepublicofKoreaBrazilJapan Percent GDPpercapita PPPGDP=grossdomesticproduct,PPP=purchasingpowerparity,PRC=People’sRepublicofChina.Note:BluedashedlineshowstheforecastforthePRCupto2040.Sources:WorldBank,WorldDevelopmentIndicators;andauthors’estimates.Toestimatethefuturecontributionofthesectorallaborshifttogrowth,wecreateanalternativescenario,underwhichtheproductivityofthethreesectorsisfrozenatthe2019level—theyear2019ischosentoavoiddistortionsbytheCOVID-19shockontheeconomy—whiletheemploymentsharesofeachsectorchange.Weassumethatthetotalemploymentwilldeclineatthesamerateastheworking-agepopulation.WethencalculateabaselinescenarioofthelevelofGDP,whereboththesectoralemploymentsharesandthesectoralproductivityarefixedat2019levels(whilethedeclineintotalworking-agepopulationistakenintoaccount).Finally,wegetthecontributionofthesectorallaborshifttoTFPgrowthasthedifferenceinGDPgrowthbetweenthealternativeandthebaselinescenario.For2001–2019,weobtainTFPgrowthastheresidualfromequation(3)above.FromtheactualGDPgrowth,wedeductthecontributionofcapital,labor,andhumancapitaltogrowth,assumingacapitalandlaborshareofoutputof0.5(Bailliuetal.2016;ILO2019).27Then,weapplyanHodrick-Prescott(HP)filtertotheresidualandgetthetrendTFP(Figure21).28AsillustratedinFigure21,TFPgrowthslowedin2001–2019.Basedonourestimates,theslowdownstemsmainlyfromcontributionofthesectorallaborshiftfadingaway.Thisisconsistentwithfindingsby27Lardy(2019)discussesthelaborsharebasedonNBSdata.Therangequotedwas54%in2002,fallingto47%in2011.Thesharerecoveredto52%by2015.However,wagegrowthinthesecondhalfofthe2010swaslowerthanthatofnominalGDP(Peschel2021).Theassumed0.5takesthisintoaccount.28Bailliuetal.(2016)showsimilarresultsfor(theresidualmeasuring)TFPgrowthupto2015whentheirdataend.UsingdatafromWorldPennTable(thatassumeslowerGDPgrowthratesforthePRC),IMF(2022a)findsasimilarpatternofchangeinTFPgrowth—i.e.,anotabledeclineinthe2020s—butatlowerrates.Furthermore,estimatesbyBrandtetal.(2022)confirmthat,comparingthe2000stothe2010s,annualTFPgrowthperworkerhasdeclinedby2percentagepointsonaverage.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 17Rozelleetal.(2020)showingthatemploymentinthesecondarysector—havingahigherproductivitythantheservicesector—declinedafter2012,whilelow-skillservicesabsorbedpartofthelaid-offindustrialworkers.Whileweassumethesectorallaborshifttocontinue,itshouldslowdownafter2030whentheemploymentshareofagriculturedeclinesataslowerrate,resultinginadownwardshiftinTFPgrowth(chart22).Figure21:TotalFactorProductivityGrowth,Figure22:DecompositionofTFPGrowth,2001–20192001–2040TrendTFP(HP-filtered)SectorallaborshiftTFP(Solowresidual)FDIAverage sImportsAverage sR&DTrendTFPPercentPercent - HP=Hodrick-Prescott,TFP=totalfactorproductivity.FDI=foreigndirectinvestment,R&D=researchanddevelopment,TFP=totalfactorproductivity.Note:Forecaststartsin2020.Source:Authors’calculations.Source:Authors’estimates.TFPgrowthaveraged3.2%in2001–2019,withthesumofthefourcomponentsaddingupto2.7%onaverage(Figure22).Thesectorallaborshiftcontributedonaverage1.6percentagepointsandR&D0.7points.RealimportsandFDItogetheronlycontributedhalfapercentagepoint,withimportsslightlymorethanFDI(0.3versus0.2points),whilehalfofapercentagepointremainsunexplainedonaverage.29For2020–2040,potentialannualTFPgrowthisestimatedtoaverage1.3%,mainlydrivenbyR&D(0.6%),followedbythesectorallaborshift,realimports,andFDI.ThecontributionofR&DtoTFPstemsfromsteadygrowthinR&Dexpenditure,resultinginR&DaccountingonaveragefornearlyhalfofTFPgrowthin2020–2040,withahighershareinthe2030swhenthecontributionfromthesectorallaborshiftwanes.30OurTFPforecastislowerthanthatofBailliuetal.(2016),mainlyduetodifferentassumptionsadoptedwithrespecttotheelasticityofR&DonTFPgrowthandregardingthesectorallaborshift.OurforecastisroughlyinlinewiththatbyRobertsandRussell(2019)extrapolatinglong-termtrends.Higgins(2020)projectsinthe“prettygood”scenarioTFPgrowthtoeasefrom1.4%in2018–2028to1.0%in2028–2038—ourestimatesareonlyslightlyhigherforcomparableperiods.29Theunexplainedpartmostlydisappearswhenlookingfrom2003onward—Figure22showsabiggapbetweensumofthecomponentsandtheoverallfor2001–2002.30CBO(2005)notesthatresearchbyGriliches(1991)arguedthatR&Dspending,assuminghighsocialreturnsfromR&D,couldaccountforalmostthree-quartersofTFPgrowthduringthepostwarperiod.
18 ADBEastAsiaWorkingPaperSeriesNo.54F.ResultsWeuseequation(4)tomodelGDPgrowthin2001–2019andthenforecastpotentialgrowthfor2020–2040:31(4),whereisthemodel-fittedGDPgrowth,andαand(1-α)arecapitalandlaborshareofGDP,withαassumedtobe0.5;isthetrendgrowthincapitalstock;isthetrendlaborgrowth(i.e.,thegrowthrateoftheworking-agepopulation);andisthetrendhumancapitalgrowth,whichisderivedfromthegrowthrateofhumancapitalindex,whileisthetrendTFPgrowthderivedfromanHP-filteredTFPgrowthestimate.32ThedecompositionofcontributionstoGDPgrowthfor2001–2019isillustratedinFigure23.Asinotherstudies(Bailliuetal.2016,RobertsandRussell2019,andHiggins2020),capitalandTFParekeydriversofgrowthin2001–2019.Thecontributionofhumancapitalwasrelativelysmall.Forthe2010s,adecliningpopulationaged15–59(Figure8)ceasesthedemographicdividendofthe2000s(andearlier).Theadverseeffectsofadeclininglabor-agepopulationwillshowmorestronglyafter2025.Figure23:DecompositionofContributionstoPotentialGDPGrowth,2001–2040 CapitalTFPLaborHumancapital PotentialGDPgrowthActualGDPgrowtht Percen - GDP=grossdomesticproduct,TFP=totalfactorproductivity.Note:Forecaststartsin2020.ThefuturecontributionofcapitaltogrowthisderivedfromtheprojectedshareofgrossfixedcapitalformationinGDP.Source:Authors’estimates.ActualGDPgrowthaveraged9.0%in2001–2019,thesameaverageaspotentialgrowthaccordingtoourmodel.Fortheperiod2006–2010ourestimateofpotentialgrowth—whichabstractsfrombusinesscycles—isbelowactualGDPgrowth.Thisisconsistentwithactualgrowthexceedingitsmedium-termpotentialbeforetheglobalfinancialcrisis(WorldBank2019;Lardy2019).31DuetothesignificantimpactoftheCOVID-19ontheeconomyin2020–2022,weuseactualdatauntil2019toavoiddistortionsfromthepandemic.32Ingrowthaccounting,thecontributionofTFPgrowthtoGDPgrowthiscalculatedasaresiduali.e.,thedifferencebetweenactualGDPgrowthandthepartofGDPgrowthexplainedbychangesincapitalandlabor(includinghumancapital).However,whenestimatingfuturepotentialGDPgrowth,TFPgrowthcannotbecalculatedasaresidual.Hence,weestimatethepastTFPtrendgrowthbysmoothingtheresidualbyanHPfilterandfindasetofproxyvariablesthatmatchtheTFPtrendgrowth.WecanthenforecastTFPgrowthbasedontheprojectionsofthesevariables.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 19Asfortheforecast,themainfindingsareasfollows:(i)PotentialGDPgrowthisestimatedat5.3%onaveragein2020–2025andthentodeclinegraduallyto2.0%onaveragein2036–2040.(ii)ConsistentwithresultsfromBailliuetal.(2016),RobertsandRussell(2019),andHiggins(2020),capitalandTFParemajorcontributorstopotentialgrowthinthelongrun.Thecontributionofcapitalisforecasttograduallydeclinefrom3.4percentagepointsonaveragein2020–2025to1.2percentagepointsonaveragein2036–2040.(iii)ThecontributionofTFPtogrowthisestimatedtobesmallerinthenext2decadesbecauseofamuchsmallercontributionfromthesectorallaborshiftandlowerreturnstoR&D.Despitethat,R&DexpenditurewillbethemaindriverofTFPgrowthin2020–2040.33(iv)Ashrinkingworking-agepopulationwillincreasinglyweighongrowth,whilehumancapitalwillcontributeonaverageabout0.4percentagepointstogrowthin2020–2040.Inanutshell,thedeclineinpotentialgrowthstemsfromalowercontributionofcapitaltogrowth,ashrinkingworking-agepopulationdraggingongrowth,andlowerTFPgrowth.Table2showsthatourestimatecomesinlowerthanBailliuetal.(2016),seeAppendix3fordetails.Fortheperiodupto2030,ourestimateisbroadlyinlinewithRobertsandRussell(2019),whilebeinghigherthanHiggins(2020).Atthesametime,ourforecastfor2031–2040averages2.3%,whichisonlyslightlyhigherthanthatofHiggins(2020)for2029–2038.Table2:ComparisonofSelectedLong-TermGrowthEstimatesforthePRCStudyPeriodGDPgrowthBailliuetal.(2016)2021–20255.9%2026–20304.9%RobertsandRussell(2019)2020–20304%–5%Higgins(2020)(prettygoodscenario)2018–20283.8%2029–20382.1%Authors’forecast2020–20255.3%2026–20303.5%2031–20352.7%2036–20402.0%GDP=grossdomesticproduct,PRC=People’sRepublicofChina.Note:Onlystudiesthatapplyasimilarmethodology.Sources:Bailliuetal.(2016);RobertsandRussell(2019);Higgins(2020);andauthors’estimates.G.PotentialImpactofRecentDevelopmentsDespitethedecliningtrendinpotentialgrowth,actualGDPgrowthwillalmostcertainlybebelowtheforecastpotentialfortheperiod2020–2025.COVID-19hasweighedoneconomicgrowthinrecentyearsandwilllikelycontinuetodosoin2023(ADB2022a).Anailingpropertymarkethasalsodraggeddowngrowth,especiallyin2022.Inaddition,thetradeconflictandtechnologicalcompetitionwiththeUSisanotherpointadverselyimpactingGDPgrowth.Thesethreefactors—andhowtheyaretakenintoaccount—couldalsoinfluencethepotentialgrowthestimate.33Asmentioned,thereisahighuncertaintyabouttheelasticityofR&DspendingonTFPgrowth.
20 ADBEastAsiaWorkingPaperSeriesNo.54COVID-19.TheeconomiceffectsofCOVID-19areassumedtobetemporary.Weassumethatthecountry’szero-COVIDpolicywillbeloosenedatonepointintime,andthusexpecttheadverseeffectsofCOVID-19-relatedrestrictionsbeconfinedtothe2020–2025periodofourestimate.However,therearepossiblechannelsforCOVID-19toaffectlong-termgrowth.First,therecouldbescarringeffectsfromhumancapitallossviatheimpactofCOVID-19-relatedlockdownsonschooling.Second,thepandemiccoulddelayinvestmentgivenincreaseduncertainty.Finally,COVID-19mighthaveadverselyimpactedlaborforceparticipationrates,possiblyeveninthelongerrun.TheadverseeffectsofCOVID-19onschoolinghavebeenlesspronouncedinthePRCthaninothercountriesintheregion(ADB2022b).34Also,inthecaseofthePRC,itseemsunlikelythatCOVID-19muchdelayedinvestmentdecisionsasFDIrecoveredquicklyfromaninitialshockin2020anddomesticmanufacturinginvestment,drivenbyanexportboom,grewsolidlyinthepastcoupleofyears.ThehousingmarketslumpisnotprimarilycausedbyCOVID-19butbyborrowingcurbsondevelopers(Box1).Asforlabormarketparticipationrates,theeffectsshouldbelimitedasstate-ownedenterprisestendnottoshedstaffduringeconomicdownturns,whilelabormigrants—thoughhithardbyCOVID-19-relatedrestrictions—willlikelyremaininthelabormarket,waitingforeconomicconditionstobrighten.However,COVID-19adverselyaffectedhouseholdconsumption,whichinturnhitservicesand,togetherwiththeexportboomandstronggrowthinindustryinrecentyears,partlyreversedtheeconomy’stransitiontoservices.35Propertymarket.Therecentpropertymarketdownturncontinuestoweighoneconomicgrowth.RogoffandYang(2022)estimatethatthedirectvalueaddedoftherealestatesector(domesticcontentonly)accountedfor11.8%oftheeconomyin2021,butwhentakingitsconnectionstoothersectorsintoaccount,thefiguremorethandoublesto25.4%(includingimportedcontentaccountingforabout3percentagepoints).Thus,propertymarketinvestmentaffectseconomicgrowthbeyondconstruction.ThecurrenthousingmarketdownturninthePRCmainlystemsfromgovernmentdeleveragingeffortsbyimposingborrowingcurbsondevelopers.Inourmodel,thedeclineinpotentialgrowthinthelatterhalfofthe2020sismostlydrivenbyalowercontributionfromcapitaltoGDPgrowth.Whilethefutureworkforcecanbedeterminedwithahighdegreeofcertainty(absentpensionreforms),thecontributionofcapitaltogrowthcouldfallfasterincasethedeleveragingthepropertymarketleadstoafaster-than-expecteddeclineinoverallinvestment(asashareinGDP).Foradiscussionontheroleofinvestmentanddebt,seeBox1.Technologicalcompetitionandpotentialdecoupling.OurassumptionstakeintoaccountthatbarrierstoacquiretechnologyhaveincreasedandthattherewilllikelybelessknowledgeexchangebetweenthePRCandtherestoftheworld,loweringreturnstoR&Dexpenditure(andTFPgrowth)inthefuture.Whilethediscountappliedisalreadysubstantial,atechnologicaldecouplingcouldhaveadverseeffectsonpotentialgrowthstrongerthanassumed.36Comparison.Ourmodelpredictspotentialgrowthof5.3%onaveragein2020–2025(from6.8%in2016–2019),decliningto3.5%in2026–2030.37Onaverage,ourestimateofpotentialgrowthis4.4%for2020–2030.ThisisbroadlyinlinewithrecentestimatesthatcouldtakeCOVID-19andamorerestrictedaccesstotechnologyintoaccount.WangandZhang(2022)seepotentialGDPgrowthat4.0%–4.5%inthisdecade.Zhu,Ng,andGe(2021)estimatepotentialgrowthat5.3%in2021–2025and4.4%in2026–2030(withahighercontributionofTFPthanourestimate).34Inthemodel,returnstoeducationareassumedtobefixovertime(butdifferbylevelofschooling);adjustingthemovertime(possiblyaccountingforeffectsfromCOVID-19)isbeyondthescopeofthispaper.Atthesametime,thecontributionofhumancapitaltogrowthissmallsothattheresultsofthemodelwouldnotchangesubstantially.35SeediscussioninBox3laterinpartIV.36Technologicaldecoupling—i.e.,theundoingofcross-bordertradeinhigh-techgoodsandservices—couldhaveasubstantialadverseimpactonthePRC’sGDPgrowthdependingonthecoalitionsofcountries(Cerdeiroetal.2021).37Inourmodel,weassumeagradualdeclineintheshareofGFCFinGDPby2030.Asweknow(andtakeintoaccount)theactualGFCFin2020–2021(at42%ofGDP),theeffectsofafallingGFCFsharewillshowmorestronglyafter2025,explainingthedropinpotentialgrowthin2026–2030.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 21Box1:GDPGrowth,Investment,andDebtAccordingtoourestimate,potentialgrowthaverages4.4%in2020–2030and2.3%in2030–2040.Animportantquestionisifpotentialgrowthcanbeachieved.Intheory,actualgrowthshouldbeabletoreachpotentialgrowth.Whilenotbeinganinputfactorintheestimateofpotentialgrowth,debt—especiallyatanexcessivelevel—canpotentiallyhurtgrowth(ReinhartandRogoff2010;Cecchetti,Mohanty,andZampolli2011).aInthePeople’sRepublicofChina(PRC),debtmorethandoubledfrom139%ofgrossdomesticproduct(GDP)attheendof2008to287%bytheendof2021(FigureB1).ItseemsunlikelythatthePRCcancontinuetoaccumulatedebtatthespeedwitnessedin2009–2019(withoutCOVID-19),whendebtroseby124percentagepointsofGDP,oronaverageover11pointsperyear.Extrapolatingthistrendwouldresultinadebt-to-GDPratioofabout500%by2040.EvenJapan,thehighest-indebtedcountrycoveredbytheBankforInternationalSettlements(BIS)statistics,onlyhaddebtequivalentto420%ofGDPbytheendof2021.FigureB1:DebtStructure,FigureB2:CreditGapandChangeinDebt,2007–20222008–2021 NonfinancialcorporationsCredit-to-GDPgapGeneralgovernmentDebt(RHS) Households change oint p DPge yG earonyea of Percenta - - r- - - - Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q GDP=grossdomesticproduct.GDP=grossdomesticproduct.Note:Credit-to-GDPgapisthedifferencebetweenthecredit-to-GDPratio(fortheprivatenonfinancialsector)anditslong-runtrend;debtofprivatenonfinancialsector.Source:BankforInternationalSettlements.https://www.bis.org/statistics/totcredit/totcredit.xlsxSource:BankforInternationalSettlements(accessed6(accessed2December2022).October2022).Economicgrowthinthe2010scouldhavebeendrivenbyunsustainableoverinvestment.TheBISprovidesestimatesofthecreditgap—i.e.,thedifferencebetweentheratiooftotalcreditrelativetoGDPanditslong-runtrend—forseveralcountries,includingthePRC(FigureB2).Inourcontext,creditgapanalysiscanbeusefulasoverinvestmentmustbefinanced.Apositivecreditgapcanbeasignofoverinvestment.Thecreditgapshotupafter2008andcamedownrapidlyin2017–2018.So,wemighthaveseenaboutadecadeofoverinvestment.In2018infrastructureinvestmentslowedsharply(Figure3inthemaintext)andnewassetmanagementrules—aimingtocurbshadowbankfinancing—wereannounced(Economist2018),whichcouldhelpexplainthequicknarrowingofthegap.Asfordebt,deleveragingeffortstargetingstate-ownedenterprisedebtshowedsomeeffectsinthesecondhalfofthe2010sbeforetheCOVID-19shockontheeconomydroveupdebtagain(FiguresB1andB2).Continuedonnextpage
22 ADBEastAsiaWorkingPaperSeriesNo.54Box1continuedForseveralreasonsitseemsunlikelythatthePRCcancontinuetheirinvestmentpatternforanother2decades.First,housingdemandisabouttopeakin2022(Yao2022).Inthisrespect,onecaninterpretthegovernmentcurbsondevelopers’borrowingasanattempttopreventthemfromoverbuildingaswellastoreiningrowthgeneratedbyinefficientinvestment.bSecond,thecountry’sphysicalinfrastructure(airports,high-speedrailway,highways,etc.)isalreadywelldeveloped.ContinuingtoinvestinitathighrateswillnotonlyyielddecliningreturnstoinvestmentintermsofGDPgrowth,butwouldalsooverstretchlocalgovernments’capacityforadditionaldebt.Aslocalgovernmentsreceiveasubstantialshareoftheirrevenuefromlandsales,thecurrentpropertymarketdownturnputspressureonlocalgovernmentfinances.Inaddition,manyinfrastructureprojectswithhighreturnshavebeenundertakenalreadysothatremainingprojectsarelessattractive.Finally,manyprovinces—especiallypoorerones—havealreadyaccumulatedasizablestockofpublicdebt,whichmakesitmoredifficultforthemtoshoulderadditionaldebt(FigureB3).cFigureB3:PublicDebtandGrossRegionalProductPerCapitabyProvince,2021 AH=Anhui,BJ=Beijing,CQNX=Chongqing,FJ=Fujian,GD= Guangdong,GRP=grossregional product,GS=Gansu,GX=GuangxiGXZhuangAutonomousRegion,GZ= Guizhou,HA=Henan,HB=Hubei,XJTJ JLHE=Hebei,HI=Hainan,HL=HBFJHEHAHeilongjiang,HN=Hunan,IM= QHLNInnerMongoliaAutonomousRegion,YNHNSNGDJXJL=Jilin,JS=Jiangsu,JX=Jiangxi,HLSCCQLN=Liaoning,NX=NingxiaHuiDebt-to-GRPratio()XZAHZJAutonomousRegion,QH=Qinghai,GSSXHNIMBJSDJSSHSC=Sichuan,SD=Shandong,SH=GZShanghai,SN=Shaanxi,SX=Shanxi,TJ=Tianjin,XJ=XinjiangUygur AutonomousRegion,XZ=TibetGRPpercapita(CNY’)AutonomousRegion,YN=Yunnan,ZJ=Zhejiang.GRP=GrossRegionalProduct.Note:Officialpublicdebt(generalandspeciallocalgovernmentdebtoutstanding)only.Sources:MinistryofFinanceandNationalBureauofStatisticsviaCEICDataCompany(accessed10October2022);andADBcalculations.Insum,thegrowthmodelofthe2010scannotbeextendedforanother2decades.Reformsareneeded,includingimprovinginvestmentefficiencyandreducingthecapitalintensityofgrowth.Notes:aSeePettis(2022)onwayshowdebtcanhurtGDPgrowthinthecaseofthePRC.bPettis(2022)seesthisasanexampleofagentsinaneconomicsectoradjustingbehaviortoavoidpotentiallosses.Beforethegovernmentclampdowntherewastheassumptionprevailingthatthedebtofdeveloperswouldberesolvedthroughimplicitgovernmentguarantees;nowpropertydevelopers,homebuyers,suppliers,contractors,banks,localgovernments,etc.,changebehaviortoprotectthemselvesfromdefaults(suchaswithholdingpayments),whichwouldleadtolossesontheirside.Thesebehavioralchangescauseeconomicgrowthtoslow.cSpecifically,morelocalgovernmentsareapproachingthe10%threshold.In2016,thePRCcreatedrulesrequiringfiscalconsolidationincaselocalgovernmentinterestpaymentsbreachthresholds:(i)interestpaymentsongeneralgovernmentbondsexceed10%offinancespendinginthegeneralpublicbudget,or(ii)interestpaymentson(localgovernment)specialbondsexceed10%ofspendinginthegovernmentfundbudget(He2022;StateCouncil2016).
IV. REFORMSTOBOOSTLONG-TERMGROWTHA.ReformAreasThispartoutlinespossiblereformstoenhancethecountry’slong-termgrowthpotential.TheadvantageofusingaCobb-Douglasproductionfunctionisthatitallowsforabreakdownofthedriversofeconomicgrowth.Ourmodelprovideabaselineforlong-termgrowth—absentmajorreformsandshocks38—whichcanhelpguidepolicymaking.Potentialgrowth,thoughhardtochange,canbeincreased,inparticularbystructuralreforms.Mirroringthemodel’sinputfactors,wefocusonreformspertainingtolabor,humancapital,capitalstock,andTFP.B.LaborThemodelassumesthattheworkforcegrowsatthesamerateastheworking-agepopulation.Iftheretirementageisraised,theworking-agepopulationwouldincreaseasmorepeoplewerecapturedbytheworking-agebracket.Inpractice,suchreformsaremorecomplexbecausecurrentlytheretirementageformenis60years,andforwomeninblue-collarandwhite-collarjobs50and55years,respectively,whiletheaverageretirementageisonlyinthemid-50s(Economist2021).Anotherissueisdemographics.Giventhecurrentlowbirthrates,anincreaseinfertilitycouldslowthespeedofdeclineoftheworkforceinthefuture,butnotreversetrendintheforeseeablefuture.Asweonlylookatuntil2040—notalonghorizonintermsofdemographics—andbirthratesareknownupto2021,agradualincreaseinthebirthratewillhardlyaffectourprojections.Also,theintroductionoftheuniversalTwo-ChildPolicyin2015hadshownnolastingeffect(Figure24),andtheexperiencefromothercountriesshowsthatthefertilityrateremainslowonceithasdeclined(Figure25).Figure24:BirthRate,Figure25:FertilityRateforSelected1960–2021Countries,1960–2020 BirthsinthePRCPRCn USioat OECD an pulRepublicofKoreapo omJapanw er p
hs r rtPe Bi PRC=People’sRepublicofChina.OECD=OrganisationforEconomicCo-operationandNote:Livehumanbirthsper1,000population.Development,PRC=People’sRepublicofChina,US=UnitedStates.Source:NBSviaCEICDataCompany(accessed2Source:WorldBank,WorldDevelopmentIndicatorsDecember2022).(accessed11July2022).38Inthebaselinemodel,weassumenomajorstructuralreformsanddonottakeintoaccountbusinesscycles.
24 ADBEastAsiaWorkingPaperSeriesNo.54Againstthisbackground,thegovernmentshouldtakemeasuresthatcouldmitigatetheimpactofdemographicagingonthelaborforce:(i)Extendworklife.Whilethelegalretirementageshouldbeincreased,itisimportantthatpeopleactuallyworklonger.Reformsmuststartwithpeopleactuallycontinuingtoworkbeyondtheirmid-50s—otherwiseanextensionofthestatutoryretirementagebysomeyearscomestonothing.39(ii)Improvehealthconditionsoftheworkforce.Improvedhealthcareandpromotinghealthyagingisneededtocopewithanagingworkforce.Whilethefocusshouldbeonasafeandsecureworkingenvironment,occupationalhealthservices,andoccupationaldiseaseprevention,effortsareneededtoscaleupscreening,earlydiagnosisandcontrollingchronicconditions,aswellasdetectingcognitiveimpairmentanddepressionattheprimarycarelevel(BaiandLei2020).(iii)Raisefemaleworkforceparticipation.40Possiblemeasuresincludestrengtheningequalemploymentopportunities,increasingmaternityleave,andimprovingsupportforchildcare,elderlycare,andsinglemothers(ADB2021b).(iv)Increaselabormobility.Thisincludesfurtherrelaxingthehouseholdregistrationsystem(hukou),improvingaccesstosocialservicesforlabormigrants(includingtheaccesstoeducationfortheirchildren),andensuringtheportabilityofsocialsecuritybenefitssothatpeopledonothavesubstantialdisadvantageswhenmoving.41(v)Increasefamilysupport.Amassiveshiftofpolicyprioritiesandresourcesmusttakeplacetosupportfamiliesinordertoraisethebirthrate.C.HumanCapitalReformsfocusingonincreasinghumancapitalcanberolledoutfasterthanthoseaimingatthelaborforce.Despitesignificantimprovementsovertime,thePRC’seducationalattainmentlagsbehindthoseinadvancedeconomies,suggestingroomforcatchingup.Whiletheforecasttakesintoaccountanincreaseintheyearsofeducation—whichserveasaproxyfortheincreaseinhumancapital—qualityofeducationisalsoimportant.Anincreaseinhumancapitalbasedonbetter—notnecessarilymore—educationwouldcontributetohigherpotentialgrowth.42Asmentioned,increasingaverageyearsofschoolingisimportantforraisinghumancapital.ThePRCsawsignificantimprovementineducationattainmentlevelsoverthepastdecades:theaverageyearsofschoolinghasincreasedsubstantiallyfrom5.7yearsin1980to8.7yearsin2015(Figure26).However,thereisstillagaptoadvancedeconomieslikeJapan,theRepublicofKorea,andtheUS.Asaresult,in2015only37.4%ofthePRC’slaborforcehadcompleteduppersecondaryeducation,comparedwith65.5%acrossG20countriesand100%inJapan(Figure27).Similarly,only18.2%ofthePRC’slaborforce39Suchreformstendtoraisethestatutoryretirementageonlybysomeyearsandareusuallyimplementedgradually.40AccordingtohouseholdsurveydatabytheSouthwesternUniversityofFinanceandEconomics,thelaborforceparticipationrateofwomenaged16–64declinedfrom61.6%in2011to58.2%in2017,whilethatofmenshowedamuchsmallerdeclinefrom76.6%to76.0%duringthesameperiod(DongandJoffre2019).41Immigrationofskilledpeopleintoareaswithlaborshortagewouldmitigatesuchshortages.Thoughunpopular,extendingworkinghourswouldalsomitigatetheeffectofashrinkingworkforce.WedonotfurtherelaborateontheseoptionsasimmigrationintothePRChas—leavingotherchallengesaside—highlanguagebarriers.Asforextendingworkinghours,diminishedproductivityinsomeareasmightstemfrominsufficientworkorinadequatequalificationofemployees,whichcannotbeaddressedbysimplyextendingworkhours.42Qualityofeducationishardtomeasure.Thoughitisnotdirectlyaninputfactorinthemodel—whileyearsofschoolingare—improvingeducationincreasespotentialgrowthbecauseitraisesworkforceskillsandtherebyproductivity.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 25hadcompletedtertiaryeducationin2015,whichislowerthanacrossG20countriesat29.8%andmuchlowerthaninJapan,theRepublicofKorea,andtheUS.However,tertiaryeducationattainmenthasbeenrisinginthePRC,reflectedinagrowingnumberofnewcollegegraduates.Arecordnumberof10.76millionwasexpectedbythegovernmentin2022(Cheng2022).Figure26:ComparisonofAverageYearsofFigure27:ComparisonofEducationSchooling,1950–2015Attainment,2015 Laborforcewithcompleteduppersecondaryeducation Percent CyeaUS PRG orapanAverageyearsofschoolingOECDJepublic GermanRofKLaborforcewithcompleted tertiaryeducation PRCJapan RepublicofKoreaUSPercent G Germany CyeaUSPRG orapanOECDJepublicGermanRofKPRC=People’sRepublicofChina,US=UnitedStatesOECD=OrganisationforEconomicCo-operationandNote:Foragegroup15–64.Development,PRC=People’sRepublicofChina,US=UnitedStates.Source:BarroandLee(2021).Source:Lietal.(2017)ThereisadisparityinyearsofschoolingbetweenruralandurbanareasinthePRC.Theaverageyearsofschoolingfortheurbanworking-agepopulationwas11.4years,versus9.1yearsinruralareasin2019(Li2021).43Theurban–ruralgapinyearsofschoolinghasbeennarrowingsince1985,buthaswidenedagainafter2015(Figure28).Also,thereisagapintheworking-agepopulation’sseniorsecondaryandtertiaryschoolcompletionrates(Figure29).43TheChinaHumanCapitalReportdefinesworking-agepopulationinthePRCasages16–54forfemaleand16–59formale.
26 ADBEastAsiaWorkingPaperSeriesNo.54Figure28:YearsofSchooling,Figure29:UrbanandRuralSchool1985–2019CompletionRates,2005,2015,and2019 Urban–ruralgapUrbanUrban TotalRuralRural t Percen Averageyearsofschooling AtleastseniorTertiarysecondaryNote:Working-agepopulationdefinedas16–54forNote:Working-agepopulationdefinedas16–54forfemaleand16–59formale.femaleand16–59formale.Source:Li(2021).Source:Li(2021).Againstthisbackground,thefollowingpolicyrecommendationsaremade:(i)Increaseyearsofschooling.Greatereffortsareneededtoincreaseyearsofschooling,especiallyinruralareas,tobolsterhighereducationalattainment.Experiencesfromhigh-incomecountriesshowthatthosewithoutpost-secondaryqualificationsaremostlikelytofacejoblossandencounterlong-termunemployment(ADB2020b).(ii)Improveruraleducationandsupportchildrenofmigrantworker.Morefundingfromthecentralgovernmentshouldbeallocatedtoruraleducationandchildrenofmigrantworkersshouldgetmoresupportfromthegovernmenttogeteasieraccesstoeducationincities(Lietal.2017).(iii)Improvethequalityofeducationandqualificationoftheworkforce.Thisincludesexpandingearlychilddevelopment,improvingthequalityoftertiaryeducation,strengtheningtechnicalandvocationaleducationandtraining(TVET),andincreasingon-the-jobandlifelongtrainingopportunities(seeAppendix5foradiscussionoftheserecommendations).D.CapitalOverviewChangestotheallocationofcapitalandcreditmightbeoneofthebiggestfactorsraisingpotentialgrowthinthePRC,buttheyarechallengingtoimplementbecausetheyrequireadjustmentstoeconomicstructuresaswellaspolicychanges.Thebankingsystemispredominatelystate-owned,withlargecommercialbanksprovidingabouthalfofallloans.Bankloans,inturn,accountforthevastmajorityofcompanyfinancing.Moreover,thestate-ownedbankingsectordominatescreditallocationandpreferstolendtostate-ownedenterprises(SOEs)asthesefrequentlybenefitfromgovernmentsupport.4444CreditflowstoSOEswere83%oftotalcredittononfinancialenterprisesin2016[latestavailabledata](Lardy2019).
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 27Inthefollowing,welookatcapitalallocationanditsefficiencyaswellasatcreditallocation.Box3attheendofthissectionwillcoverstructuralchange.ReturnstoinvestmentTheimpactofinvestmentongrowthhasdeclinedinthePRCaftertheglobalfinancialcrisis.Theincrementalcapitaloutputratio(ICOR),whichmeasureshowmuchcapitalisneededtogenerateoneextraunitofoutput,hasincreased.In2000–2009,theratioaveraged3.7,followedby5.8onaveragein2010–2019(Figure30).45In2019,itexceeded7.46Thus,morecapitalisneededtogenerategrowth.AnobviousquestioniswhyreturnstocapitalhavedecreasedandtheICORhasincreased.Marginalproductivityofcapitalgenerallydeclineswithahighercapitalstockperworker.RealreturnoncapitalhastrendeddowninthePRCthelastquartercentury(IMF2022a).Thedecline,inparticularoverthelastdecade,couldbedrivenbythecompositionofinvestment47and/ortheownershipofinvestment.Thestateholdingshareintotalinvestmentpickedupafter2015,whilethatofprivatelyheldcompaniesslightlyeased,raisingconcernsoverallocationefficiency(Figure31).Figure30:IncrementalCapitalOutputFigure31:FixedAssetInvestmentbyRatio,2000–2019OwnershipType,2004–2017ICOR StateholdingPrivateholding Percent ICOR=incrementalcapitaloutputratio.Note:Sharesdonotaddupto100%asthereareothercategoriesinthestatistic.Sources:NBSviaCEICDataCompany(accessed20Sources:NBSviaCEICDataCompany(accessed26September2022);andADBcalculations.September2022);andADBcalculations.45Calculatedasinvestmentratio(GFCFas%ofGDP)dividedbyGDPgrowth(Kwan2004).46AsGDPgrowthslowedsharplyin2020,whileinvestmentwasthemaincontributortogrowthasconsumptiontookahitfromCOVID-19,theratiospikedthatyear,butreversedin2021(notshowninchart).47Theavailablestatisticsoninvestmentaresketchy.EstimatesfromHerd(2020)until2016showthattheshareofcapitalformationinthebusinesssectordeclinedovertimeandthesharesofinfrastructureandhousingincreased,whilethecapital-outputratiointheinfrastructuresectorincreasedthemost,followedbythehousingandthebusinesssector,whichexplainstheriseintheICOR(Brandtetal.2020).
28 ADBEastAsiaWorkingPaperSeriesNo.54State-ownedenterprisesandsupply-sidestructuralreformsTheshareofemployeesworkingatSOEshasdeclinedovertime(Figure32).Atthesametime,SOEassets(as%ofGDP)areseveraltimeshigherinthePRCthaninothercountries(IMF2021b)andhaveincreasedaftertheglobalfinancialcrisis(Figure33).However,informationisonlyreadilyavailableonindustrialstateholdingenterprises,i.e.,enterpriseswherethestatehasadominantequityshareorotherwiseactualcontrol.48Thesetendtobelessefficientthanprivateenterprisesandaboutoneinfourofthemisloss-making(Box2).Figure32:EmploymentbyCategoryofFigure33:SOEAssets,andReturnonEmployer,1980–2019AssetsbySector,2003–2020SOEsCollectivePrivateOtherAgricultureSOEassets(RHS) ConstructionIndustry Services of G DP PercentPercent SOE=state-ownedenterprise.GDP=grossdomesticproduct,SOE=state-ownedenterprise.Note:NonfinancialSOEsonly.NodataonreturnonassetsofindustrialSOEsin2017–2018.Sources:MinistryofHumanResourcesandSocialSecuritySources:MinistryofFinanceandNBSviaCEICDataviaCEICDataCompany(accessed16September2022);CompanyandWIND(accessed2December2022);andandADBcalculations.ADBcalculations.Thereturnonassets(ROA)ofSOEsintheservicesectorhasfirstrisenbeforetheglobalfinancialcrisisand—afterarisefollowedbyalong-stretcheddecline—stoodin2020againapproximatelyatits2003value(Figure33).TheROAofindustrialSOEshasdeclinedovertimeuntilthemid-2010s,andthenimprovedbynearly1percentagepointfrom2.0%in2015to2.9%in2018–2019,reflectingsupply-sidestructuralreformsin2016–2018.49Supply-sidestructuralreformsfocusedonthesteel,coalandelectricitysectors,andzombieenterprisestherein.In2016,theNationalDevelopmentandReformCommissionreleaseddetailed5-yearplansforthesteelandcoalindustries,aimingatconsolidatingthesector.Otherindustriestargetedforcapacityreductionincludedcement,glass,andagriculture(Boulter2018).Atthesametime,SOEsintheservicesectorhavegainedlessattentionandhavenotseentheirROArise(Figure33).Positiveeffectsofstructuralreformsincludedclosingsomezombiefirmsandhighlyenvironment-polluting,energy-inefficientcompanies.Itshouldalsobenotedthatthegovernmentinvestedinindustrialupgradingsothatsupply-sidereformswerenotlimitedtoclosingfirms.Thereformsalsocontributed,aspartofa48SOEsareenterpriseswhoseentireassetsarestate-ownedandareregisteredasanunincorporatedeconomicorganization.49Kroeber(2020,chapter14)notesthatsupply-sidestructuralreformsstartedtocollidewitheconomicgrowthtargetsin2018andwereshelved.Boulter(2018)notesthattheprofitmargininthesteelsectorimprovedin2016–2018.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 29broaderdeleveragecampaign,toa[temporary]stabilizationincorporatedebt(Boulter2018andFigure41).Theoverallpositiveexperienceofsupply-sidestructuralreformssuggestsexploringpossibilitiestoextendthemtootherindustries.ThereremainseveralSOE-relatedissuesthatneedtobeaddressed.Dataavailableonindustrialstateholdingenterprisespaintsthefollowingpicture(Box2):(i)Theshareofstateholdingenterprisesinindustrydeclinedrapidlyintheearly2000s,whileinthe2010sthisdevelopmentdecelerated.(ii)Theshareofloss-makingindustrialstateholdingenterprisesdeclinedinthefirstdecadeofthe2000s,butthistrendstoppedinthe2010sandthesharehashoveredaroundatoneinfour.Annuallossesbyindustrialstateholdingenterprisesaveraged0.5%ofGDPinthe2010s.(iii)Industrialstateholdingenterpriseshave,onaverage,lowerreturnsonassetsandonequitythantheirprivatepeers.(iv)Industrialstateholdingenterprisesusedtohaveahigherliabilities-to-assetratioaswellasahigherliabilities-to-equityratiothantheirprivatepeers.However,industrialstateholdingenterpriseshavereducedtheirliabilities-to-assetratioinrecentyears,whilethatofprivateindustrialcompaniesrosesharplyin2018.TherearereformneedswithrespecttoSOEs.Fourareasarefrequentlyhighlighted:50(i)ClarifySOEs’scopeandfunction.WhilethefunctionandscopeofSOEsneedsclarifications,inparticulartheareasinwhichthegovernmentwantstoretainSOEsshouldbesmallandwell-specified.Competitionshouldbestrengthenedbymakingmarketentryeasierforprivatecompaniesintheotherareas.ThiscouldincreasepressureoninefficientSOEstoincreasetheirefficiency.Monopoliesshouldbekepttoaminimumandtheuseofadministrativemonopolies(i.e.,exclusiverightsgrantedbyregulations)reduced.(ii)Leveltheplayingfield.ThereshouldbealevelplayingfieldfortheprivatesectortocompetewithSOEs,inparticularintermsofaccesstolandandcredit.ThisincludesremovingimplicitguaranteesforSOEsandtheirpreferredaccesstobankcredit.(iii)SeparatesocialfunctionsfromSOEs.InthePRC,manySOEsalsoperformsocialfunctions,suchasstabilizingemploymentandprovidingsocialservicesthatareinothercountriesfrequentlyfallintothecategoryofbasicpublicservice.SocialfunctionsshouldbeseparatedfromSOEstoincreasetheircompetitivenessandbetterrealizeeconomiesofscaleandscope.(iv)ImproveSOEmanagement.ThemanagementandsupervisionofSOEsmustbestrengthenedtoraisetheirefficiency.Elementssuchasaligningincentivemechanismsandgraduallyadoptingmarket-basedsalariesarementionedinthe14thFive-YearPlan.50SeeforinstancemorecomprehensivediscussionsinWorldBankandDRC(2013and2019);IMF(2019,2021b,and2021c);ADB(2021d);andOECD(2022a).
30 ADBEastAsiaWorkingPaperSeriesNo.54Box2:IndustrialStateHoldingEnterprisesFigureB4:IndustrialStateHoldingEnterprisesTheshareofstateholdingenterprisesinindustrialassetsSincetheglobalfinancialcrisis,theshareofloss-makingfellsharplyinthe2000s.stateholdingenterpriseshoversaroundone-quarter.ShareofstateholdingenterprisesShareofloss-makingstateholdingenterprisesLossesbystateholdingenterprises(RHS) ofGDP Percent Percent PrivateindustrialenterpriseshavehigherreturnsonassetsStateholdingenterprisespayahighershareoftheirpre-taxandonequity.profitsoninterestpayments.ROAstateholdingenterprisesStateholdingenterprisesROAprivateenterprises Privateenterprises ROEstateholdingenterprisesROEprivateenterprises t t PercenPercen Theliabilities-to-assetratioofcentralSOEshasbeenhigherLiabilities-to-equityratiosofstate-holdingenterprisesthanthoseofindustrialstateholdingenterprisesafter2008.andprivateenterpriseshaveconvergedinrecentyears.StateholdingenterprisesStateholdingenterprisesPrivateenterprisesPrivateenterprises CentralSOEs t PercenPercent GDP=grossdomesticproduct,ROA=returnonasset,ROE=returnonequity,SOE=state-ownedenterprise.Note:Thedataisnotcontrolledforindustry.IMF(2022b)controlscapitalproductivityinmanufacturingforindustrycompositionofSOEs,withsimilarresultstotheonesinthechartaboveonreturnonassets.CentralSOEsundertheStateCouncilincludenon-industrialSOEs.Sources:NBSandMinistryofFinanceviaCEICDataCompanyandWIND(accessed16September2022);andADBcalculations.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 31CreditallocationWhilecreditallocationcantheoreticallybeanalyzedfromvariousangles,thereisonlylimitedcredit-relatedinformationavailableforthePRC.Furthermore,manyrelevantstatisticsendin2019(bythelatest).Thus,inthefollowingweanalyzebroadtrendsincredit,usingavailableinformationonbusinessloans,thebankingsectorstructure,andtheevolutionofkeyinterestratesanddebt.Fromthestatisticsontotalsocialfinancing—abroadcreditaggregateincludingbankloans,shadowbankfinancing,equityfinancing,andgovernmentandcorporatebonds(andsomesmallerotheritems)—somebroadtrendsrelevanttocompanyfinancingcanbeinferred.Figure34:SelectedFinancingSources,Figure35:ShadowBankFinancing,2002–20212002–2021LoansShadowbankfinancingShareincreditoutstandingCorporatebondsEquityShadowbankfinancingoutstanding(RHS) increasey t earonyear Percen - Note:Loansbyfinancialinstitutions.ThesumoftheNote:Creditdefinedasthesumofloansbyfinancialfourcategoriesisnormalizedto100%.Shadowbankinstitutionsandshadowbankfinancingoutstanding.financingcomprisesentrustloans,trustloans,andbanks’DefinitionofshadowbankfinancingasinFigure34.acceptancebills.Sources:People’sBankofChinaviaCEICDataCompanySources:People’sBankofChinaviaCEIC(accessed20(accessed20September2022);andADBcalculations.September2022);andADBcalculations.Theshareofequityfinancinghasstayedverylow,whilebondfinancingcouldgainshareovertime(Figure34).Secondly,shadowbankfinancing—judgingfromtheregulatedpartcoveredbythestatistics—hasbeenreinedinsincethemid-2010s(Figure35).Theshareofshadowbankfinancinginoverallcredit—definedhereassumofbankloansplusregulatedshadowbankfinancing—hasfallentobelow9%in2021,downfromabout20%in2013and2014.Businessloans.In2019,outstandingbusinessloansequaledto56.4%ofbankloans(or87.3%ofGDP).Amongbusinessloansoutstanding,thetopthreecategoriesweremanufacturing(9.6%);transport,storage,andpostalservices(9.1%);andleasingandcommercialservice(8.4%).Threeothercategorieshadasharemorethan5%:realestate(6.4%),waterconservancy,environmentandutilitymanagement(6.2%),andwholesaleandretailtrade(6.1%).Thesetopsixcategoriesaccountedfornearlyhalf(45.8%)ofbusinesscreditoutstanding.Totalloanshaverisenfasterthanbusinessloansupto2019,thelatestavailableyear(Figure36).Smallandmedium-sizedenterprise(SME)loansgrewaspercentageofGDPbut,after2012,expandedonlybroadlyinlinewithbusinesscreditsothatitsshareinbusinessloansstagnatedandevendeclinedasashareintotal
32 ADBEastAsiaWorkingPaperSeriesNo.54loans.In2017,therewasachangedinstatisticalcategoriesandSMEloanswerediscontinued.Loanstomicroandsmallbusiness,anewlyintroducedcategory,haveremainedvirtuallyflatasapercentageofGDP.Figure36:LoansOutstandingbyCategory,Figure37:MaturityofFinancing,2009–20212013–2020TotalBusiness(total)SMElendingBusinessloans HouseholdsSMEMSE ofGDP GDP=grossdomesticproduct,MSE=microandsmallenterprise,SME=smallandmedium-sizedenterprise.Note:TotaldomesticloansandhouseholdloansbyNote:Chartshowstheratioofshort-termloanstofinancialinstitutionsfromPeople’sBankofChina;othersmedium-andlong-termloans.Short-termloanshaveafromOECD.maximummaturityof1year.Sources:OECD(2022b),People’sBankofChinaviaCEICSources:OECD(2022b);andADBcalculations.(accessed20September2022);andADBcalculations.Intermsofmaturity,SMEshavebecomeslightlylessdependentonshort-termlending(Figure37).However,after2017therehasbeennofurtherprogress,whichmaybeexplainedbyrisingeconomicuncertainty.Bankingsectorstructure.Theshareofassetsoflarge(state-owned)bankshasdeclineduntilthemid-2010s,butstabilizedafter(Figure38).Conversely,ruralandcitycommercialbanksaswellasjoint-stockcommercialbankscouldgainshareuntilthemid-2010s.Figure38:MarketSharebyTypeofBank,Figure39:KeyInterestRates,2003–20212008–2022Joint-stockcommercialbanksPolicybanksAveragelendingrateYuebaoCitycommercialbanksOtherHouseholdsavingsdepositsrateRuralcommercialbanksLargecommercialNetinterestmargin(RHS) banks Percent Percent Percent QNote:Byassets.LargecommercialbankscompriseICBC,Note:Quarterlydata;lastobservationQ22022.Yu’eBOC,CCB,ABC,andBoCom.baoisapopularmoneymarketaccountinthePeople’sRepublicofChina.Sources:ChinaBankingandInsuranceRegulatorySources:People’sBankofChinaandChinaBankingandCommissionviaCEIC(accessed20September2022);InsuranceRegulatoryCommissionviaCEICandWINDandADBcalculations.(accessed20September2022);andADBcalculations.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 33Financialrepressionandassetpriceinflation.Banks’netinterestmargintendstomovewiththeaveragebanklendingrate(Figure39).Thisratehasbeenonadecliningtrendsince2012,whilethehouseholddepositratehasremainedunchangedforseveralyearsatalowlevel.Asaresultofalackofsuitableformsofinvestments—givenlowdepositrates,anuncertainoutlookonthestockmarket,andaclosedcapitalaccount—householdpurchasedproperty,whichnotonlydroveuppropertypricesbutalsoresultedinabiggershareofcreditflowingtohouseholds(Figure36).Thisresultedinapickupinhouseholddebtdrivenbymortgageborrowinginthesecondhalfofthe2010s(Figure40).Meanwhile,corporatedebtaspercentageofGDPhasdeclinedinthesecondhalfofthe2010s,beforepickingupagaininthewakeoftheCOVID-19pandemic(Figure41).Figure40:HouseholdLoansbyType,Figure41:CorporateDebt,2007–20222007–2021HouseholdoperatingloansNominalGDPOtherconsumerloansCorporatedebtMortgageloans Corporatedebtoutstanding(RHS) Householddebt(RHS) of of G G DPDP CNYtrillion increaseyearonyear QQQQQQQQQQQQQQQQ - GDP=grossdomesticproduct.GDP=grossdomesticproduct.Note:Missingdatapointsbefore2013wereestimatedNote:Nonfinancialcorporates.bylinearextrapolation.HouseholddebtfromBankforInternationalSettlements(BIS).Sources:People’sBankofChinaviaCEIC;BIS(bothSources:BankforInternationalSettlementsandNBSviaaccessed2December2022);andADBcalculations.CEICDataCompany(accessed19September2022);andADBcalculations.Fromtheinformationabove,wecanstatethefollowingoncreditallocationinthePRCbasedontheavailablestatistics:(i)Bytypeoffinancing,shadowbankingfinancingcapturedbythestatisticshasdeclinedandtherelianceonbankloanshasincreasedsincethemid-2010s,whilethesharesofbondandequityfinancingvirtuallystagnatedinrecentyears.(ii)Businessloansoutstanding(aspercentageofGDP)havehardlyincreasedoverthepastdecade.Mostoftheincreaseinloanstakenwerebyhouseholds,whichusedthemoneymostlyformortgagesinthesecondhalfofthe2010s(Figure40).(iii)LoanstoSMEscouldnotincreasetheirshareafter2012(until2017,latestavailabledata).LoanstomicroandsmallenterpriseshaveremainedflataspercentageofGDPin2017–2019.(iv)Theratioofshort-termtomedium-andlong-termSMElendingstagnatedafter2017,wheneconomicconditionsbecamefraughtwithhigheruncertainty.(v)Bigbankscouldstabilizetheirassetshareinthesecondhalfofthe2010s.Thesharesofcityandruralcommercialbankshavenotincreasedfurther.
34 ADBEastAsiaWorkingPaperSeriesNo.54(vi)Banks’profitability—asmeasuredbythenetinterestmargin—hasdeclinedoverthepastdecade.Lowdepositratescouldnotpreventthisdecline,whichwasdrivenbyfallingbanklendingrates(Figure39).Againstthisbackdrop,thefollowingrecommendationscanbemade:(i)Strengthenequityandbondfinancing.Encourageequityfinancingandbondissuanceasawayofcorporatefinancing,especiallyforlargercompanies,whilecontinuetoreininshadowbanking.(ii)Shiftbusinessloanstotheprivatesector.Increasetheshareofloanstotheprivatesector,whilereducingthatofSOEs.Todoso,removeimplicitguaranteesforSOEsandendtheirpreferentialborrowingconditions(IMF2019).(iii)Strengthenlendingtomicroandsmallenterprises.Increasetheshareofbusinessloanstomicrobusinessesandsmallcompanies,especiallythosethatgothithardbytheadverseimpactoftheCOVID-19pandemicontheeconomy.(iv)Movetolonger-termfinancing.Ahighshareofshort-termloanspotentiallyleadstohightransactioncosts,whilebanksshouldconcentrateonimprovingtheirabilitytoassesscreditrisks(ADB2021c).(v)Improvebanks’efficiency.Loweringhouseholddepositrateswilllikelynotbeasolutionforfallingnetinterestmarginsofbanks,butcouldspuranotherroundofpropertyinvestment.Improvingbanks’(operational)efficiencyisabetteroption(ADB2021c).TimeframesregardingtheimplementationoftheserecommendationscanbefoundinAppendix6.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 35Box3:StructuralChangeOverthepast4decades,theeconomyofthePeople’sRepublicofChina(PRC)hasshiftedawayfromagricultureandtowardservices,withstronggrowthinthelattersectorinthe2010s(FigureB5).Atthesametime,theshareofindustryjuststartedtodeclineinthepastdecade,andeventhenshowedsignsofstabilizationinrecentyears.Theseshiftsarelinkedtothegrowthmodeloftheeconomy.Effortsweremadetoincreasedomesticdemand,especiallyaftertheglobalfinancialcrisisthatendedexport-ledgrowth—thePRCranacurrentaccountsurplusofalmost10%ofgrossdomesticproduct(GDP)in2007.However,inrecentyearstheshifttoserviceslostdynamics.ServicescouldonlyimprovetheirshareinGDPmarginallyfrom52.4%in2016to53.3%in2021ashouseholdconsumptionstagnatedasashareinGDP(FigureB6).FigureB5:SectorShareinGDP,FigureB6:HouseholdIncomeand1980–2021Consumption,andInvestment,1980–2021Agriculture Industry Services P DG PercapitaconsumptionPercent of Disposableincomepercapita Grossfixedcapitalformation GDP=grossdomesticproduct.GDP=grossdomesticproduct.Source:NBSviaCEICDataCompany(accessed13JulySource:NBSviaCEICDataCompany(accessed14July2022).2022).Inthefirsthalfofthe2010s,householdconsumption(asashareinGDP)pickedup,whichcoincidedwithincomeinequalityimprovingandwagesinlow-skillsectorsoutgrowingthatinhigh-skillones(Peschel2021).However,theshareofhouseholdconsumptioninGDPhasnotincreasedfurtherinrecentyears,butdefactostagnated(38.7%ofGDPin2016versus38.4%in2021)reflectingastagnatingdisposableincomeshare(61.6%ofGDPin2016versus61.7%in2020).Facilitatingstructuralchangefromindustrytoservicesrequiresonthedemandsidestrongerhouseholdconsumption,forwhichincreasesinhouseholdincomeiscrucial.Inthiscontext,moreaffordablehousingcouldsupporthouseholdconsumption.Second,thereisaneedtoreduceincomeinequalitytoboosthouseholdconsumption.Third,basicpublicservicesandsocialsecurityneedtobeimprovedtoreducetheneedforprecautionarysaving.Thisrequireshighertransferstoless-developedregionsandadditionalgovernmentspendingtocopewithneedsofarapidlyagingsociety(Peschel2021).Onthesupplysidethegovernmentwouldneedtoputgreateremphasisontheservicesectorineconomicplanning.Currently,effortsinindustrialpolicygounmatchedintheservicesectorthatwouldbenefitfromasimilarstatustomanufacturingintermsoffiscalincentives,resourceallocation,andopenness.The14thFive-Year-Plan(2021–2025)envisageslessrelianceonheavyindustriesbyproposinganewtargettoincreasetheshareofthestrategicemergingindustries(i.e.,advancedmanufacturing,includinghigh-endmachineryandequipment,advancedmaterials,andelectricvehicles)from11.5%ofGDPin2019tomorethan17%ofGDPby2025(ADB2021d).Whilepreviousgrowthinserviceswasconcentratedinfinancialservicesandrealestate,thereispotentialforthefurtherdevelopmentofthehighervalue-addedservicessuchasresearchanddevelopment,informationandcommunicationtechnology-relatedservices,andconsumerservices(WorldBank2021).
36 ADBEastAsiaWorkingPaperSeriesNo.54E.TotalFactorProductivityOverviewThissectiondiscussespoliciesinareaswhicharecoveredbytheexplanatoryvariablesinthemodel:(i)imports—which,inabroadersense,relatestotradeopenness;(ii)FDI;(iii)sectorallaborshift;and(iv)R&D.Asmentionedabove,thebiggestinfluenceonfutureTFPgrowthhasR&D,followedbythesectorallaborshift.ImportsandFDIletaneconomybenefitfromimportedtechnology.ThesectorallaborshiftinthePRCispurelydomesticasimmigrationcanbeneglectedbysize.Finally,thoughR&Dinvestmenttakesplaceinthecountry,researchfrequentlyhaslinkageswiththerestoftheworld.Havingsaidthis,thissectioncannotexclusivelyfocusondomesticpoliciesbutwillalsoneedtotakeintoaccount,totheextentpossible,howthePRCpositionsitselfinaninternationalcontextwithrespecttotrade,FDI,andR&D.Inthefollowing,wewilllookattheseissues,butbeforewebrieflydiscusspoliciessupportingthesectorallaborshift.SectorallaborshiftTheshareofindustryinGDPhasdeclinedsubstantiallyinthedecadeaftertheglobalfinancialcrisis(Box3).Asseeninthemodelabove,thespeedofmigrationfromagricultureintoindustryandservicesisexpectedtomoderateovertime.Thisisbecausetheshareofworkforceemployedinagriculturehasalreadyfallensubstantially.Also,industryhasstartedsheddingworkforcesothatthemigrationroutefromruralareastocomparativelywell-paidindustryjobshasnarrowed.Thus,rural–urbanmigrationinthefutureismorelikelytotakeplaceintolow-skilledservicesectorjobs,whichdependondomesticconsumptionandtendtobemorevolatile.TradeopennessandforeigndirectinvestmentImportsarefrequentlyusedfortheproductionofexportsinadvancedeconomiesexhibitingahighleveloftradeopenness,whichismeasuredasthesumofmerchandiseexportsandimportsaspercentageofGDP.WhilethePRC’stradeopennesshasrisensharplyintheearly2000supto2007,reachingalevelsimilartoGermanyandtheRepublicofKorea,ithasfallenbacksinceand,by2021,wasmuchclosertoJapanandtheUS(Figure42).Figure42:TradeOpenness,1980–2021Figure43:ProcessingTrade,1993–2021PRCGermanyProcessingexports JapanRepublicofKorea ProcessingimportsUnitedStates t ofGDPPercen GDP=grossdomesticproduct,PRC=People’sRepublicofChina.Note:TradeopennessmeasuredassumofmerchandiseNote:ProcessingexportsandimportsasashareintotalexportsandimportsaspercentofGDP.merchandiseexportsandimports,respectively.Source:WorldBank,WorldDevelopmentIndicatorsSources:GeneralAdministrationofCustomsviaCEIC(accessed14September2022).DataCompany(accessed14July2022);andADBcalculations.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChina 37InthecaseofthePRC,processingtradehasdeclinedoverthepastdecade(Figure43).ThisdevelopmentwasdrivenbythePRCestablishingdomesticmanufacturingvaluechainsandthusbecominglessdependentonimportsofintermediarygoodsformanufacturingexports.Also,wagesinthePRCroserapidlysothatprocessingtradebecamelessattractiveforforeigncompanies.FDIinflowintothePRC,measuredaspercentageofGDP,hasdeclinedoverthepast2decades,whilethecountry’srestrictionsonFDIhaveremainedhighcomparedtoOECDaswellasnon-OECDcountries(Figures44and45).Figure44:FDINetInflowintothePRC,Figure45:FDIRestrictivenessIndex,2000–20202003,2006,2010,2015,and2020FDInetinflow PRCOECDaverage Non-OECDaverage DP G of FDI=foreigndirectinvestment,GDP=grossdomesticFDI=foreigndirectinvestment,OECD=Organisationproduct,PRC=People’sRepublicofChina.forEconomicCo-operationandDevelopment.PRC=People’sRepublicofChina.Note:Non-OECDcomprises46economies.Sources:WorldBankviaCEICDataCompany(accessed2Source:OECDDatabase,https://stats.oecd.org/Index.December2022);andADBcalculations.aspx?datasetcode=FDIINDEX(accessed5September2022).ResearchanddevelopmentOnechallengeforthePRC’sR&Dpolicyismisuseoffunds.AccordingtoestimatesbyBoeingandPeters(2022),usingPRCfirm-leveldatafor2001–2011,42%ofgranteesmisappropriatedR&Dsubsidies(accountingfor53%oftotalsubsidies).Thishasreducedtheeffectivenessofthecountry’sR&Dpolicybymorethanhalf(BoeingandPeters2022).ThoughthegapinR&DspendingbetweenthePRCanddevelopedeconomiesisexpectedtonarrowoverthenextdecades,expenditureonbasicresearch(as%ofGDP)inthePRChasremainedlowsofar.In2020,thePRCspent0.14%ofGDPonbasicresearch,lessthantheUS(0.5%ofGDP),Japan(0.4%ofGDP),ortheRepublicofKorea(0.7%ofGDP)(OECD2021).Furthermore,inthePRC,enterprises’basicresearchinvestmentremainslowascompanieswouldfocusondevelopingnewtechnologiesandproductswithclearcommercialapplicationinsteadoftechnologiesthatmightbecomecommerciallyviableinthelongterm(Zheng,Zhuang,andWang2020).Zheng,Zhuang,andWang(2020)seeinsufficientintellectualproperty(IP)rightsprotectionasacontributingfactortothelowqualityofinnovationinthePRC.AspunishmentforIPinfringementisinsufficient,companieswouldratherfocusonimitationandapplicationactivitiesthanonbasicresearchandinnovativeproducts,whichiscostlywhenIPinfringementmakeitdifficultforcompaniestoreaptherewardsfortheirinvestment.Second,sincecompanieswouldoftensplitapatentintomultiplesmall
38 ADBEastAsiaWorkingPaperSeriesNo.54ones,limitedpatentreviewresourceswouldbeconsumedbylow-qualitypatents(andhigh-qualityoneslosingout).Finally,therewouldbeanotablegapinIPprotectionacrossregionsinthePRCwithdifferencesinthelevelofeconomicdevelopmentandmarketoperationefficiency.LocalprotectionismwouldresultinunequalIPprotection,withIPinfringementinless-developedregionslikelytofacelessliabilityissuesthaninadvancedregions.PolicyrecommendationsAgainstthebackdropofalessdynamicsectorallaborshiftinthefuture,thedeclineinthecountry’stradeopennessformorethanadecade,persistingrestrictionsonFDIinmanyareas,andongoingchallengestoimprovethequalityofR&DinthePRC,thefollowingpolicyrecommendationscanbemade:(i)Loosenresidencepermitrestrictions.Looseninghukou(residencepermit)restrictionscouldoverallhelpimprovelaborallocation(Lam,Liu,andSchipke2015).Also,theaccessoflabormigrantstohealthservicesandeducationfortheirchildrenincitieswheretheyliveneedstobeimproved.Furthermore,theportabilityofsocialsecuritybenefits,suchaspensionsandhealthinsurance,shouldbeincreased.(ii)Facilitatenewtradeandinvestmentagreements.Lookingforward,thePRC’scommitmentsinthecontextofnewtradeandinvestmentagreementscanhelpincreaseopennessandcouldalsofacilitatedomesticreforms(IMF2022aand2022b).Whilethereisanintentiontoadvanceinnovationinmultipleareastobecomelessdependentonforeigntechnology,itisrecommendedtospecializeinanumberofareas,whilesourcinggloballyinothers—anapproachpracticedbymostadvancedeconomies(ADB2021d).ThisapproachleavesroomforimportsintothePRC,makingthecountryatradingpartnerforothercountries.(iii)Loosenrestrictionsonforeigndirectinvestment.TopotentiallyincreaseFDIinflow,loosenrestrictionsonsuchinvestments,e.g.,bysubstantiallyrevisingandshorteningthenegativelistforFDI.Anotherpointisreducingjointventurerequirements.SuchrequirementsarefrequentformultinationalfirmsseekingFDIinthePRCtopartnerwithadomesticcompanyandformaninternationaljointventure,whicharecharacterizedbyexplicittechnologytransferrequirements(Jiangetal.2019).(iv)StrengthenIPrightsprotection.TheanalysisaboveshowsthatstrengtheningIPrightsprotectioniskeytoimproveincentivesforcompaniestoinvestinbasicresearchandinnovativeproducts.IPprotectionstandardsneedtobeunifiedacrossthecountryandpatentreviewstrengthened.Tothisend,thecountrywouldneedtoinvestinanIPrightsprotectionregimeinlinewithinternationalbestpractices,includingitsinstitutionalrequirements.Whilethiswouldnotonlyhelptoincreasethequalityofpatents,itwouldalsobeasteptowardincreasingcompanies’incentivestoputinnovationbeforeapplication.AtablesummarizingthepolicyrecommendationsmentionedinthischaptercanbefoundinAppendix6.
V. OUTLOOKONPOLICYCHALLENGESANDREFORMSThepreviouschapterhaspresentedanarrayofreformspertainingtostrengtheningthelaborforce,humancapitalformation,capitalandcreditallocation,andmeasurestoenhanceTFPgrowth.Theseareasrelatetobiglong-termchallengessuchasdemographicaging,theurban–ruralgap,improvingthecapitalandcreditallocation,andshiftingthecountry’sgrowthmodeltowarddomesticsourcesofgrowth.Finally,effortstoadvancedomestictechnologydependonsuccessfulR&Dpolicies.Demographicagingisanissuethatmustbeaddressed.Somereforms,suchasanincreaseofsupportforfamilies,willonlyshoweffectinthelongrun.Therefore,mitigationmeasuresareneeded,suchasextendingtheworklife,which,besideschangestoretirementrules,includesimprovinghealthcaresothatemployeesstayhealthy.Atthesametime,raisingtheretirementageisunpopularbutnecessarygiventherapidagingofsocietyandrisinglifeexpectancy.Anagingsocietywouldbenefitfromincreasedlabormobilitytobetterallocatetheshrinkingworkforce.Hukourestrictionsshouldbeloosened.Inaddition,thenearly200millionmigrantworkers(andtheirfamilies)wouldneedbetterandeasieraccesstothebasicpublicservicesandeducationinthecitieswheretheywork.Furthermore,intermsofeducation,stepsareneededtoimprovethequalityoftertiaryeducationandon-the-jobtraining.Besidesimprovementsineducation,thementionedreformsrequirechangestoandthestrengtheningofsocialpolicies.Theurban–ruralgapputsaburdenonhumancapitalformationandrequiresattentionbypolicymakers.Increasedeffortstoboosteducationinruralareasareneeded,asarestepstoexpandearlychildhoodeducation.Suchreformswouldrequirehigherfiscaltransferstopoorerregionsandneedtobeembeddedinbroaderreformsofintergovernmentalfiscalrelations(Peschel2021).Thecapitalandcreditallocationmustbeimproved.Measuresareurgentlyneededtoimprovecapitalallocationandefficiency,therebyboostingpotentialgrowth.Asanalyzedabove,thePRC’sinvestmentshareinGDPwillhavetomoderateasthelevelofdebthassharplyrisenoverthepastdecadeandfurthermassiveinvestmentinthehousingmarketandthecountry’sinfrastructurewouldnotbesustainable.SOEreformsareneeded.Continuouslyloss-makingSOEswillneedtoexitthemarkettoavoidazombificationoftheeconomy.Inthecomingyears,whentheworkforcewillstarttoshrinkquickly,thereareever-fewerargumentstocontinuesustainingloss-makingSOEs.However,laid-offworkerswouldneedhelpandsupporttofindandbequalifiedfornewjobs,whichrequirestemporaryfinancialsupportand(re)training.Creditwillneedtoincreasinglybeallocatedtotheprivatesector,inparticulartomicroandsmallenterprises.Tothisend,bankswillneedtoadjusttheirlendingpreferences,whichrealisticallyrequiresmovingawayfromimplicitguaranteesforSOEsandtheirprivilegedaccesstocredit.Suchreformsnotonlytaketimetoimplement,buttheyalsorequirewillingnesstoadvancethem.Thesereformsmightalsohave(unwanted)sideeffects,suchassomeSOEsgoingoutofbusiness.Furthermore,thebankingsectorwillneedtodevelopandimprovecreditriskassessmentcapabilities.Insum,reformstoraisecapitalefficiencyrequirebothhighwillingnessandefforts.Atthesametime,theywouldhaveamajorimpactonpotentialgrowth.Withadecliningshareofindustryandarisingshareofservicesintheeconomy,growthshouldbedrivenlessbyinvestmentandmorebydomesticconsumption,fueledbyanincreaseinhouseholdconsumption—forwhichimprovementsinsocialpoliciesareneededtoreducetheneedforprecautionarysaving.
40 ADBEastAsiaWorkingPaperSeriesNo.54AnotherareaofreformisimprovingthedomesticR&Defficiency.Asseenintheanalysisabove,R&Dreformshingeuponcomplexinstitutionalchangesandenforcementissues.However,improvingIPprotectionseemstobeacrucialstepforchangingincentives(andpotentiallybehavior)thatareneededtoimprovethequalityofdomesticR&D.Furthermore,R&D(andTFP)benefitsfromaccesstoforeigntechnologyandanopenexchangeofideaswiththerestoftheworld.Asforthepacingofreforms,itmustbenotedthatsomereformsdependonmanybuildingblocksthatrequirecoordinatedeffortsovermanyyears(seeAppendix6).OneshouldalsonotethatimprovingpotentialgrowthwilllikelymeanlessgovernmentinfluenceastheSOElandscapewillneedtobestreamlined,abiggershareofcreditwillneedtogototheprivatesector,andhouseholdswillneedgainspendingpowertofacilitatethetransitiontoaconsumption-basedeconomy.Toachievethelatter,thegovernmentwillneedtostrengthensocialpoliciesandredistribution,requiringadditionaldomesticresourcemobilizationandbroaderreformstointergovernmentalfiscalrelations.WhiletherehavebeenseveralchallengestothePRC’seconomicdevelopmentinthepast4decades,therewerealsomanystrongtailwinds,suchasthedemographicdividendandthecountry’saccessiontotheWorldTradeOrganization.Asdemographicshasturnedtoheadwinds(andwillincreasinglydosointhefuture)anddebthasincreasedsubstantiallyafter2008,comprehensivedomesticreformshavebecomeapressingissueaswellasabigchallenge.ThestrengthofthePRC’seconomicdevelopmentmodelhasbeenflexibility,combinedwiththewillingnesstolearnandtoadjust.Withthisinmind,initiatingcomprehensivedomesticreformsiskey—adjustmentscanbedonealongtheway.
APPENDIXESAppendix1:Barro’sEstimateofConvergenceBarro(2016and2017)looksingreaterdetailatconditionalconvergenceforthePeople’sRepublicofChina(PRC).Fittingamodelthatconsidersseveralfactors,51BarroestimatesthepasteconomicgrowthratesforthePRCandcomparesthemtotheactualgrowthrates(FigureA1),beforeprovidingexplanationsfortheobserveddeviationsfromtheforecastvalues(TableA1).FigureA1:RealPerCapitaGDPGrowthinthePRC:Barro’sEstimateofConvergence,1960–2020 PercapitagrowthrateFittedvaluePercent – – – – – – – – – – – – GDP=grossdomesticproduct,PRC=People’sRepublicofChina.Sources:Barro(2017);NBSviaCEICDataCompany(accessed07June2022);andADBcalculations.51Factorsincludelifeexpectancyatbirth,totalfertilityrate,indicatorsofruleoflawanddemocracy,investment-to-GDPratio,government-consumption-to-GDPratio,femaleandmaleaverageyearsofschoolattainment,(trade)openness,andinflation.
42 Appendix1TableA1:Barro’sExplanationofthePRC’sPerCapitaGrowthPatternPeriodDeviationBarro’sExplanation(percentagepoints)1960–1965-2.9ThemodelomitssomeaspectsofthePRCcommunistsystemthatwereadversetoeconomicgrowththroughthelate1970s,e.g.,theGreatLeap1965–1970-2.2Forwardof1958–1963andtheCulturalRevolutionof1966–1976.1970–1975-3.31975–1980-2.41980–19852.1Theconvergenceeffectisoffsetbychangesintheexplanatoryvariables:e.g.,thepredictedgrowthraterisesthroughoutbecauseofhigherlife1985–1990-4.8expectancy,through2000becauseoflowerfertility,1990–19951.3through2005duetogreaterinternationalopenness,through2010becauseofahigherratioofinvestmenttoGDP,1995–2000-2.6andupto1995–2000duetobetterlawandorder.2000–20051.82005–20101.7After2005,theconvergenceforcestrengthensduetothesharplyrisingpercapitaGDPandpartlybecausetheexplanatoryvariablesprovideless2010–20151.5ofanoffsettingeffect.Theoffseteffecteventuallytendstobecappedbecauseoflimitstorisinglifeexpectancy,fallingfertility,risinginterna-2015–20200.8tionalopenness,improvingruleoflawandeaseofdoingbusiness,etc.GDP=grossdomesticproduct,PRC=People’sRepublicofChina.Note:Deviationisthedifferencebetweenactualandfittedvalue(inpercentagepoints).Sources:Barro(2017),withauthors’additionofthedeviationforthe2015–2020period.
Appendix2 43Appendix2:HumanCapitalIndexConstructingthehumancapitalindexFollowingBailliuetal.(2016),weconstructahumancapitalindexforeachyeartindexas(1),wheresisthetotalyearsofschoolingandφ(s)measurestheefficiencyofaunitoflaborwithsyearsofeducationrelativetoonewithoutanyschooling.Foreachyeart,thehumancapitalindexisconstructedbycombiningeducationalattainmentdifferentiatedbyagegroupsandtheirrespectiveshareinpopulationasfollows:Equation(2)isthemeasureofeducationforeachagegroup:(2),whereNsistheshareofpopulationwithnoeducation.Tp,Ts,andTtaretheshareofagegroupiwithiiihighesteducationattainmentlevelofprimary,secondary,ortertiary.Cp,Cs,andCtaretheshareofiiipopulationofagegroupiwhohascompletedhighesteducationattainmentlevelofprimary,secondary,ortertiary.Pisthenumberofpeopleinagivenagegroupandpisthenumberofpeopleinallagegroups.itReturnstoeducationareassumedtobeconstantovertime.Equation(3)istheshareofpopulationineachagegroupintheoverallpopulationinyeart:(3).Substitutingequation(2)intoequation(3),andthen(3)into(1),weobtainthehumancapitalindex.EducationalattainmentFortheagegroup15–24,totalandcompletedprimaryschoolattainmentdatafrom2016to2040isassumedbasedonhistoricaltrend.Fortheolderagegroups,weknowtheirprimaryschoolattainmentratesandcarryforwardtheratesastheyage.Accordingtothe14thFive-YearPlan(2021–2025)forNationalEconomicandSocialDevelopmentandtheLong-RangeObjectivesThroughtheYear2035,thegrosstertiaryenrollmentinthePeople’sRepublicofChina(PRC)isexpectedtoreach60%by2025(NDRC2021).Furthermore,thePRC’sEducationModernization2035Planstatesthatthecountrytertiaryenrollmentshouldincreaseto65%in2035(StateCouncil2019).Given58.4%grosstertiaryenrollmentin2020,weassumeenrollmentforyearsbetween2021and2025inawaythatitreaches60%by2025.Assumingaconstantgrowthrateof0.5percentagepointsperyearfor2026–2040,weprojectgrosstertiaryenrollmentratetobe67.5%in2040(seeFigure10,maintext).52Fortotalandcompletedtertiaryschoolattainment,wesimplifyandassumeitwillfollowthesamegrowthrateasenrollmentratebutwitha4-yearlag(drop-outratesareverylowfortertiaryeducation).52Afterithasrisensharplyinthepastdecade,tertiaryenrollmentwillnowflattenasruralareascontinuelaggingbehind.
44 Appendix2TotalsecondaryschoolattainmentcanbeestimatedastheresidualofNs+Tp+Ts+Tt=100,whereNsistheshareofpopulationwithoutformaleducation,whichisforecastbasedonhistoricaltrend,andTp,Ts,andTtaresharesofpopulationthatattendedprimary,secondary,andtertiaryeducation,respectively.Foragegroup15–24,secondarycompletionrateisassumedtofollowthegrowthrateofthetotalsecondaryattainment.Fortheotherfouragegroups,i.e.,25yearsandolder,thehighest-levelschoolattainmentiscarriedforwardfor2015–2040frompreviouslevelsastheyage(asreturningtotertiaryeducationlaterinlifeisrareinthePRC).
Appendix3 45Appendix3:ComparisonofResultsWhileHiggins(2020)andRobertsandRussell(2019)onlyprovideaggregatedresults,Bailliuetal.(2016)providedetailsoftheirestimates.Here,wecomparethemtoourfindings.For2016–2019,ourestimateofthecapitalcontributiontogrowthishigherthantheforecastinBailliuetal.(2016),whileourestimateoftotalfactorproductivity(TFP)islowerthantheirs(TableA3.1).Also,ourestimateofhumancapital,usingthelatestBarro-Leedataset,yieldsalowercontributionofhumancapitaltogrowthfor2001–2019TableA3.1:ComparisonofPotentialGDPGrowthEstimateswithBailliuetal.(2016)ActualBailliuGDPAuthors’etal.Periodgrowthestimate(2016)ContributionKLHCTFPKBailliuLBailliuHCBailliuTFPBailliu2001–20059.8%10.4%10.5%5.2%4.9%0.9%0.9%0.3%0.9%3.9%4.0%2006–201011.3%10.3%11.3%5.6%6.0%0.3%0.4%0.1%0.7%4.3%4.1%2011–20157.9%8.2%7.6%4.8%5.5%-0.1%-0.1%0.6%0.7%2.9%1.5%2016–2019/2020*6.6%6.8%7.0%5.1%3.3%-0.1%-0.1%0.5%0.6%1.3%3.3%2001–2019/2020*9.0%9.0%9.8%5.2%5.5%0.4%0.4%0.4%0.8%3.2%3.2%GDP=grossdomesticproduct,K=capital,L=labor,TFP=totalfactorproductivity.*Upto2019forourestimatesandupto2020forBailliuetal.(2016),usingtheirforecastvalues.Sources:Bailliuetal.(2016);andauthors’estimates.Fortheoverlappingforecastperiod(2020–2030),weestimatealowercontributionfromTFPtogrowth,whileinourviewthecontributionofcapitalishigherin2020–2025—theshareofgrossfixedcapitalformationhasnotfallenasfastasexpectedinBailliuetal.(2016)—butclosetoBailliuetal.(2016)for2026–2030(TableA3.2).Whileourestimatesarequitesimilarforlaborandhumancapital,thefuturetotalfactorproductivitygrowthisestimatedmuchhigherbyBailliuetal.(2016).Weassumethattheshareofemploymentinindustrywillcontinuetofall,resultinginasmallerfuturecontributionfromsectorallaborshifttoTFPgrowth.Also,weassumealowercontributionfromresearchanddevelopmentspendingonTFPgrowthfrom2020sothatourestimatedTFPcontributiontogrowthisonlyhalfoftheirs.TableA3.2:ContributionstoPotentialGDPGrowth,2020–2040Authors’Bailliuetal.Periodforecast(2016)ContributionKLHCTFPKBailliuLBailliuHCBailliuTFPBailliu2020/2021–2025*5.3%5.9%3.4%2.6%-0.2%-0.3%0.5%0.5%1.7%3.0%2026–20303.5%4.9%2.1%1.9%-0.5%-0.5%0.4%0.6%1.5%2.9%2031–20352.7%1.5%-0.5%0.6%1.1%2036–20402.0%1.2%-0.4%0.2%1.0%GDP=grossdomesticproduct,K=capital,L=labor,TFP=totalfactorproductivity.*2020-2025forauthors’forecastand2021–2025forBailliuetal.(2016).Sources:Bailliuetal.(2016);andauthors’estimates.
46 Appendix4Appendix4:EstimatingCapitalStockGrowthTheestimateofthegrowthrateofthecapitalstockdeterminesthecontributionofcapitaltogrossdomesticproduct(GDP)growthinthemodel.Thecapitalstockgrowthrateisfrequentlysensitivetodeflatorchoice.Itcouldalsobesensitivetothedefinitionofcapitalstock,inparticularifhousingisincludedornot.Deflatorchoice.Toconstructatimeseriesinconstant2000CNY,Bailliuetal.(2016)firstchoseastartingvalueforthecapitalstock.Then,theyusethecontributionofinvestmenttoGDPgrowthtoestimateinvestmenteachyear,anddeductdepreciationtoobtainacapitalstockestimateanditsgrowthrate.Anotherrouteistoestimateinvestmentviagrossfixcapitalformation—asdoesIMF(2021a).Here,thecapitalstockestimateisavailableincurrentCNYonly.Possibledeflatorchoicesarethefixedassetinvestment(FAI)deflatorandtheGDPdeflator.Accordingtoourcross-check,contributionsofcapitaltogrowthforpastyearschangedependingonthedeflator.ByusinganFAI-deflatedcapitalstock,thecontributionofcapitaltogrowthincreases,whilethecontributionoftotalfactorproductivity(TFP)togrowthfalls.53However,thecontributionofcapitaltogrowthderivedwouldbetoohighusingtheFAIdeflator.Housingstock.Separatingthehousingcomponentfromthecapitalstockisachallengeastheshareofpropertyinvestmenthasnotbeenstableovertimeandthedepreciationratesofhousingandothercomponentsofthecapitalstockdiffer.Herd(2020)providesanestimationofcapitalstockinfoursectors(housing,infrastructure,government,andbusiness)from1953to2016.Theshareofhousingintotalcapitalstockhasfirstfallenintheearly2000sandthenincreasedafter2010(Figure13,maintext).Inaddition,thereareseveralotherfactorsmakingitdifficulttoestimatethefuturehousingstock:(i)theshareofhousinginoverallinvestmentusedtodifferovertime(RobertsandRussell2019),(ii)demographicssuggeststhattheneedfornewhousingconstructionwilldecline,and(iii)recentgovernmentmeasuresaimedatcurbinginvestmentgrowthinthepropertysector,possiblywithlastingimpact.Bailliuetal.(2016)foundthattheslopeofthecapitalstockwithhousingissteeperin,suggestingthatthegrowthrateofthecapitalstockwithhousingishigher,whichimpliesthatthestockofhousinghasgrownfasterthantherestofthecapitalstock.BasedonthedatasetbyHerd(2020),wefindthattheaverageannualgrowthrateofthecapitalstockincludinghousingisonlyminimallyhigherthanthatwhenexcludinghousingfor2001–2016(12.75%versus12.69%),implyingthatgrowthinthehousingstockwasnotmuchfaster.Thus,thechoiceofdeflatorhasamuchhigherimpactonthegrowthrateofthecapitalstock,andthusonthecontributionofcapitaltogrowththanexcludingthehousingstockornot.Higgins(2020)reachedasimilarconclusionbycross-checkinghisresultsusingdifferentPennWorldTabledatasets.53SeeHiggins(2020)forarelateddiscussion.
Appendix5 47Appendix5:ImprovingtheQualityofEducationThequalityofeducationcouldberaisedbyexpandingearlychildhooddevelopment,changingthefocusoftheeducationsystem,strengtheningtechnicaltraining,andpromotinglifelonglearning.Earlychilddevelopment.Educationandnutritioninthefirstyearsoflifehavelifelongeffectsonthebrainandbody(WorldBankandDRC2013).Withrespecttoeducation,the2015ProgramforInternationalStudentAssessmentscienceresultsshowthatstudentsinthePeople’sRepublicofChina(PRC)whoreceivedearlychilddevelopment(ECD)educationperform2.5yearsaheadofthosewithout(WorldBankandDRC2019).However,earlychildhoodeducationisyettobeincludedinthePRC’scompulsiveeducationsystem.Thecountry’sinvestmentinECDeducationwasonly0.4%ofgrossdomesticproduct(GDP)in2016,comparedwith2.3%onaverageacrossOrganisationforEconomicCo-operationandDevelopment(OECD)countries(WorldBankandDRC2019).DespitethePRC’searlychildhoodeducationenrollmenthavingincreasedfrom56.6%in2010to83.4%in2019,thecountrystilllagsbehinditsneighbors,suchasJapan(at92%)andtheRepublicofKorea(at94%)(ADB2021e).Inaddition,Lietal.(2017)pointoutthatover30%ofchildrenaged0–3bornandraisedinthePRC’sruralareasexhibitcognitivedelayduetolackofmodernparentingandstimulation.Therefore,thecountryshouldprioritizeearlychildhoodeducationandhealthprograms,especiallyinruralareas(includinglowertuitionfeesforpreprimaryschoolsandbetterruralpreprimaryschooleducation).Educationsystem.ThePRC’seducationsystemisconsideredasexam-driven,whichcanprofitthedevelopmentofpersistenceandself-control.However,itmayinhibitintrinsicmotivationandstiflecreativity,whichcouldbeaddressedbyshiftingtomorebalancedcurriculaandpayinggreaterattentiontodevelophigher-orderskills(WorldBankandDRC2019).Inturn,thecountrycouldbenefitfromamoreinnovativeandcreativelaborforce,whichwouldsupportmovinguptheglobalvaluechainandhelpacceleratethepaceofnewtechnologyadoption.Sincetheexpansionofcollegeeducationin1999,thePRChaswitnessedasurgeintertiaryschoolenrollmentrates.Lietal.(2017)arguesthatthefastexpansionoftertiaryschoolenrollmentrateswasfollowedbyadeclineinqualityofthecountry’scollegeeducation.Loyalkaetal.(2016)assessessamplesofcollegestudentsinengineeringfromthePRC,theRussianFederation,andtheUnitedStatesbytestingtheiracademicandthinkingskills.TheassessmentshowsthatPRCstudentsperformwellinstandardizedtests,butshownofurthergrowthincognitiveskillsafter2yearsofcollege.Thus,whilegraduationratesarerising,thequalityoftertiaryeducationshouldnotbeneglected.Technicalandvocationaleducationandtraining.ThoughthePRChasthelargestvocationaleducationsystemintheworldbysize,withover11,000vocationalschoolsand10milliongraduateseachyear,thesystemhasyettomeetlabormarketdemands.Thefocusofthetechnicalandvocationaleducationandtraining(TVET)systemneedstobemarket-oriented,andtheschool–industrylinkmustbestrengthenedtoraiseitsattractivenessforhighschoolgraduates(WorldBankandDRC2019).Moreover,atcurrentstage,TVETandacademicstudiesareseenastwoseparatesystemsinthePRC.Pathwaysbetweenthosetwoshouldbeopened.Lifelonglearning.Finally,itisessentialforworkerstocontinuelearningthroughouttheirworkinglife.Theeducationandtrainingreceivedrightbeforeenteringtheworklifeandduringthefirstyearsinthejobareunlikelytosufficethroughoutworklife.Thus,companieswillhavetoprovideon-the-jobtrainingopportunities.However,thelackofcapacitybyandincentivesforsmallandmedium-sizedenterprises(SMEs)toofferthiskindoftraining,givenahighlaborturnover,couldbeaddressedbyproviding“trainingcontracts”thatwouldrequireworkerstostayinthejobforacertainperiodorpaybackthetrainingcost(WorldBankandDRC2019).Additionally,increasingpublicfundingforjob-relatedtrainingcanhelpretrainworkersandupgradetheirskills.
48 Appendix6Appendix6:SummaryofPolicyRecommendationsThefollowingtablesummarizesthepolicyrecommendationsmadeinpartIV.TableA6:SummaryofPolicyRecommendationsAreaMeasureTimeFrameforImplementationLaborGraduallyraisethelegalretirementageShorttermFocusonraisinglaborparticipationinthemid-50sShorttomediumtermImprovehealthcareandpromotinghealthyaging,focusingonsafeMediumtolongtermandsecureworkingenvironment,occupationalhealthservices,andoccupationaldiseasepreventionRaisefemaleworkforceparticipationbystrengtheningequalMediumtolongtermemploymentopportunities,increasingmaternityleave,andimprovingsupportforchildcare,elderlycare,andsinglemothersUndertakeeffortstoincreasethebirthrateLongtermIncreaselabormobilitybyfurtherrelaxingthehouseholdregistrationMediumtermsystem(hukou),andimproveaccesstosocialservicesforlabormigrantsandtheirchildrenHumanCapitalImproveruraleducation,especiallyuppersecondaryattainmentandMediumtermtertiaryeducationExpandearlychilddevelopmentMediumtolongtermImprovethequalityoftertiaryeducationMediumtolongtermStrengthentechnicalandvocationaleducationandtraining(TVET)MediumtermIncreaseon-the-jobandlifelongtrainingopportunitiesShorttermCapitalState-ownedClarifythescopeandfunctionofSOEs.TheareasinwhichtheShorttermenterprisesgovernmentwantstoretainSOEs,shouldbesmallandwell-specified.(SOEs)StrengthenthemarketeconomybymakingmarketentryeasierforMediumtolongtermprivatecompanies.Thisincludestoletnonessentialloss-makingSOEsleavethemarket.Monopoliesshouldbekepttoaminimum.LeveltheplayingfieldfortheprivatesectortocompetewithSOEs,Mediumtolongterminparticularintermsofaccesstolandandcredit.ThiswouldincluderemovingimplicitguaranteesforSOEsandtheirpreferredaccesstobankcredit(moreonthatbelow).SeparatesocialfunctionsfromSOEssothateconomiesofscaleandShorttomediumtermscopecanbebetterrealizedundersuchseparation.ImproveSOEmanagement.ThemanagementandsupervisionofMediumtermSOEsshouldbestrengthenedtoraisetheirefficiency.
Appendix6 49AreaMeasureTimeFrameforImplementationCreditStrengthenequityandcorporatebondfinancing,especiallyforlargerShorttomediumtermcompaniesIncreasetheshareofbusinessesloans,inparticulartomicroandsmallMediumtolongtermenterprises(MSEs)Increasetheshareofloanstotheprivatesector(insteadofSOEs)MediumtolongtermRemoveimplicitguaranteesforSOEsandendtheirpreferentialMediumtermborrowingconditionsMovetolonger-termfinancing,whilebanksshouldconcentrateonShorttomediumtermimprovingtheirabilitytoassesscreditrisksImprovebanks(operational)efficiencyMediumtermStructuralFocusmoreonservicesectordevelopmenttofacilitatestructuralShorttomediumtermchange(Box3)changefromindustrytoservices.AddressincomeinequalitytoboosthouseholdconsumptionMediumtolongtermStrengthenbasicpublicservicestoandsocialsecuritytoreducetheMediumtolongtermneedforprecautionarysavingsTotalFactorProductivitySectorallaborSocialsecurityreforms(e.g.,centralizingunemploymentbenefitsandMediumtolongtermshiftincreasetheportabilityofhealthinsuranceandpensions)toimprovelabormobilityLoosenhukou(residencepermit)restrictionstofacilitatelaborShorttomediumtermmigrationtourbanareasTradeopennessContinuetoengageinandadvanceinternationaltrade(andservices)MediumtolongtermagreementsSpecializeincertainindustriesandkeeproomtoimportinothersMediumtolongtermFDILoosenrestrictionsonforeigndirectinvestment(FDI)inflowsbyShorttomediumtermrevisingandshorteningthenegativelistforFDIReduceremainingjointventurerequirementsthatpotentiallyholdShorttomediumtermbackinvestorsR&DReininthemisappropriationofresearchanddevelopment(R&D)ShorttermfundsInvestmoreinbasicresearch,especiallyonenterpriselevelMediumtolongtermIncreasethequalityofpatentsissuedbystrengtheningtheirreviewMediumtolongtermStrengthenIPrightsprotectionMediumtolongterm
REFERENCESAcemoglu,D.,D.Y.Yang,andJ.Zhou.2022.PowerandtheDirectionofResearch:EvidencefromChina’sAcademia.NBERChineseEconomyWorkingGroupMeeting,October2022.ADB(AsianDevelopmentBank).2020a.AsianDevelopmentOutlook2020Update:WellnessinWorryingTimes.CountrychapteronthePeople’sRepublicofChina.Manila.ADB.2020b.Post-2020PovertyReductionPolicyOptionsforthePeople’sRepublicofChina.ObservationsandSuggestions.No.2020-03.Manila.ADB.2021a.RisingGlobalInflationandConsumerPricesinthePeople’sRepublicofChina.ObservationsandSuggestions.No.2021-03.Manila.ADB.2021b.TheChallengesofPopulationAginginthePeople’sRepublicofChina.ObservationsandSuggestions.No.2021-02.Manila.ADB.2021c.AsianDevelopmentOutlook2021:FinancingaGreenandInclusiveRecovery.CountrychapteronthePeople’sRepublicofChina.Manila.ADB.2021d.The14thFive-YearPlanofthePeople’sRepublicofChina—FosteringHigh-QualityDevelopment.ObservationsandSuggestions.No.2021-01.ManilaADB.2021e.ReportandRecommendationofthePresidenttotheBoardofDirectors.ProjectNumber:53060-001.Manila.ADB.2022a.AsianDevelopmentOutlook2022Update:EntrepreneurshipintheDigitalAge.CountrychapteronthePeople’sRepublicofChina.Manila.ADB.2022b.SpecialTopicoftheAsianDevelopmentOutlook2022.FallingFurtherBehind:TheCostofCOVID-19SchoolClosuresbyGenderandWealth.Manila.Ahmed,S.,C.Ricardo,D.A.Dias,N.Gornemann,J.Hoek,A.Jain,E.Liu,andA.Wong.2019.GlobalSpilloversofaChinaHardLanding.October.InternationalFinanceDiscussionPapers.No.1260.BoardofGovernorsoftheFederalReserveSystem.Arsov,I.andB.Watson.2019.PotentialGrowthinAdvancedEconomies.GlobalEconomy.Sydney,NSW:ReserveBankofAustralia.Bailliu,J.,M.Kruger,A.Toktamyssov,andW.Welbourn.2016.HowFastCanChinaGrow?TheMiddleKingdom’sProspectsto2030.StaffWorkingPaper2016-15.Ottawa,ON:BankofCanada.Bai,C.andX.Lei.2020.AgingandSocialPolicyinanEraofDemographicTransition,inDollar,D.,Y.Huang,andY.Yao,eds.China2049–EconomicChallengesofaRisingGlobalPower(pp.69–91).Washington,DC:BrookingsInstitutionPress.Barro,R.J.2016.EconomicGrowthandConvergence,AppliedEspeciallytoChina.WorkingPaper21872.Cambridge,MA:NationalBureauofEconomicResearch.Barro,R.J.2017.EconomicGrowthintheWorldandinChina.July.VoxChina.https://voxchina.org/show-3-20.html
References 51Barro,R.J.andJ.Lee.2021.EducationalAttainmentforTotalPopulation,1950-2015.Barro-LeeDataset.http://barrolee.com/Boeing,P.andB.Peters.2022.MisappropriationofR&DSubsidies:EstimatingTreatmentEffectswithOne-SidedNoncompliance.DiscussionPaper.No.21-081/04.Cologne:LeibnizCentreforEuropeanEconomicResearch.Boulter,J.2018.China’sSupply-sideStructuralReform.DecemberBulletin.Sidney:ReserveBankofAustralia.Brandt,L.,J.Litwack,E.Mileva,L.Wang,Y.Zhang,andL.Zhao.2020.China’sProductivitySlowdownandFutureGrowthPotential.PolicyResearchWorkingPaperNo.9298.Washington,DC:WorldBank.Brandt,L.,J.Litwack,E.Mileva,L.Wang,Y.Zhang,andL.Zhao.2022.RecentProductivityTrendsinChina:EvidencefromMacroandFirm-LevelData.China:AnInternationalJournal,Volume20,Number1,February2022,pp.93-113.Brooks,R.andS.Barnett.2006.What’sDrivingInvestmentinChina?IMFWorkingPaper.WP/06/265.Washington,DC:InternationalMonetaryFund.CBO(CongressionalBudgetOffice).2005.R&DandProductivityGrowth:ABackgroundPaper.Washington,DC:TheCongressoftheUnitedStates.Cecchetti,S.G.,M.S.Mohanty,andF.Zampolli.2011.TheRealEffectsofDebt.September.BISWorkingPaper.No352.Basel:BankforInternationalSettlements.Chen,W.,X.Chen,C.T.Hsieh,andZ.Song.2019.AForensicExaminationofChina’sNationalAccounts.BrookingsPaperonEconomicActivity.BPEAConferenceDrafts,7–8March.Cheng,S.2022.MinistryPrioritizingFindingJobsforNewGraduates.ChinaDaily.28April.https://www.chinadaily.com.cn/a/202204/28/WS6269eb3da310fd2b29e59cd6.htmlClark,H.,M.Pinkovskiy,andX.Sala-i-Martin.2017.China’sGDPGrowthMayBeUnderstated.WorkingPaper23323.Cambridge,MA:NationalBureauofEconomicResearch.CerdeiroD.A.,J.Eugster,R.C.Mano,D.Muir,andS.J.Peiris.2021.SizingUptheEffectsofTechnologicalDecoupling.IMFWorkingPaper.WP/21/69.Washington,DC:InternationalMonetaryFund.Dollar,D.andS.J.Wei.2007.Das(Wasted)Kapital:FirmOwnershipandInvestmentEfficiencyinChina.IMFWorkingPaper.WP/07/9.Washington,DC:InternationalMonetaryFund.Dollar,D.,Y.Huang,andY.Yao.2020.China2049–EconomicChallengesofaRisingGlobalPower.Washington,DC:BrookingsInstitutionPress.Dong,X.Y.andV.M.Joffre.2019.InclusiveGrowthinthePeople’sRepublicofChina:ADeepLookAtMenAndWomen’sWorkAmidDemographic,Technological,andStructuralTransformations.ADBEastAsiaWorkingPaperSeries.No.23.Manila:AsianDevelopmentBank.Griffith,R.,S.Redding,andJ.V.Reenen.2004.MappingtheTwoFacesofR&D:ProductivityGrowthinaPanelofOECDIndustries.TheReviewofEconomicsandStatistics.86(4),pp.883–895.
52 ReferencesGriliches,Z.1991.TheSearchforR&DSpillovers.NBERWorkingPaper.No.3768.Cambridge,MA:NationalBureauofEconomicResearch.Guisinger,A.Y.,M.T.Owyang,andH.G.Shell.2018.ComparingMeasuresofPotentialOutput.FederalReserveBankofSt.LouisReview.100(4),pp.297-316.Haider,F.,R.Kunst,andW.Franz.2020.TotalFactorProductivity,ItsComponentsandDrivers.Empirica.No.48,pp283–327.He,W.2022.LocalFiscalConsolidationLooms.22February.GavekalDragonomics.Herd,R.2020.EstimatingCapitalFormationandCapitalStockbyEconomicSectorinChina:TheImplicationsforProductivityGrowth.PolicyResearchWorkingPaper.No.WPS9317.Washington,DC:WorldBank.Higgins,M.2020.China’sGrowthOutlook:IsHigh-IncomeStatusinReach?EconomicPolicyReview26.No.4.October.NewYork:FederalReserveBankofNewYork.InternationalLabourOffice(ILO).2019.TheGlobalLabourIncomeShareandDistribution.DataProductionandAnalysisUnit,DepartmentofStatistics.July.Geneva.InternationalMonetaryFund(IMF).2016.WorldEconomicOutlook:SubduedDemand–SymptomsandRemedies.October.Washington,DC.IMF.2019.People’sRepublicofChina:SelectedIssues.IMFCountryReportNo.19/274.Washington,DC.IMF.2021a.IMFInvestmentandCapitalStockDataset.May.Washington,DC.IMF.2021b.People’sRepublicofChina:SelectedIssues.IMFCountryReportNo.21/12.Washington,DC.IMF.2021c.2020ArticleIVConsultationwiththePeople’sRepublicofChina.IMFCountryReport.No.21/6.Washington,DC.IMF.2022a.2021ArticleIVConsultationwiththePeople’sRepublicofChina.IMFCountryReport.No.22/21.Washington,DC.IMF.2022b.People’sRepublicofChina:SelectedIssues.IMFCountryReport.No.22/22.Washington,DC.Jiang,J.andK.Yi.2015.HowRichWillChinaBecome?ASimpleCalculationBasedonSouthKoreaandJapan’sExperience.FederalReserveBankofMinneapolis.https://www.minneapolisfed.org/article/2015/how-rich-will-china-become.Jiang,K.,W.Keller,L.D.Qiu,andW.Ridley.2019.China’sJointVenturePolicyandtheInternationalTransferofTechnology.VoxChina.https://www.voxchina.org/show-3-115.html.Johnson,P.andC.Papageorgiou.2020.WhatRemainsofCross-CountryConvergence?JournalofEconomicLiterature.58(1).pp.129-175.Kwan,C.,H.2004.WhyChina’sInvestmentEfficiencyisLow–FinancialReformsareLaggingBehind.RIETI.https://www.rieti.go.jp/en/china/04061801.html.
References 53Kroeber,A.R.2020.China’sEconomy:WhatEveryoneNeedstoKnow(Chapter14).2ndEdition.NewYork.OxfordUniversityPress.Lam,W.R.,X.G.Liu,andA.Schipke.2015.China’sLaborMarketinthe“NewNormal”.IMFWorkingPaper.WP/15/151.Washington,DC:InternationalMonetaryFund.Lardy,N.2019.TheStateStrikesBack–TheEndofEconomicReforminChina.Washington,DC:PetersonInstituteforInternationalEconomics.Lee,H.,J.Lee,andH.Kim.2011.ForeignDirectInvestment,TechnologyDiffusionandHostCountryProductivityGrowth.WorkingPaper.No.272.Manila:AsianDevelopmentBank.Li,H.,P.Loyalka,S.Rozelle,andB.Wu.2017.HumanCapitalandChina’sFutureGrowth.JournalofEconomicPerspectives.31(1)1,pp.25-48.Li,H.Z.2021.ChinaHumanCapitalReport2021[InChinese].HumanCapitalandLaborEconomyResearchCenterofCentralUniversityofFinanceandEconomics.Loyalka,P.,E.Kardanova,I.Chirikov,L.Liu,GLi,H.Wang,E.Enchikova,H.Shi,andN.Johnson.2016.DevelopingInstrumentstoAssessandComparetheQualityofEngineeringEducation:TheCaseofChinaandRussia.Assessment&EvaluationinHigherEducation.41(5),pp.770–786.Ma,X.andI.Iwasaki.2021.ReturntoSchoolinginChina:ALargeMeta-analysis.EducationEconomics.29(4),pp.379–410.Maliszewski,W.andL.Zhang.2015.China’sGrowth:CanGoldilocksOutgrowBears?IMFWorkingPaper.WP/15/113.Washington,DC:InternationalMonetaryFund.Mankiw,N.G.,D.Romer,andD.N.Well.1992.AContributiontotheEmpiricsofEconomicGrowth.TheQuarterlyJournalofEconomics.May.pp.408-427.McMorrow,K.andW.Roger.2009.R&DCapitalandEconomicGrowth:TheEmpiricalEvidence.EuropeanInvestmentBankPapers.14(1),pp.94-118.NationalDevelopmentandReformCommission(NDRC).2021.The14thFive-YearPlan(2021–25)forNationalEconomicandSocialDevelopmentandtheLong-RangeObjectivesThroughtheYear2035[inChinese].Chapter13.http://www.gov.cn/xinwen/2021-03/13/content_5592681.htm.OrganisationforEconomicCo-operationandDevelopment(OECD).2021.MainScienceandTechnology.Volume2021/2.Paris.OECD.2022a.OECDEconomicSurveysChina.March.Paris.OECD.2022b.FinancingSMEsandEntrepreneurs2022:AnOECDScoreboard.Paris.Peschel,D.2021.ReducingInequalityinthePeople’sRepublicofChinaThroughTaxandFiscalReforms.ADBEastAsiaWorkingPaperSeries.No.46.Manila.AsianDevelopmentBank.Pettis,M.2022.HowDoesExcessiveDebtHurtanEconomy?8February.CarnegieEndowmentforInternationalPeace.https://carnegieendowment.org/chinafinancialmarkets/86397.
54 ReferencesPowell,T.,L.Sheiner,andD.Wessel.2021.WhatIsPotentialGDP,andWhyIsItSoControversialRightNow?Brookings.https://www.brookings.edu/blog/up-front/2021/02/22/what-is-potential-gdp-and-why-is-it-so-controversial-right-now/.Reinhart,C.M.andK.S.Rogoff.2010.GrowthinaTimeofDebt.NBERWorkingPaper.No.15639.Cambridge,MA:NationalBureauofEconomicResearch.Roberts,I.andB.Russell.2019.Long-termGrowthinChina.GlobalEconomy.December.Sydney,NSW:ReserveBankofAustralia.Rogoff,K.andY.C.Yang.2022.ATaleofTier3Cities.IMFWorkingPaper.WP/22/196.Washington,DC:InternationalMonetaryFund.Rozelle,S.,Y.Xia,D.Friesen,B.Vanderjack,andN.Cohen.2020.MovingBeyondLewis:EmploymentandWageTrendsinChina’sHigh-andLow-SkilledIndustriesandtheEmergenceofanEraofPolarization.ComparativeEconomicStudies.62(4),pp.555–589.Sasaki,T.,T.Sakata,Y.Mukoyama,andK.Yoshino.2021.China’sLong-TermGrowthPotential:CanProductivityConvergenceBeSustained?WorkingPaper.No.21.Tokyo:BankofJapan.StateCouncil.2016.NoticeonLocalGovernmentDebtRiskEmergencyResponsePlan[inChinese].GeneralOfficeoftheStateCouncil.http://www.gov.cn/zhengce/content/2016-11/14/content_5132244.htmStateCouncil.2019.EducationModernization2035Plan[inChinese].Chapter4.https://www.uta.edu.cn/fzghc/2021/1224/c1958a108579/page.htmSummers,L.andL.Pritchett.2014.AsiaphoriaMeetsRegressiontoTheMean.NBERWorkingPaper.No.20573.Cambridge,MA:NationalBureauofEconomicResearch.TheEconomist.2018.LightontheShadows.14June.TheEconomist.2019.Old,notyetRich.2November.TheEconomist.2021.VanguardoftheNon-WorkingClass—At54,China’sAverageRetirementAgeIsTooLow.26June.UnitedNations.2019.WorldPopulationProspects.DepartmentofEconomicandSocialAffairs,PopulationDivision.Wang,T.andN.Zhang.2022.ChinaEconomicPerspectives:EconomicPolicyTakeawaysfromthe20thPartyCongress.GlobalResearchandEvidenceLab.UBS.WorldBank.2019.CyclicalRisksandStructuralImperatives.ChinaEconomicUpdate.December.Washington,DC.WorldBank.2021.RebalancingActfromRecoverytoHigh-qualityGrowth.ChinaEconomicUpdate.December.Washington,DC.WorldBankandDRC(DevelopmentResearchCenteroftheStateCouncil).2013.China2030,BuildingaModern,Harmonious,andCreativeSociety.Washington,DC.
References 55WorldBankandDRC(DevelopmentResearchCenteroftheStateCouncil).2019.InnovativeChinaNewDriversofGrowth.Washington,DC.Yao,R.2022.ThisTime,PropertyHasPeaked.29March,GavekalDragonomics.Zheng,S.,Q.Zhuang,andY.Wang.2020.China’sInnovationCapacityin2049.InDollar,D.,Y.Huang,andY.Yao,eds.China2049–EconomicChallengesofaRisingGlobalPower.pp.235–258.Washington,DC:BrookingsInstitutionPress.Zhu,E.andT.Orlik.2022.WhenWillChinaBetheWorld’sBiggestEconomy?MaybeNever.Bloomberg.11February.https://www.bloomberg.com/news/articles/2022-02-11/when-will-china-be-the-world-s-biggest-economy-maybe-neverZhu,H.B.,G.Ng,andT.T.Ge.2021.China:FromPost-PandemicRecoverytoSustainableGrowth.EconomicResearch.J.P.Morgan.Zhu,M.,L.Zhang,andD.Peng.2019.China’sGrowthPotential—AStocktakingandReassessment.IMFWorkingPaper.WP/19/263.Washington,DC:InternationalMonetaryFund.
TheLong-TermGrowthProspectsofthePeople’sRepublicofChinaEconomicgrowthinthePeople’sRepublicofChinahasmoderatedoverthepastdecade,evenbeforethecoronavirusdisease(COVID-19)pandemic.AcriticalquestionistowhichlevelofeconomicgrowththecountrywillreturntoafterCOVID-19and,morebroadly,howitslong-termgrowthprospectsare,givenarapidlyagingsociety,theongoinghighdependenceofgrowthoninvestment,andachangedinternationalenvironment.Thispaperestimatesthecountry’spotentialgrossdomesticproductgrowthupto2040basedonaCobb-Douglasproductionfunctionandprovidespolicyrecommendationsonstructuralreformstoboostpotentialgrowth.AbouttheAsianDevelopmentBankTHELONG-TERMGROWTHADBiscommittedtoachievingaprosperous,inclusive,resilient,andsustainableAsiaandthePacific,whilesustainingitseffortstoeradicateextremepoverty.Establishedin1966,itisownedby68members—49fromtheregion.Itsmaininstrumentsforhelpingitsdevelopingmembercountriesarepolicydialogue,PROSPECTSOFTHEPEOPLE’Sloans,equityinvestments,guarantees,grants,andtechnicalassistance.REPUBLICOFCHINADominikPeschelandWenyuLiuNO.56ADBEASTASIADecember2022WORKINGPAPERSERIESASIANDEVELOPMENTBANK6ADBAvenue,MandaluyongCity1550MetroManila,PhilippinesASIANDEVELOPMENTBANKwww.adb.org
6WDUW
此文档下载收益归作者所有
