劳动力市场管制改革对农村人口迁移的影

1LaborMarketDeregulationsandMigrationConsequencesinChina:EvidencefromtheHouseholdRegistrationReformProgram(PaperpresentedattheThirdChinaEconomicsAnnualConference,20-21December2003,FudanUniversity,ShangHai,China)LaborEconomicsandPopulationEconomicsSectionHuangYi&WangGewei*ChinaCenterforEconomicResearchPekingUniversityJuly8,2003AbstractBasedonanempiricalstudyofChinaTownHouseholdRegistrationReformProgram(CTHRRP)inZhejingProvince,theauthortriestodiscussthelabormarketderegulationandmigrationconsequencesofruralpopulation.UsingtheGeneralandadjusteddifference-in-differences(DID)estimator(Abadie,2001,Duflo,2003),instrumentvariable(IV)-2SLSandprobitmodels,weshowthatpeasantsdidn’tswarmintocitiesandthelabormarketderegulationreformshavenosignificanteffectsonruralincomeandmacro-economyofreformtown.Theauthorfindthegovernmenttakeadvantageofhukoureformasachancetogivemorepowerandrents,whichresultinthenegativeeffectsofthefiscalexpenseincrement.Finally,thissuggeststhatattemptstoderegulatethelabormarketandpromotethemigrationofruralpeoplesmayendbyhurtingtheruralsector.KEYWORDS:LaborMarketDeregulationPolicy,Migration,DifferenceinDifferencesEstimatorJELCLASSIFICATIONNUMBERS:C14,J61,J68lTheauthor’scorrespondence:ChinaCenterforEconomicResearch,BeijingUniversity,Beijing100871,China.Phone:[phone];fax:[phone];email:[e-mail]the2002“ChinaPublicPolicyGrantsCompetition”,theFordFoundation.lIamgratefultoProf.ZhouQiren(myadvisor),ZhaoYaohui,HuDayuan,WangDingding,ShengMinggao,ZhaoZhong(PekingUniversity),AlbertPark(MichiganUniversity),LouiseFox(WorldBank),JohnKnight,TaoRan(OxfordUniversity),LinaSong(NottinghamUniversity),LuMing(FudanUniversity),Yao,Xianguo,LuoWeidongandChenLing(ZhejiangUniversity)forhelpfulcommentsandinsightfulsuggestions.ThecriticismsandsuggestionsfromparticipantsoftheLaborWorkshopatCCERarealsogratefullyacknowledged.Allpossibledeficienciesinthispaper,however,areminealone.2:90,Generalandadjusteddifference-in-differences(DID)(Abadie,2001,Duflo,2003),2SLSProbit,“”“”Difference-in-DifferencesEstimator20036831/41994Rawski,2002,290Gale.Johnson12001500Johnson,200232:19784%198531%,199843%198519986%55%Johnson,20013J490,;Difference-In-DifferenceGeneral-DID2SLSProbit20“”/”4(Oliver.BlanchardandJustinWolfers,2000)“DoLaborMarketRegulationsAffectLaborEarringsinEcuador”DannaMaclsaacMartinRama19974GordonBetcherman,AmyLuinstra,andMakotoOgawa2001LaborMarketRegulation:InternationalExperienceinPromotingEmploymentandSocialProtection.530%rigidBotero,Juan,SimeonDjankov,RafaelLaPorta,FlorencioLopez-de-Silanes,andAndreiShleifer200385HernandodeSoto198920005self-entrepreneurshipOliverBlanchard,19976Djankov,Simeon,RafaelLaPorta,FlorencioLopez-deSilanes,AndreiShleifer191492561662394622280670LiangandMichael,1997,Zhao,1997,1999ChengandSelden1994Solinger,1999ChanandLi,1999,AuandHenderson,2002Johnson7740Johnson,2000,2002Chan,2000Johnson“”,:12001500Johnson,2002“”“”“”“”“”8“”“”“”2002“”“”“”“”8820“”2001“90200180%(1)901232512529934123121231995127199620011200211888819902001881070TreatmentGroup(ControlGroup)2.1—2.24.2Difference-In-Difference(DID)ModelTwo-Wayfixed-effectsModelHsiao,1986,Difference-In-Difference(DID)ModelWooldridge,2002DIDEricaFieldAnneCaseEricaFieldetal,2002AlbertParkAlbertParketal,2002DID“”“”AlbertoAbadie(2001)SemiparametricDifference-in-DifferencesestimatorsSusanAtheyGuidoW.ImbensNonlineardifferenceindifferencesestimatesEstherDuflo(Dufloetal,2003)“Howmuchshouldwetrustdifferenceindifferencesestimates”DIDDIDDID—1992—19951996—1998ABdB10d20011yd2dBd2dBubdbd=+++⋅+yd2dB⋅1dOLS1ˆdDIDdifferences-in-differences,1Ayy,2Ayy,1By11,2By1ˆd1,2,1,2,1ˆ()()BBAAyyyyd=---33DID0.03380.14700.11990.1199-0.03380.0861DID(systematic)CausalEffectDID0.20583DIDDID(0.0091vs.0.0861)t12DIDt1012DiffDiff7.47517.50890.03388.34468.37830.0338(0.0512)(0.0772)(0.1470)(0.0205)(0.0247)(0.0585)After8.22108.34090.1199After8.91998.94460.0247(0.0503)(0.0723)(0.1440)(0.0208)(0.0519)(0.0612)Diff0.74600.83200.0861Diff0.57530.5662-0.0091(0.0718)(0.1058)(0.2058)(0.0292)(0.0575)(0.0847)DiffDiff10.202310.54770.34537.60908.83351.2245(0.0777)(0.1292)(0.2235)(0.2421)(0.3156)(0.6916)After10.187910.54080.3529After7.72898.90861.1798(0.0796)(0.1349)(0.2292)(0.2493)(0.3243)(0.7122)Diff-0.0144-0.00690.0076Diff0.11990.0752-0.0447(0.1112)(0.1868)(0.3202)(0.3475)(0.4525)(0.9927)DiffDiff10.044910.52660.48175.77086.33820.5674(0.1827)(0.2457)(0.5223)(0.1376)(0.2721)(0.3990)After10.960811.46890.5081After6.53767.15510.6175(0.1856)(0.3443)(0.5367)(0.1317)(0.1486)(0.3752)Diff0.91590.94230.0264Diff0.76680.81690.0501(0.2604)(0.4230)(0.7489)(0.1905)(0.3100)(0.5477)12DiffDiffDID3BeforeAfter13319957,,”,DIDControllableExperimentquasi-experimentDIDDID110-1ProbitPseudoR2Loglikelihood19951994DID7091994FPNFPAVIVFAINCNFAINCFIFE(F2np)451%11MIT——OLSDoandIyer(2002)145%DID4Probit7091994-0.0106**(0.0042)1.6034***(0.5949)-0.3560(0.4199)-18.5776**(7.4993)N70PseudoR20.49loglikelihood-13.77***p0.01;**p0.05;*p0.1;probitPseudoR20.20.4PseudoR20.49121280002000015IV2SLSDID2SLS4.4GeneralDiff