How Extensive Is the Brain Drain? - International Monetary Fund

One important implication of the brain drain is that investment in education in a developing country may not lead to faster economic growth if a large number of ... What'sNew SiteMap SiteIndex ContactUs Glossary Home AbouttheIMF Research CountryInfo News Videos DataandStatistics Publications ShareThis SearchFinance&Development AdvancedSearch AboutF&D Subscribe BackIssues WriteUs CopyrightInformation UsethefreeAdobeAcrobatReadertoviewapdffileofthisarticle FreeEmailNotification Receiveemailswhenwepostnew itemsofinteresttoyou.Subscribeor Modifyyourprofile AquarterlymagazineoftheIMFJune1999,Volume36,Number2 HowExtensiveIstheBrainDrain? WilliamJ.CarringtonandEnricaDetragiache Howextensiveisthe"braindrain,"andwhichcountriesandregionsaremoststronglyaffectedbyit?Thisarticleestimatestheextentofmigration,bylevelofeducation,fromdevelopingcountriestotheUnitedStatesandotherOECDcountries. Perhapstheoldestquestionineconomicsiswhysomecountriesarerichwhileothersarepoor.Economictheoryhasemphasizedthatdifferencesintheeducationallevelsofthepopulationareanimportantpartoftheanswerandthatimprovedschoolingopportunitiesshouldraiseincomesindevelopingcountries.Yet,whilethereislittledoubtthathighlyeducatedworkersinmanydevelopingcountriesarescarce,itisalsotruethatmanyscientists,engineers,physicians,andotherprofessionalsfromdevelopingcountriesworkinCanada,theUnitedStates,andWesternEurope.Thisphenomenon,oftenreferredtoasthe"braindrain,"wasnoticedasearlyasthe1960sandhasbeenacontentiousissueintheNorth-Southdebateeversince.Oneimportantimplicationofthebraindrainisthatinvestmentineducationinadevelopingcountrymaynotleadtofastereconomicgrowthifalargenumberofitshighlyeducatedpeopleleavethecountry.Also,effortstoreducespecificskillshortagesthroughimprovededucationalopportunitiesmaybelargelyfutileunlessmeasuresaretakentooffsetexistingincentivesforhighlyeducatedpeopletoemigrate. Buthowextensiveisthebraindrain?Whichcountriesandregionsareespeciallyaffected?Dohighlyeducatedprofessionalsfromdevelopingcountrieslivingabroadrepresentasizableproportionofthepoolofskilledworkersintheircountriesoforiginortoosmallanumbertoworryabout?Unfortunately,attemptstoanswertheseimportantquestionsquicklycomeupagainstaformidablebarrier:thereisnouniformsystemofstatisticsonthenumberandcharacteristicsofinternationalmigrants.Also,sourcecountriestypicallydonotkeeptrackofemigrants'characteristics,and,althoughsomereceivingcountriesdo,theirdefinitionsofimmigrationdiffer.Thus,itisdifficulttomeasurepreciselytheflowandlevelsofeducationofimmigrants.Further,ithasonlyrecentlybecomepossibletomeasurethestockofeducatedworkersineachsourcecountry—thepoolfromwhichbrainpowerisdrained. Estimatingthebraindrain Despitethelackofsystematicdataaboutinternationalmigrants,estimatesofthestockofmigrantsbyeducationallevelinmembercountriesoftheOrganizationforEconomicCooperationandDevelopment(OECD)canbeconstructedusingavarietyofdatasources.Theresultingestimatesarelessthanperfectinmanyrespects,buttheysignificantlyimproveourknowledgeofthemagnitudeofthebraindrain.Thestudyonwhichwebasedthisarticle(CarringtonandDetragiache,1998)coversmigrationfrom61developingcountriesaccountingforabout70percentofthetotalpopulationofdevelopingcountries.Becauseofthelackofdata,wehavenotattemptedtoestimatetheextentofeitherthebraindrainfromtheformerSovietUnionandEasternEurope,eventhoughcasualevidencesuggeststhatitissubstantial,ormigratoryflowsamongdevelopingcountries.Wefollowedatwo-stepprocedure:first,estimatesofthebraindraintotheUnitedStateswereconstructedusing1990U.S.censusdataandothersourcesofinformation.Then,theseestimateswereused—togetherwithdataonmigrantstoOECDcountriesotherthantheUnitedStatesdrawnfromtheOECD'sContinuousReportingSystemonMigration—toestimatetheextentofthebraindraintoallOECDcountries.WhiletheresultingestimatesshouldbereasonablypreciseformigrationtotheUnitedStates(whichaccountsfor54.3percentofthetotalmigrationfromthedevelopingcountriesinoursampletoallOECDcountries),theyaremuchmoretentativeforthebraindraintoallOECDcountries. TheU.S.censusreportswhetherindividualspolledareforeignbornand,iftheyare,whattheircountryoforiginis;thenumberofyearsofschoolingreceivedisalsoreportedforeachindividual.Afterindividualsunder25yearsofageareeliminatedtoensurecompatibilitywiththedataoneducationalattainmentdescribedbelow,allforeign-bornindividualsinthecensusareputintooneofthreebroadeducationalcategories:primary(0to8yearsofschooling),secondary(9to12yearsofschooling),andtertiary(morethan12yearsofschooling).Afurtheradjustmentinvolvessubtractingfromthegroupofforeign-bornindividualswithatertiaryeducationallgraduatestudentsinU.S.universities,usingdatafromtheInstituteofInternationalEducation.Thisprocedureyields,foreachdevelopingcountryinthesample,thenumberofmigrantsintheUnitedStatesineachofthethreeeducationalcategories.Toassesstheextentofthebraindrainfromeachcountryconsidered,theseestimatesmustbecomparedwiththenumberofindividualsineacheducationalgroupwhoremainintheirhomecountry.Doingthisrequiresabreakdownbyeducationalcategoryofthepopulationofeachdevelopingcountryinthesample,forwhichwerelyonadatasetrecentlyassembledbyRobertBarroandJong-WhaLee(BarroandLee,1993),whichprovidesthebestestimatesavailabletodateofeducationalattainmentforindividualsmorethan25yearsofageinalargesampleofcountries. BraindraintotheUnitedStates ThefirststrikingfeatureoftheU.S.migrationdataisthatimmigrationflowsofindividualswithnomorethanaprimaryeducationarequitesmall,bothinabsolutetermsandrelativetoothereducationalgroupings(about500,000individualsoutofatotalof7millionimmigrants).Foreign-bornindividualswithlittleornoeducation,however,maybeundercountedbythecensusiftheyareinthecountryillegallyordonotspeakEnglish.ThelargestgroupofimmigrantsintotheUnitedStates(about3.7million)consistsofindividualswithsecondaryeducationfromotherNorthAmericancountries(understoodheretoincludeCentralAmericanandCaribbeancountries),primarilyMexico.Perhapssurprisingly,thesecondlargestgroup(almost1.5millionindividuals)consistsofhighlyeducatedmigrantsfromAsiaandthePacific.TotalimmigrationfromSouthAmericaand,especially,Africaisquitesmall.Itisnoteworthy,however,thatimmigrantsfromAfricaconsistprimarilyofhighlyeducatedindividuals(about95,000ofthe128,000Africanmigrants). AmongthecountriesinAsiaandthePacific,thebiggestsourceisthePhilippines,with730,000migrants.Ofthese,thegreatmajorityhaveatertiaryeducation.ThesecondlargeststockofmigrantsisfromChina(400,000),whichissplitalmostequallybetweenthesecondaryandtertiaryeducationalgroups.BothIndiaandKoreahaveseenmorethan300,000peoplemigratetotheUnitedStates.Itisstrikingthatmorethan75percentofIndianimmigrantshaveatertiaryeducation,comparedwithonly53percentofKoreanimmigrants. ThebiggestmigratoryflowsfromAfricatotheUnitedStatesarefromEgypt,Ghana,andSouthAfrica,withmorethan60percentofimmigrantsfromthosethreecountrieshavingatertiaryeducation.MigrationofAfricanswithonlyaprimaryeducationisalmostnil.ThepictureisquitedifferentforthemigratoryflowsfromtheWesternHemisphere:Mexicoisbyfarthelargestsendingcountry(2.7million),withthelargemajorityofitsmigrants(2.0million)havingasecondaryeducationandfewerthan13percenthavingatertiaryeducation.ThispatternisalsoobservedforthesmallercountriesofCentralAmerica,butnotforthetwoCaribbeancountriesforwhichwehaveinformation,forwhichmigrantswithatertiaryeducationareamoresubstantialpercentageofthetotal(42percentforJamaicaand46percentforTrinidadandTobago).Finally,migrationfromSouthAmericatotheUnitedStatesisrelativelysmallinabsolutenumbers,withimmigrantssplitalmostequallybetweenthesecondaryandthetertiaryeducationalgroups. Ineachsendingcountry,howdothenumbersofemigrantscomparewiththesizeofthepopulationwithagiveneducationalattainment?Formostcountries,peoplewithatertiaryeducationhavethehighestmigrationrate,withtheexceptionsoftheCentralAmericancountries,Ecuador,andThailand(inThailand,peoplewithasecondaryeducationandthosewithatertiaryonehaveapproximatelythesamemigrationrates)(seechart).Thus,migrantstotheUnitedStatestendtobebettereducatedthantheaveragepersonintheirhome(thatis,thesending)country,andtheproportionofveryhighlyeducatedpeoplewhomigrateisparticularlyhigh.Also,migrationfromCentralAmericaseemstofollowasomewhatdifferentpatternthanmigrationfromotherdevelopingcountries,inthatthehighestmigrationrateisforpersonswithasecondaryeducation,ratherthanthosewithatertiaryeducation. ThebraindraintotheUnitedStatesfrommanyCentralAmericanandCaribbeancountriesissubstantial:forpersonswithatertiaryeducation,immigrationratesforvirtuallyallthesecountriesareabove10percent,andsomeappeartobe50percentorevenhigher.InSouthAmerica,thecountrywithbyfarthelargestbraindrainisGuyana,fromwhichmorethan70percentofindividualswithatertiaryeducationhavemovedtotheUnitedStates;fortherestoftheregion,theimmigrationratesforthiseducationalgrouparemuchlower.TheIslamicRepublicofIranhashadasubstantialdrainofhighlyeducatedindividuals(morethan15percent)andsohasTaiwanProvinceofChina(8–9percent). BraindraintootherOECDcountries ToconstructestimatesofthebraindrainfromdevelopingcountriestoOECDcountries,wehavereliedontheOECDContinuousReportingSystemonMigration.Unfortunately,unliketheU.S.census,thisdatasourcedoesnotreporttheyearsofschoolingthatmigrantshavereceived.Forlackofanypracticalalternatives,wehaveassumedthatthedistributionofimmigrantsbyeducationalcategoryfromeachsourcecountryisthesamefortheUnitedStatesasforotherOECDcountries.Althoughthisistheonlyfeasibleapproach,whichoftenproducesnumbersthatareconsistentwithanecdotalevidence,therearesomeinstancesinwhichityieldsimplausibleresults,particularlyforcountrieswithlowratesofimmigrationtotheUnitedStatesbuthighratestooneormoreoftheotherOECDcountries.ImmigrantstotheUnitedStatesfromsuchcountriesarelikelytobebettereducatedthanimmigrantstootherOECDcountries,whothusmaybemorerepresentativeofthesourcecountry'spopulation. AsecondproblemwiththedataforOECDcountriesotherthantheUnitedStatesliesinthedifferentcriteriaforclassifyingindividualsasimmigrants.AlthoughAustralia,Canada,andtheUnitedStatesdefineanimmigrantasapersonwhowasbornabroadtononcitizens,mostEuropeancountriesdefineimmigrantstatusbasedontheethnicityorimmigrationstatusoftheparent.AthirddifficultywiththeOECDdataisthattheydidnotpermitustoexcludeimmigrantsundertheageof25.Finally,theOECDrecordsimmigrantsfromonlythetop5or10countriesfromwhichtheycometoeachOECDcountry.Thus,forexample,theOECDfiguresforCanadawouldincludespecificinformationonthenumbersofimmigrantsfromChinaandMexico,butnotthosefromJamaicaandElSalvador.Thisisaproblemwhenemigrationflowsaresignificantforthesourcecountrybutsmallforthereceivingcountry.Thus,particularlyforsmallcountries,ourestimatesofimmigrationtoOECDcountriesotherthantheUnitedStatesmaybeseriouslyunderstated. If,asaruleofthumb,weconsiderestimatestobeunreliablewhenmigrantstotheUnitedStatesaccountforlessthanone-thirdofthetotalofimmigrantstoallOECDmembercountries,thenallestimatesforimmigrationfromtheAsianandPacificcountriesarereliablewiththeexceptionsofthoseforMalaysiaandSriLanka.Turkeyisalsoanexception.Amongtheremainingcountries,theextentofthebraindraintoallOECDmembersissubstantial—anditincreasessignificantlycomparedwiththeU.S.data—fortheIslamicRepublicofIran,Korea,and,toalesserextent,thePhilippines.FortheIslamicRepublicofIran,thefractionofthepopulationwithatertiaryeducationlivinginOECDcountriesisaround25percent;forKorea,15percent;andforthePhilippines,about10percent.ForPakistan,themigrationrateofindividualswithatertiaryeducationismorethan7percent,whileforIndiaitisabout2.7percent;thesefigures,however,failtotakeintoaccountthesizableflowofprofessionalsfromtheIndiansubcontinenttoBahrain,Kuwait,Oman,Qatar,andtheUnitedArabEmiratesandthereforeneglectanimportantcomponentofthebraindrainfromtherelevantsourcecountries.ThemigrationrateofhighlyeducatedindividualsfromChinaisabout3percent. ForAfrica,theestimatesareunreliableforAlgeria,Senegal,andTunisia,fromwhichmigrantsgomainlytoFrance.Formostothercountriesinthesample,however,migrationtoOECDcountriesotherthantheUnitedStatesisquitesmall,sotheresultsderivedfortheUnitedStatesremainessentiallyvalid.Thereare,however,someexceptions:forGhana,themigrationrateofhighlyeducatedindividualsisadramatic26percent;forSouthAfrica,itismorethan8percent;forEgypt,thebraindrainincludes2.5percentofsuchindividualsemigratingtotheUnitedStatesandanother5percentemigratingtootherOECDcountries.ForcountriesintheWesternHemisphere,thebulkofmigrationistotheUnitedStates,andinclusionofflowstotherestoftheOECDmakeslittledifference.TheonlyexceptionisJamaica,whichhasaconsiderablestockofmigrantslivingintheUnitedKingdom.ThedrainfromJamaica'spopulationwithsecondaryeducationis33percent,whilethatfromitspopulationwithtertiaryeducationismorethan77percent. Conclusion Ourestimatesshowthatthereisanoveralltendencyformigrationratestobehigherforhighlyeducatedindividuals.WiththeimportantexceptionsofCentralAmericaandMexico,thehighestmigrationratesareforindividualswithatertiaryeducation.Anumberofcountries—especiallysmallcountriesinAfrica,theCaribbean,andCentralAmerica—lostmorethan30percentofthisgrouptomigration.WehavealsofoundasizablebraindrainfromIran,Korea,thePhilippines,andTaiwanProvinceofChina.Thesenumberssuggestthatinseveraldevelopingcountriestheoutflowofhighlyeducatedindividualsisaphenomenonthatpolicymakerscannotignore. Moreresearch,especiallyempiricalstudies,isneededtoevaluatetheimpactofthebraindrainonsourceeconomiesandonworldwidewelfare,aswellasthereasonsforsuchmigration.Inregardtothelattersubject,immigrationpoliciesinOECDcountriestendtofavorbetter-educatedpeople,whichmayexplainwhytheeducationalcompositionoftotalmigrationisskewedtowardthebettereducatedbutcannotexplainwhysomanyskilledworkersarewillingtoleavedevelopingcountries.Wagedifferentialsmaybepartoftheexplanation,butthisraisesthequestionofwhataccountsforsuchdifferentials.Differencesinthequalityoflife,educationalopportunitiesforchildren,andjobsecuritymayalsoplayarole,asmaythedesiretointeractwithabroadergroupofsimilarlyskilledcolleagues.Anotherimportantissueistheextenttowhichthebenefitsofeducationacquiredbycitizensofdevelopingcountriesareexternalitiesthatindividualscannotbeexpectedtotakeintoaccountwhenmakingtheirprivatedecisions.Ifsuchexternalitiesaresubstantial,asisemphasizedbythe"newgrowththeory,"thenpoliciestocurbthebraindrainmaybewarranted. Ourresearchalsoindicatesseveralwaysinwhichestimatesofthebraindraincouldbeimprovedusingexistingdata.Thefirstwouldbetousecensusinformationforotherlargeimmigrantreceivingcountries,suchasAustralia,Canada,France,andGermany.TogetherwiththeUnitedStates,thesefourcountriesaccountforabout93percentoftotalmigratoryflowstoOECDcountries,sotheresultingfigureswouldbeaverygoodapproximationofthetotal.Anotherpromisingdirectionforfutureresearchwouldbetotrytoobtain,fromcensusdataorothersources,moredetailedinformationabouttheoccupationalcategoriesofhighlyskilledmigrants,inordertoassesswhetherthebraindrainfromagivencountryisespeciallymarkedforparticularprofessionalgroups.Thistypeofanalysiscouldbeusefulforevaluatingtheproblemsthatpolicyprograms—suchashealthsectorreform,financialliberalization,orcivilservicereform—mayencounterindevelopingcountries. Suggestionsforfurtherreading: RobertJ.BarroandJong-WhaLee,1993,"InternationalComparisonsofEducationalAttainment,"JournalofMonetaryEconomics,Vol.32(March),pp.363–94. WilliamJ.CarringtonandEnricaDetragiache,1998,"HowBigIstheBrainDrain?"IMFWorkingPaper98/102(Washington). WilliamJ.CarringtonisanEconomistattheU.S.BureauofLaborStatistics,Washington,DC.TheresearchonwhichthisarticleisbasedwascarriedoutwhilehewasaVisitingScholarintheIMF'sResearchDepartment. EnricaDetragiacheisanEconomistintheCommoditiesandSpecialIssuesDivisionoftheIMF'sResearchDepartment. Home What'sNew SiteMap SiteIndex AbouttheIMF Research CountryInfo News Videos DataandStatistics Publications CopyrightandUsage PrivacyPolicy HowtoContactUs عربي 中文 Français 日本語 Русский Español