Date post: | 31-Aug-2018 |
Category: |
Documents |
Upload: | vuongkhanh |
View: | 214 times |
Download: | 0 times |
1
AUser’sGuidetothe2016NationalAcademyReporton“TheEconomicandFiscalConsequencesofImmigration”
GeorgeJ.Borjas*September22,2016
Introduction
TheNationalAcademyofSciences(NAS)hasjustpublishedamajorreportonTheEconomicandFiscalConsequencesofImmigration.TheNationalAcademypanelthatpreparedthereportconsistedofabout20socialscientists,includingeconomists,sociologists,anddemographers.TheprojectwasledbyFrancineBlau,aprofessorofeconomicsatCornell,andChristopherMackie,whoisastudydirectorwiththeCommitteeonNationalStatisticsattheNAS.FranandChrisdidanamazingjobbringingthisverycomplicatedprojecttofruitionoverathree-yearperiod.Theywere(very)patient,fair,professional,andmadesurethatalltheworkdonebythemembersofthepanelwassomehowweavedintoacohesivewhole--andthat'snosmallfeat!AnddidIsaytheywereverypatient?
Fulldisclosure:IwasamemberoftheNASpanelthatpreparedthereport,butanythingIsayinthisseriesofpostsreflectsonlymytakeaboutwhatisinthereportandwhat,Ithink,isimportant.Thesepostshavenotbeenreadorvettedbyanyoneinthepanel.
Ithinkthereporthasfourmajorconclusions.Butitisnear300,000wordslongwithbigchunksofitwrittenin"technical-ese,"comprehensibleonlytotrainedeconomistsandlikelytoappealonlytoimmigrationgeeks.SoI'mgoingtowritefivepoststhattogethermakeup"AUser'sGuide."TheUser'sGuidewilllinktothemaintablesandfiguresinthereportthatdocumentthoseconclusions.
LetmestartbygivingabriefoutlineofmyUser'sGuide.Allquotesbelowarefromthereport'ssummary:
1 Therehasbeenaslowdowninassimilationduringtheimmigrants'lifetime(User'sGuide,1)."AstimespentintheUnitedStateslengthens,immigrants’wagesincreaserelativetothoseofnativesandtheinitialwagegapnarrows.However,thisprocessofeconomicintegrationappearstohaveslowedsomewhatinrecentdecades;therateofrelativewagegrowthandEnglishlanguageacquisitionamongtheforeign-bornisnowslightlyslowerthanitwasforearlierimmigrantwaves."
2 Immigrationhasaharmfuleffectontheearningsoflow-skillworkers(User'sGuide,2):"Whenmeasuredoveraperiodof10yearsormore,theimpactof
*ProfessorofEconomicsandSocialPolicy,HarvardKennedySchool;ResearchAssociate,National
BureauofEconomicResearch;andProgramDirector,ProgramonLaborMobility,IZA.
2
immigrationonthewagesofnativesoverallisverysmall.However,estimatesforsubgroupsspanacomparativelywiderrange,indicatingarevisedandsomewhatmoredetailedunderstandingofthewageimpactofimmigrationsincethe1990s.Totheextentthatnegativewageeffectsarefound,priorimmigrants—whoareoftentheclosestsubstitutesfornewimmigrants—aremostlikelytoexperiencethem,followedbynative-bornhigh-schooldropouts,whosharejobqualificationssimilartothelargeshareoflow-skilledworkersamongimmigrantstotheUnitedStates."
3 Immigrantsandtheirdependentchildrencreateafiscalburden(User'sGuide,3;andUser'sGuide,4):"Onaverage,individualsinthefirstgenerationaremorecostlytogovernments,mainlyatthestateandlocallevels,thanarethenative-borngenerations...For2013,thetotalfiscalshortfall(i.e.,theexcessofgovernmentexpendituresovertaxes)was$279billionforthefirstgenerationgroup...Viewedoveralongtimehorizon(75yearsinourestimates),thefiscalimpactsofimmigrantsaregenerallypositiveatthefederallevelandnegativeatthestateandlocallevels."Butthesefiscalimpactestimatesare,rightly,stampedwithaConsumerWarninglabel:"Assumptionsplayacentralroleinanalysesofthefiscalimpactsofimmigration."
4 Thebottomline.(User'sGuide,5):TheNASreportdoesnotconductthefinal(andobvious)calculationthataddsuptheeconomicgainsandcomparesthatnumberwiththefiscalburden.Butanyonewithapencilandaproverbialback-of-an-envelopecandosousingthenumbersinthereport.TheonlytimetheNAScomesclosetoestimatingthetotalgainsiswhenitreportsthe"immigrationsurplus"--theincreaseintheaggregatewealthofnativesresultingfromtheproductivecontributionsofimmigrants.Althoughmuchisleftoutwhencalculatingthistheory-basedsurplus,itseemsevident(atleasttome)thatthebottomlineisverysimple:Theeconomicimpactofimmigrationis,atbest,anetwashfortheaveragenative-bornperson.Thegainsaccruingfromtheimmigrants'productivecontributionsareprobablyoffsetbythefiscalburden.Buteventhoughthemythicalaveragepersonisunaffected,somegroupsgainalotandsomegroupslosealot.
Finally,letmere-emphasizethatthisUser'sGuidefocusesontopicsthatIpersonallyfindinterestingandimportant.There'smuchmoreinthereport,including(longanddense)discussionsofimmigration,innovation,andeconomicgrowth,wherethefoundationalresearchisstillaworkinprogress.Nevertheless,theyprovideanexcellentintroductiontomanyresearchandpolicyquestions.
1.Assimilation
ItiswellknownthatimmigrantshaveaneconomicdisadvantagewhentheyfirstentertheUnitedStates.ManyarenotfluentinEnglish;theyarenotfamiliarwithhowtheUSlabormarketworks;andonandon.Soitisnotsurprisingthat,atfirst,theyearnfarlessthan
3
natives.Overtime,theimmigrantslearnthelanguage,acquirenewskills,andbeginto"catchup"orassimilatetothenativenorm.
Economicassimilationisobviouslyanimportantcomponentofanyassessmentoftheimpactofimmigration.Anddespiteallthehypeclaimingthatimmigrantstodayareassimilatingjustaswellasearlierwavesdid(includinganotherNASreportfocusingspecificallyonassimilationandreleasedjustoneyearago;here'sthatreportandthemediaspin),thenewNASreportgivesafarmorerealisticandmeasuredassessmentofthesituation.
Chapter3,whichsummarizestrendsintheskillsofimmigrants,isquitedetailed.Butthereisonetable(Table3-12)andonefigure(Figure3-6)thatspeakvolumesaboutwhatreallymatters.Here'sthetable:
Thistablereportstheage-adjustedpercentwagegapbetweenspecificimmigrantwavesandnativesatdifferentpointsintime.Thetableobviouslyshowsthattherewasalotofwagegrowthforimmigrantswhoarrivedinthe1960sand1970s,butfarlessforimmigrantswhoarrivedinthe1980sand1990s.Forexample,the1965-1969arrivalshada23.5%wagedisadvantageatthetimeofarrival,andthishadnarrowedtoa12.0%disadvantageafter10years.Butthe1995-99arrivalshada27.3%wagedisadvantageatthetimeofarrival,anditwasstill26.9%after10years.Asthereportmodestlyputsit:
Maleimmigrantswhoarrivedbetween1965and1969experiencedrapidgrowthintheirrelativewages,whichallowedthemtoclosethegapwithnative-bornpeers.Thisindicationofeconomicintegrationhasshownsignsofslowinginmorerecentdecades.Therelativewageprofilehasflattenedsomewhatacrossrecentarrivalcohorts,indicatingaslowingrateofwageconvergence.
Andthenhere'sthefigure:
4
Itshowsthatimmigrantswhocameinthe1970sbecamefluentinEnglishatafasterratethantheimmigrantswhoarrivedinthe1980sand1990s.Thisishowthepaneldescribesthefinding:
Thesetrendsgenerallycorroboratethefinding...thatearliercohortsofimmigrantsexperiencedmorerapidlanguageassimilationthanrecentcohorts.
TheNASanalysisoftrendsinimmigrantskillsbringstotheimmigrationdebateanewandimportantfact:Therehasbeenaslowdownintheeconomicassimilationofimmigrants.Eventhoughwedonotyetfullyunderstandwhythisslowdownoccurred,thereisacrucialquestionlurkingunderneath:Whatdoestheassimilationslowdownmeanforthefuture?
Anaside:BewareofadjectivesandadverbsintheNASreport.LookatTable3-12andFigure3-6,andaskyourself:Istheassimilationslowdownnumericallysignificant?Thisishowthereport'ssummarydescribesit:"TherateofrelativewagegrowthandEnglishlanguageacquisitionamongtheforeign-bornisnowslightlyslowerthanitwasforearlierimmigrantwaves."Maybeit'sjustme,butthefindingthatthereisnowagegrowthwhatsoeverbythetimeyougettothe1995-1999waveisatadstrongerthantheclaimthatassimilationis"nowslightlyslower."
5
2.LaborMarketImpact
Chapter5ofthereport,entitled"EmploymentandWageImpactsofImmigration,"weighsinatover32,000words(forcontext,that'soverhalfthelengthofmynewbook,WeWantedWorkers).Iamcynicalenoughtoknowthatmostofthepeoplewhowillbothertowadethroughtheverbiagearefishingfor"talkingpoints"thatwillsupporttheirideologicalpointofview.Butthey'llbemissingsomething.Thisis,byfar,thebestandmostextensivesurveyofadifficultandvoluminousliterature.Thereport'semphasisonthediversityoffindings,andthemanycaveatsthatgoalongwiththosefindings,reflectsthedoubtsanduncertaintyintheexistingacademicliterature.
Havingsaidthat,westillneed"stylizedfacts"tohelpusthinkaboutthisissue.Theverylongchapteronlyhastwotables(inthemaintext).Andthosetwotablessummarizethekeyinsightsfromtheliterature.Soletmedescribewhatthosetwotablesdo,andnotethetake-awaypoints.
Table5-1,copiedverbatimfromtablesthatIpublishedinmytechnicalbookImmigrationEconomics,summarizestheevidencefrom"structuralmodels"ofthelabormarket.InplainEnglish.Let'sassumethatalltheimmigrantswhoarrivedbetween1990and2010enteredthecountryallatonce.Wearethengoingtostreamthesedatathroughamathematicalmodelthatpurportstodescribehowthelabormarketworks.Thismathematicalexercisethenletsus"see"howthemarketreactsinthe"shortrun"(theinstantaftertheimmigrantsarrive)andthe"longrun"(afterthemarkethasfullyadjustedtotheirentry).
6
AlthoughIampersonallyresponsibleforintroducingthistypeofstructuralsimulationinthesecondhalfofmy2003QuarterlyJournalofEconomicspaper,I'mnotabigfanofit.Why?Becausethemathematicalmodelbuildsinmanyassumptions,andassumptionsmatter.Thisopensupthedoorforalotofmischiefandobfuscation,asdifferent
7
researchersplaywithdifferentassumptionsandendupproducingdifferentanswers.Letmequotethereportonwhatitiswelearnfromthistypeofstructuralanalysis:
Thekeydifferencesinthestructuralstudiesliteraturecanbelinkedbacktothestudies’modelingassumptions.Allowingcapitaltoadjust(i.e.,movingfromashort-runtoalong-runscenario)reducestheestimatednegativeeffectsacrosstheboard[GoingfromPanelAtoPanelBofthetable]...Thesimulationsalsoshowthatallowingforimperfectsubstitutionbetweenimmigrantsandnatives[goingfromScenario1toScenario2]doesnotgreatlyattenuatethewageimpactofimmigrationonhighschooldropouts.Thereisstilla2to5percentwageloss,dependingonwhetheronelooksatthelongrunorshortrun...Thescenariothatdoesleadtoamuchlowernegativeorevenpositiveimpactofimmigrationonthelowestskilledworkersistheonethatalsoincorporatesthepossibilitythathighschooldropoutsandhighschoolgraduatesareperfectsubstitutes[goingfromScenario1toScenario4].
Letmetranslateallthis.Twoassumptionshavebeenusedtoclaimthatimmigrantshaveonlyatrivialwageeffectonlow-skillnatives.Thefirstisthatlow-skillimmigrantsarenotproductive"clones"oflow-skillnatives--sothattheentryoflow-skillimmigrantsmayactuallybemakingthelow-skillnativesmoreproductive.ThisispreciselytheclaimfirstmadebyOttavianoandPeriadecadeago.Wenowknow,asPeri-coauthorEthanLewisconcludesinfootnote7ofhissurvey,thatthistypeofcomplementarityis,atbest,"verymodest."Notsurprisingly,theNASreportsthataccountingforthisissue"doesnotgreatlyattenuatethewageimpactofimmigrationonhighschooldropouts."Whatreallymattersisaddingintheotherassumption:thathighschooldropoutsandhighschoolgraduatesareproductiveclones.This,asthereportacknowledges,istheassumptiononeneedstogetthedatatofinally"confess"thatlow-skillworkersarenotharmedbyimmigration.
(Forthegeekyreader.Scenario2inTable5-1assumeslow-skillimmigrantscomplementlow-skillnatives;Scenario3assumeshighschooldropoutsandhighschoolgraduatesareproductiveclones;andScenario4assumesboth).
Theothertableinthechapter(Table5-2)skipsallthatmathandallthoseassumptions,andinsteadsummarizeswhatwefindwhenwesimplycorrelatewageswithimmigration(acrosscitiesorskillgroups).
8
Thistableisa"let-the-data-decide"kindoftable(inPanelsAandB).Ithinkthisisafarmorecredibleapproach.AndthisiswhattheNASreportsaysaboutthosecorrelations:
Somenotablepatternsemerge...Nativedropoutstendtobemorenegativelyaffectedthanbetter-educatednatives(asindicatedbycomparingresultsfordropoutswiththeoverallresultsforallworkersorallmenorwomen).Theresultsinthetablealsosuggestthatthisnegativeeffectmaybecompoundedfornativeminorities.AltonjiandCard(1991)foundmore-negativeresultsforlow-educationblacksthanlow-educationwhites...CortésexaminedanumberofgroupsandfoundthelargestnegativeeffectsforHispanicdropoutswithpoorEnglish,aswellaslargernegativeeffectsforHispanicdropoutsthanforalldropouts.Thiscouldbebecausenativedropoutminoritiesaretheclosestnativesubstitutesforimmigrants.
InplainEnglish:theactualdataindicatethatthosenativeswhoaremostlikelytobeaffectedbytheimmigrantsbecausetheysharesimilarskillsare,infact,thenativesmostaffectedbythoseimmigrants.ThereisadeliciousironyinTable5-2thatIcannotresistpointingout.Lookandseewhicheconomisthasproducedthemostnegativeimpactofimmigrationonthewageoflow-skillworkers.IthappenstobenoneotherthanDavidCard.[FullDisclosure:Anotherpanelmemberwasresponsiblefortheconstructionofthetable].
And,aftereverythingissaidanddone,theNASreportconcludes:
9
Whenmeasuredoveraperiodof10yearsormore,theimpactofimmigrationontheoverallnativewagemaybesmallandclosetozero.However,estimatesforsubgroupsspanawiderrangeandsuggestsomerevisionsinunderstandingofthewageimpactofimmigrationsincethe1990s...Theintensiveresearchonthistopicoverthepasttwodecades,summarizedinTable5-2,displaysamuchwidervariationintheestimatesofthewageimpactonnativeswhoaremostlikelytocompetewithimmigrants,withsomestudiessuggestingsizablenegativewageeffectsonnativehighschooldropouts...Thus,theevidencesuggeststhatgroupscomparabletotheimmigrantsintermsoftheirskillmayexperienceawagereductionasaresultofimmigration-inducedincreasesinlaborsupply,althoughtherearestillanumberofstudiesthatsuggestsmalltozeroeffects.
Letmeaddanimportantcaveattothisquote.Thezeroaveragewageeffectinthelongrun("10yearsormore")isbasedonthestructuralestimatesreportedinTable5-1.Takealookatthelastcolumnofthattableandnotethatthelong-runimpactofimmigrationontheaveragewageofworkersisalwaysexactlyequalto0.0percent,regardlessofwhichscenariowelookat.Whataremarkablestatisticalcoincidence!
Asthepanelitselfacknowledges,however,thiszerowageeffectisbuiltinbythemathematicsofthemodel:"Inthecaseofstructuralstudies,whencapitalisassumedtobeperfectlyflexible,[average]wageeffectsonnativesarezero,althoughthisresultisbuiltinbytheoreticalassumptions."Putbluntly,claimsthatthelong-runeffectofimmigrationontheaveragewageis"smallandclosetozero"havenothingtodowiththedata.Thatresultisinsteadaby-productofamathematicalassumptionusedtoconstructthemodelofthelabormarket.
And,tomakemattersworse,thismathematicalassumptioncascadesovertoeveryothernumberreportedinTable5-1.Afterall,thewageeffectsforthevariousskillgroupsmustaverageouttozero.Thismeansthateachparticularwageimpactneedsto"alignitself"aroundzerosothattheweightedaverageoftherelevantnumbersindeedaddsuptothemathematicallybuilt-in0.0longrunwageeffect.Putbluntly:Table5-1shouldcomestampedwithabigUsersBewaresign.
3.Short-RunFiscalImpact
TheNASpanelcalculatedtheshort-runfiscalimpactbycomparingthecostofprovidingpublicservicestoimmigrantswiththetaxesthatthoseimmigrantspayinaparticularyear.BothChapter8andChapter9giveestimatesoftheshort-runfiscalimpact.Chapter8includesfederalexpendituresandtaxeswhencalculatingtheimpactin2013,whileChapter9focusesontheimpactatthestateandlocallevelfortheyears2011-2013.
Table8-2isthekeyshort-runtableinChapter8.TheNASpanelusedeightalternativescenariostocalculatethefiscalimpact.
10
IthinkScenarios1and5arethemostinterestingones,androughlydefinetheextremes.Scenario1assumesthatthecostofprovidingpublicgoods(suchasnationaldefense)isthesameforanimmigrantasitisforanative,whileScenario5assumesthatimmigrantsdonotincreasethecostofpublicgoodsatall.Thereare55.5millionpersonswhoareeitherimmigrantsortheminorchildrenofimmigrants(thatis,theirdependents).AndthisishowthereportdescribesthefiscalimpactofthisgroupunderScenario1:
Thetotalfiscalburdenis$279billionforthefirstgenerationgroup(averageoutlaysof$15,908minusaveragereceiptsof$10,887,multipliedby55.5millionindividuals).
ItisequallyeasytoestimatethefiscalburdeninScenario5,whereimmigrantsareassumednottoincreasethecostofpublicgoodsatall.Theaverageoutlayisthen$11,669andtaxreceiptsremainat$10,887,creatingafiscalburdenof$43.4billion.ThereporttriestoputthesestatisticsincontextbynotingthattheUnitedStatesrunsafiscaldeficitexceedingoveratrilliondollarsayear,sothateveryoneisafiscalburden.Leftunsaidisanequallyimportantpoint:Someburdensareavoidable,andsomeburdensarenot.Wemaynotbeabletodomuchaboutthefiscalburdenofthenative-bornpopulation.Buttherearemanyobviouspolicyoptionsavailabletoensurethatthealready-largeburdenisnotfurtherincreasedbyimmigration.
Chapter9presentsmoredetailedestimatesoftheshort-runfiscalimpact,focusingonexpendituresandtaxesatthestateandlocalgovernmentlevel.Thepanelallocatedexpenditureson"local"publicgoods(suchaspublicsafety,hospitals,andlibraries)onaper-capitabasistoimmigrantsandnatives.Table9-6isthekeytablethatsummarizestheevidence,andreportsthetaxesandexpendituresforthefirstgeneration(theimmigrants
11
andtheirdependents).Hereistherelevanttable,whereIcutoutmostofthestate-specificrowstomakeitmorereadable(hereisthefulltable):
Nationwide,thetypicalimmigrantgeneratesafiscalshortfallatthestate-locallevelof$1,600annually.Thereare36.1millionsuchfirst-generation"households"(seeTable9-13inthereport),sothatthetotalshortfallisover$57billion(or$1,600times36.1million).Thisishowthereportdescribesthatfinding:
Firstgenerationindependentpersonunits(whichincludefirstandsecondgenerationchildrenassignedtoindependentfirstgenerationpersons)costthestatesonnetabout$1,600each...Theseestimatesofthefiscalimpactimplythatthetotalannualaggregateimpactofthefirstgenerationandtheirdependents,averagedacross2011-13,isacostof$57.4billion.
Thedataaresounambiguousthatitisveryeasytosummarizewhattheysay.Onayear-to-yearbasis,thetaxesthatimmigrantpaysimplydonotcoverthepublicexpenditurestheytrigger.Andtheshortfallseemstobeatleast$50billionannually.
4.Long-RunFiscalImpact
Bylookingonlyatexpendituresandtaxesduringagivenyear,thecalculationoftheshort-runfiscalimpactignoresthatsomeofthoseexpendituresactuallyyieldareturn.Thecostofsendingthechildrenofimmigrantstoschooltodayleadstohigherearningsforthosechildreninthefuture.Plustheagingofthenative-bornpopulationiscreatingseverefiscalproblems,asthereisnotenoughmoneytofundtheliabilitiesinSocialSecurityandMedicareunlesswedrasticallyraisetaxesorcutbenefits.Immigrationbringsinnewtaxpayerswhocanfundsomeofthoseliabilitiesinthefuture.
Chapter8oftheNASreportpresentsthecalculationofthelong-runfiscalimpact.Toseehowthisisdone,imaginethefollowingsequenceofevents.Animmigrantarrivestoday,
12
payingtaxesandreceivingpublicservices.Thatimmigranthaschildren.Thosechildrenmaybecostly,buttheyeventuallygrowupandpaytaxes.Thechildrenhavechildren,andtheprocessgoeson.Thepaneldidthiscalculationby"tracking"theimmigrantandalldescendantsoverthe75-yearperiodafterarrivalandaddingupallthetaxespaidandexpendituresincurred.Thedifferencebetweentotaltaxesandtotalexpendituresisthelong-runfiscalimpact.
Table8-12isthekeytableinthechapter.
13
Thetablehastwopanels.Thetoppanelreportsthelong-runimpactifwetrackedthetypicalimmigrantwhoarrivedbetween2011and2013,whilethebottompanelreportsthelong-runimpactifwetrackedthetypicalimmigrantnowlivingintheUnitedStates.Thetrackingoftheimmigrantwhoarrivedbetween2011and2013canbeverymisleading--justimaginewhatthefiscalimpactwouldlooklikeiftheUnitedStatessuddenlydecidedtoadmit300,000refugeesandwethentrackedthattypicalimmigrant.Thereareblipsinwhotheimmigrantsarefromyeartoyear,with"good"and"bad"years.Toavoidslantingthenumbersinanyparticulardirection,itisfarpreferabletotracktheaverageimmigrantinthecountry.
Thetableusesfourdifferentscenariostocalculatethelong-runimpact.Thescenariosdifferonwhattheyassumeaboutwhetherimmigrantsincreasethecostofpublicgoods,andonwhattheyassumeaboutthepathoftaxesandexpendituresoverthenext75years.AndthepanelhelpfullyaddedniceyellowhighlightstoTable8-12thatisolatethekeynumberresultingfromeachscenario.
Thelong-runfiscalimpactispositiveonlyifimmigrantsdonotaffectthecostofpublicgoodsandweassumethatfuturetaxratesandbenefitpaymentswillfollowtheprojectionsmadebytheCongressionalBudgetOffice(CBO).Ifyougetridofeitherofthoseassumptions,thepositivelong-termimpactofanimmigrantcontributinganetof+$58,000overthenext75yearsbecomesalossaslargeas-$119,000.TheroleofassumptionsingeneratingtheanswerleadstoseveralboldedbulletpointsintheNASreport:
Forward-lookingprojectionsofthenetfiscalimpactofanadditionalimmigrantanddescendantsgeneratearelativelywiderangeofpossibleresults.
Thefuturepathoffiscalpolicyisimportantforassessingthefiscalimpactsofimmigrants.
Thetreatmentofspendingonpublicgoodsisimportantforassessingthefiscalimpactofimmigrants.
Letmetranslate.Assumptionsmatter,anddifferentassumptionsleadtowildlydifferentanswers.IthinkthereisanelephantintheroomthattheNASreportalludesto,butcannotbringitselftosayoutloud.SoIwill:Allestimatesofthelong-runfiscalimpactareuseless!
Itisextremelyeasytomanipulateassumptionsandendupwitheitherlargepositiveorlargenegativelong-runimpacts.Doyouwantalargepositivenumber--assomepeopleinthedebatesurelydo?Thenpickayearwherethenewimmigrantslookparticularly"good,"assumethattaxeswillgoupinthefuture,andignorepublicgoods.Doyouwantalargenegativenumber--assomeotherpeopleinthedebatesurelydo?Thenpicka"bad"year,assumeimmigrantsincreaseexpendituresinpublicgoods,andassumethattaxesandexpendituresstayintheircurrentpathfortheremainderofthiscentury.
Iwouldalsoadd:Don'tbefooledbytheCBO"experts"whoclaimtoknowhowtaxesandexpenditureswillevolveoverthenext75years.Thosesameexpertscouldn'tevenpredict
14
Obamacareenrollmentjustafewmonthsago.Whatdotheyreallyknowabouttaxesandexpendituresintheyear2075?
Onefinalpoint.Table8-12alsoreportsthelongrunfiscalimpactforimmigrantsineacheducationgroup.Regardlessofscenario,itisobviousthatlow-skillimmigrantsimposeafiscalburdeninthelongrun,butthatimmigrantswithatleastacollegeeducationarefiscallybeneficial.
5.TheBottomLine
Immigrantshavebothalabormarketimpactandafiscalimpact.Dotheeconomicgainsgeneratedbyworkingimmigrantsoutweighthefiscalburdenthatimmigrantsimpose?TheNASreport(probablywisely)avoidsputtingtwoandtwotogether,butthereportcontainsallthenecessaryingredientstoletusdoitourselves.Solet'stakeacrackatit.
Thereisafiscalburden.Acrossalllevelsofgovernments,theannualburdenrangesfromaminimumof$43billiontoamaximumof$299billion,dependingonwhatisassumed(Table8-2ofthereportshowsalleightscenarios).AsInotedearlier,theestimatesofthelong-runimpact,whichincorporatesthetaxesandexpendituresoftheimmigrantanddescendantsovera75-yearperiod,areuselessandeasilymanipulatedtoproducewhateverlargepositivenumberorlargenegativenumberonewants.
Nowlet'sfindoutwhatthereportsaysaboutthe"immigrationsurplus,"theincreaseinwealthaccruingtothenativepopulationasaresultofimmigration.Asimmigrantsenterthelabormarketandreducethewageofnatives,theyincreaseprofitsfortheemployers.Plustheimmigrantsthemselvesproduceadditionaloutput,generatingevenmoreprofits.Intheend,theaggregatewealthofnatives--bothworkersandfirms--rises,andthereisaredistributionofwealthfromworkerstofirms.ThereportpresentsitsestimateoftheimmigrationsurplusinChapter4:
Immigrantlaboraccountsfor16.5percentofthetotalnumberofhoursworkedintheUnitedStates,which...impliesthatthecurrentstockofimmigrantsloweredwagesby5.2percentandgeneratedanimmigrationsurplusof$54.2billion,representinga0.31percentoverallincreaseinincomethataccruestothenativepopulation.
Thisshortparagraphcontainsalotofimportantinformation.First,theimmigrationsurplusisrelativelysmall,about$54billion.Unfortunatelythereportdoesnotgiveatransparentestimateofthesizeofthewealthtransferfromworkerstofirms,reportinginsteadthat,onaverage,wageswentdownby5.2percent.Itwouldbebetteriftheyhadreportedthenumberofdollarsinvolvedinthattransfer.Thatnumber,itturnsout,wouldbeabout$500billion.(Ageekyfootnoteattheendofthispostexplainshowyoucangetthatnumber).So,yes,immigrantscreatedanadditional$54billionworthofnewwealth,butabyproductofthatcreationwasawealthtransferofhalf-a-trilliondollars.
15
Thereportcautionsthatthisisanestimateoftheshort-runeconomicgains,basedontheassumptionthattheeconomyhasnotyetadjustedtotheentryofimmigrants.Obviously,the42millionimmigrantsnowinthecountryenteredoveraperiodofmanyyears.Andeconomictheoryimpliesthat,overtime,ascapitaladjusts,theimmigrationsurplusdwindlesdowntoalmostnothing.Asthereportputsit:
Overthecourseofdecades...capitalhashadplentyoftimetoadjust,andsotheseestimatescanatbestbedescribedasupperlimitsthatexaggeratetherealimpactofimmigrationonnativewagesandoverallincomes.
Iwouldaddanotherhugecaveattothe$54billionestimateofthesurplus.Itignoresalltheexternalitiesthatimmigrantscreatealongtheway.Theexternalitiesarebothgood--theentryofextremelyhigh-skillimmigrantssurelyacceleratesinnovation,makesusmoreproductive,andhasabeneficialimpactoneconomicgrowth.Andbad--theentryofsomehigh-skillimmigrants,suchasthosewhoenrolledinflightschoolsandlearnedtoflyplanesandthenflewthemonSeptember11,2001,canmakeusallmuchworseoff.Thepaneldidnoteventrytoquantifythevalueofallthemanypositiveandnegativeexternalities(and,infact,neitherhasanybodyelse).So,intheend,allwereallyhavetogoonisanestimatedsurplusof$54billionintheshortrun.
Ifwethentakethereport'sestimatesofthesurplusandthefiscalburdenatfacevalue,itisself-evidentthat:
Theimpactofimmigrationontheaggregatewealthofnativesis,atbest,awash.
Instead,theimpactofimmigrationisdistributional.Thosewhocompetewithimmigrantsareeffectivelysendingbillionsandbillionsofdollarsannuallytothosewhouseimmigrants.
Toreiterate,theimpactofexternalitiescanradicallychangethisconclusion(ineitherdirection).Butnotethatevenifbeneficialexternalitiesdominated,theywouldhavetobeawfullyimportant--theywouldneedtoquintuplethecurrentestimateoftheimmigrationsurplusfrom$54billionto$270billion--tooffsetthehigh-endestimatesofthefiscalburden.
______________________________________________________
Forthegeekyreader:
Thisishowtocalculatethat$500billiontransferinthebackofanenvelope.ThecalculationoftheimmigrationsurplusreportedinChapter4oftheNASreportassumesthatGDPis$17.5trillion;that65%ofGDPgoestoworkers;andthat16.5%percentoftheworkforceisforeign-born.Thereportalsosaysthat"thecurrentstockofimmigrantsloweredwagesby5.2percent."
16
Becauseonly65%ofGDPgoestoworkers,thatmeansthatthetotalearningsofallworkersis$11.4trillion(or0.65×17.5).Butbecauseonly16.5%ofworkersareforeign-born,thefractionoftotalearningsthatgoestonativeworkersis$9.5trillion(or0.835×11.4).TheNASreportsaysthatnativeearningsfellby5.2percent,sothatthewagetransferfromnativeworkerstoemployersis$494billion(or0.052×9.5).