Time200gana-894
FINALREPORTGrowingWageInequality,theMinimumWage,andtheFutureDistributionofRetirementIncomeDamirCosicRichardW.
JohnsonKarenE.
SmithJuly2018PROGRAMONRETIREMENTPOLICYABOUTTHEURBANINSTITUTEThenonprofitUrbanInstituteisaleadingresearchorganizationdedicatedtodevelopingevidence-basedinsightsthatimprovepeople'slivesandstrengthencommunities.
For50years,Urbanhasbeenthetrustedsourceforrigorousanalysisofcomplexsocialandeconomicissues;strategicadvicetopolicymakers,philanthropists,andpractitioners;andnew,promisingideasthatexpandopportunitiesforall.
Ourworkinspireseffectivedecisionsthatadvancefairnessandenhancethewell-beingofpeopleandplaces.
CopyrightJuly2018.
UrbanInstitute.
Permissionisgrantedforreproductionofthisfile,withattributiontotheUrbanInstitute.
CoverimagebyTimMeko.
ContentsAcknowledgmentsivExecutiveSummaryvWageInequality,theMinimumWage,andFutureRetirementIncomes1Background4ExplainingRisingInequality6ImpactofMinimumWageIncrease8Methods9DYNASIM411EstimatingtheEducationPremiumModel12EstimatingtheWageModel16DYNASIM'sEmploymentandHoursModel18SimulatingtheMinimumWage19ValidatingSimulationOutcomes25Results27SharedLifetimeEarningsProjections28NetPerCapitaAnnuityIncomeatAges67to7531Conclusions39Appendix42Notes54References55AbouttheAuthors58StatementofIndependence60IVACKNOWLEDGMENTSAcknowledgmentsThisreportwasfundedbytheUSDepartmentofLabor'sEmployeeBenefitsSecurityAdministration,undercontractDOL-OPS-14-D-0020,ordernumberDOL-OPS-16-T-00148.
Wearegratefultothemandtoallourfunders,whomakeitpossibleforUrbantoadvanceitsmission.
TheviewsexpressedarethoseoftheauthorsandshouldnotbeattributedtotheUrbanInstitute,itstrustees,oritsfunders.
FundersdonotdetermineresearchfindingsortheinsightsandrecommendationsofUrbanexperts.
FurtherinformationontheUrbanInstitute'sfundingprinciplesisavailableaturban.
org/fundingprinciples.
ThisreportusestheUrbanInstitute'sDynamicSimulationofIncomeModel4(DYNASIM4).
KarenSmithandMelissaFavreaultdirectDYNASIM4development.
DouglasMurraywasDYNASIM4'sleadprogrammerfrom1998through2017,andDamirCosichasbeenleadprogrammersince2018.
CurrentandrecentfundersofDYNASIM4includetheUSDepartmentofLabor,AARP,AlfredP.
SloanFoundation,AndrewW.
MellonFoundation,FordFoundation,KaiserFamilyFoundation,NationalInstituteonAging,OfficeoftheAssistantSecretaryforPlanningandEvaluationattheUSDepartmentofHealthandHumanServices,RockefellerFoundation,SCANFoundation,andSocialSecurityAdministration.
TheauthorsgratefullyacknowledgevaluablecommentsonpreviousdraftsfromMelissaFavreaultandresearchstaffattheOfficeofPolicyResearchattheEmployeeBenefitsSecurityAdministrationandeditorialassistancefromRachelKenney.
EXECUTIVESUMMARYVExecutiveSummaryWageinequalityintheUnitedStateshasbeenrisingsincethe1970s.
Althoughitsgrowthhasslowedinrecentyears,theforcesdrivingwageinequalitypersist.
Increasedwageinequalityhasespeciallyhurtworkersnearthebottomofthewagedistribution,whereinflation-adjustedearningshavefallenoverthepastfewdecades.
Changesinthedistributionofwagesshapeworkers'lifetimeearningsandaffectthedistributionofretirementincome.
Peoplewhoexperiencehighwageinequalityduringtheirworkingyearsarelikelytoexperiencehighretirementincomeinequality,becauseSocialSecuritybenefitsaretiedtolifetimeearnings,andpeople'sabilitytosaveforretirementdependsonhowmuchtheyearn.
Thisstudyexaminestheeffectofrisingwageinequalityonlifetimeearningsandfutureretirementincomeandevaluatesthepotentialeffectivenessofthefederalminimumwageinmitigatingthedistributionalconsequencesofrisingwageinequality.
TheanalysisusestheUrbanInstitute'sDynamicSimulationofIncomeModel4(DYNASIM4)toprojectfutureearningsandretirementincomeovera70-yearperiodundertheassumptionthatwageinequalitycontinuesrisingandcomparesoutcomeswiththebaselineinwhichwageinequalityremainsconstantattoday'slevel.
Thestudyalsoprojectsfutureearningsandretirementincomeundertheassumptionthatthefederalminimumwageincreasesin2017fromitscurrentvalueof$7.
25perhourto$12perhourandsubsequentlyadjustswithinflation.
Theinvestigatorsprojectwageinequalitybasedonrecenttrendsinthegrowthoftheeducationpremium—thehourlywagegapbetweencollege-educatedworkersandthosewithlesseducation—whichhasbeenoneofthemostsignificantandpersistentdriversofwageinequality.
Intheprojections,theeducationpremiumgrowsatadiminishingrate,peaksaround2070,andremainsconstantthereafter.
Atitspeak,theeducationpremiumraisesprojectedwageinequalityabout15percentabovethebaseline.
Thestudycomparesmeanincomeatages67to75byquintilesoflifetimeearnings.
Theincomemeasureaddstocashincomethevalueofimputedrentfromowner-occupiedhousingandtheincomestreamthatahouseholdwouldreceivebyannuitizing80percentofitsfinancialassetsandsubtractstaxesandMedicarepremiums.
Themeasurealsodividesthisamountbytwoformarriedcouples.
Theresultsshowthattheeffectsofrisingwageinequalityreverberateintoretirement.
Ifthehourlywagegapbetweencollegeandhighschoolgraduatescontinuestogrow,ourprojectionsshowthefollowingforpeopleages67to75:VIEXECUTIVESUMMARYLifetimeearningsinthetopfifthofthedistributionwouldrise2percentin2045,5percentin2065,and8percentin2085,relativetoprojectedlevelsunderthebaselinewhentheeducationpremiumdoesnotchange.
Annualretirementincomesinthetopfifthofthelifetimeearningsdistributionwouldrise3percentin2045,5percentin2065,and7percentin2085.
Inthebottomfifthofthedistribution,lifetimeearningswouldfall2percentin2045,5percentin2065,and9percentin2085.
Annualretirementincomesinthebottomfifthofthelifetimeearningsdistributionwouldfall3percentin2045,6percentin2065,and13percentin2085.
Theselossesexceedthepercentagedeclineinlifetimeearnings,despiteSocialSecurity'sprogressivebenefitformulathatreplacesalargershareofpreretirementearningsforpeoplewithlimitedlifetimeearningsthanforthosewithmorelifetimeearnings.
Theinequalityofretirementincomewouldriseovertime,pushingtheGinicoefficient.
03pointshigherthanunderthebaselinein2085.
Ahigherminimumwagewouldmitigatebutnoteliminatetheeffectsofgrowingwageinequalityonfutureretirementincomes.
Immediatelyraisingthefederalminimumwagefrom$7.
25to$12perhourandsubsequentlyadjustingitforinflationwouldraiselifetimeearningsandfutureretirementincomeforpeopleinthebottomtwo-fifthsofthelifetimeearningsdistribution,buttheimpactwoulddissipateovertime,despitethebuilt-ininflationadjustment.
Asaveragewagesgrowfasterthanprices,theshareofworkerswhowouldotherwiseearnlessthantheminimumwagewouldfalleachdecade.
Forretireesinthebottomfifthofthelifetimeearningsdistribution,thesimulatedincreaseintheminimumwagewouldoffset57percentoftheretirementincomelosttorisingwageinequalityin2065and37percentoflostretirementincomein2085.
Boostingtheminimumwageandmaintainingitsvalueovertimeasaveragewagesrisewouldhelplow-wageworkersandimprovetheirretirementprospects,butabroaderpolicyagendaisneededtoprovideallworkerswithapathtoretirementsecurity.
Optionsincludeworkertrainingandapprenticeshipprograms;financialliteracytraining;taxreformsthatwouldbetterincentivizeretirementsavingsforlow-wageworkers;andSocialSecurityreforms,suchascreatingameaningfulminimumbenefitorchangingthebenefitformulatoimproveprogressivity.
WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES1WageInequality,theMinimumWage,andFutureRetirementIncomesWorkersenteringthelaborforcetodayfacehistoricallyhighratesofwageinequality,andtheyarelikelytocontinuetorise.
ThedistributionofwagesintheUnitedStateshaschangeddramaticallysincethe1970s.
Afteraperiodofrelativestabilitythroughoutthe1950sand1960s,wageinequalitygrewrapidlyoverthefollowingthreedecades.
InequalityisoftenmeasuredbytheGinicoefficient,whichrangesfrom0,indicatingcompleteequality,to1,indicatingcompleteinequality,withonepersonreceivingallincomeorwealth.
1TheGinicoefficientformen'searningsincreased0.
8pointsbetween1950and1979and10.
9pointsbetween1980and2004(figure1).
Becausethedriversofeconomicinequality,includingtechnologicaladvances,increasedinternationaltrade,anddecreasedunionization,persist,inequalitywilllikelycontinuetorise,despitearecentslowdowninitsgrowth.
Changesinthedistributionofwagesshapeworkers'lifetimeearningsandfutureretirementincome.
Cohortsthatexperiencehighwageinequalityduringtheirworkingyearsarelikelytoexperiencehighretirementincomeinequality.
TheGinicoefficientforincomeatages65to70forthecohortbornbetween1940and1949,whoworkedthroughouttheperiodofrisingwageinequality,wasjustbelow0.
4,almostfourpointshigherthantheGinicoefficientforthecohortbornbetween1920and1929,whoexperiencedrisingwageinequalityonlyattheendoftheirworkingyears(OECD2017).
Futureretireeswillhaveworkedduringperiodsofevenhigherwageinequality,suggestingthatretirementincomeinequalitywillgrow,despitetheequalizingeffectofSocialSecurity'sprogressivebenefitformula(Smith2003).
Risingwageinequalityhasbeenespeciallydetrimentaltoworkersnearthebottomofthewagedistribution.
Whiletheaveragerealwagehasbeenrising,theconcentrationofgainsamonghigher-wageworkershasreducedwagegrowthforlower-wageworkers.
Oneofthemainreasonsforthisdeteriorationhasbeenthegrowingdemandforhighlyskilledworkers,whichraisestheirrealwagesandreduceswagesforotherworkers.
Autor(2014)foundthattheriseinthepremiumtopostsecondaryeducationaccountedforabouttwo-thirdsoftheobservedincreaseinwageinequalitybetween1980and2005.
Thesteadypaceoftechnologicalprogressandtherelatedgrowthinthedemandforhighlyskilledworkerssuggestthatthiseducationpremiumwillcontinuetorise.
2WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE1GiniCoefficientofEarnings,1937–2004URBANINSTITUTESource:Kopczuk,Saez,andSong(2010).
Anotherfactorthatshapesthebottompartofthewagedistributionistheminimumwage.
Therealvalueofthefederalminimumwagefellsharplyduringthe1980sandhasfluctuatedeversince.
Severalstudiesfoundthatthedecreaseintheinflation-adjustedvalueofthefederalminimumwagefrom1979to1988explainedasubstantialpartofthedropintheratioofthe10thpercentileoftheearningsdistributiontothe50thpercentileoverthatperiod(Lee1999;DiNardo,Fortin,andLemieux1996;Autor,Manning,andSmith2016).
Raisingtheminimumwagecouldmitigategrowingearningsinequality.
Inadditiontoboostingcurrentearningsforlow-wageworkers,raisingtheminimumwagecouldimprovetheirfutureretirementincomesbecausehigherwageswouldallowworkerstoaccumulateadditionalSocialSecuritycredits,andtheymightbeabletosavesomeoftheirincreasedearningsforretirement.
Toexploreretirementprospectsfortoday'sworkersandthestepsthatpolicymakersmighttaketoimprovethem,weexaminedhowwageinequalityandthefederalminimumwagemightshapefutureretirementincomes.
Ourstudyaddressedtworesearchquestions:0.
30.
320.
340.
360.
380.
40.
420.
440.
460.
480.
51940195019601970198019902000GiniYearMenWomenAllworkersWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES3HowwouldrisingwageinequalityaffectretirementincomesforfutureretireesHowmuchwouldanincreaseinthefederalminimumwagemitigatetheimpactofrisingwageinequalityonthedistributionofretirementincomesWeapproachedthesequestionsbysimulatingfutureretirementincomesbasedonalternativescenariosoffuturewagegrowthandminimumwagelevels.
UsingtheDynamicSimulationofIncomeModel4(DYNASIM4),theUrbanInstitute'sstate-of-the-artdynamicmicrosimulationmodel,wesimulatedthreescenarios:1.
Abaselinescenariothatassumedthatthecurrentlevelofwageinequalityandthecurrentnominalfederalminimumwageof$7.
25continuedthroughoutthesimulationperiod2.
Ahigh-inequalityscenariothatraisedwageinequalityoverthesimulationperiodwhilekeepingtheminimumwageatitscurrentlevel3.
Apolicyinterventionscenariothatraisedwageinequalityandincreasedthefederalminimumwageto$12.
00in2017,adjustedforinflationinsubsequentyearsOurprojectionshighlightedthefuturedistributionoflifetimeearnings,retirementsavings,andretirementincome.
Toaddressthefirstresearchquestionabouthowrisingwageinequalitymightaffectfutureretirementincome,weestimatedthedifferenceinthedistributionsofretirementincomeunderscenarios1and2.
Ouranalysisfocusedonthefutureevolutionoftheeducationpremium,definedasthewagedifferentialbetweenworkerswithandwithoutafour-yearcollegedegree.
Raisingtheearningsgrowthrateforcollegegraduatesinscenario2,whilemaintainingthesamegrowthrateforthemeaneconomy-widewageasinscenario1,boostedtheearningsgapbetweenless-educatedworkersandtheirbetter-educatedcounterparts.
ThisincreaseinwageinequalityaffectedthedistributionofretirementincomethroughchangesinSocialSecuritybenefits,employer-sponsoredretirementplans,andotherretirementsavings.
BecauseSocialSecuritybenefitsarebasedonworkers'lifetimeearnings,thedistributionofbenefitswillbecomemoreunequalwhenthedistributionofearningsbecomesmoreunequal,althoughtheprogressivebenefitformulaprovidessomeprotectiontolow-wageworkers.
Similarly,becauseworkerssaveforretirementoutoftheirearnings,retirementsavingsinequalitywillincreaseasearningsinequalitygrows.
Ouranalysisconsideredonlythegrowthininequalitythatarisesfromtheincreasingeducationgapinwages,whichisthemostimportantsourceofeconomicinequality.
Weaddressedthesecondresearchquestion,abouthowraisingthefederalminimumwagemightmitigatetheimpactofgrowingwageinequalityonretirementincomes,bycomparingretirement4WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESincomedistributionsunderscenarios2and3.
Althoughthemeanwagegrewatthesamerateunderthetwoscenarios,anincreaseinthefederalminimumwageunderscenario3compressedthewagedistribution.
Workingthroughthesamemechanismsasabove(SocialSecuritybenefits,employer-sponsoredretirementplans,andotherretirementsaving),thischangeinthewagedistributionalteredthedistributionofretirementincome.
Thedifferenceintheretirementincomedistributionsunderthetwoscenariosindicateshowaminimumwageincreasewouldaffectthedistributionofretirementincomeandpotentiallymitigatetheimpactofrisingwageinequality.
Oursimulationsshowthatretirementincomeswillbecomeincreasinglyunequaliftheearningsgapbetweencollegegraduatesandhighschoolgraduatescontinuestogrow.
Risingwageinequalitywouldreducemeanincomesforadultsages67to75inthebottomfifthofthelifetimeearningsdistributionby3percentin2045,6percentin2065,and13percentin2085,whileincreasingincomesforthoseinthetoplifetimeearningsquintileby3percentin2045,5percentin2065,and7percentin2085.
Raisingthefederalminimumwageto$12perhourin2017andtyingittosubsequentpriceinflationwouldboostlifetimeearningsforlow-wageworkersand,forolderadultsinthebottomfifthofthelifetimeearningsdistribution,offsetmuchoftheretirementincometheywouldotherwisehavelosttogrowingwageinequality.
However,theminimumwage'simpactwoulderodeovertimeasaveragerealwagesgrow.
In2045,theminimumwageincreasewouldcompletelyoffsetthelossinretirementincomefromrisingwageinequality.
In2065,itwouldoffsetjustlessthanone-halfoftheloss,andin2085,itwouldoffsetlessthanfour-tenthsoftheloss.
Raisingtheminimumwagewouldbelessbeneficialforthoseinthesecondandthirdquintilesofthelifetimeearningsdistribution.
BackgroundRetirementincomedependslargelyonhowmuchpeopleearnedwhentheywereyounger.
Aspeopleearnmoreovertheirlifetime,theyaccumulatemorecreditstowardfutureemployer-sponsoreddefinedbenefit(DB)pensionbenefits—iftheyarecoveredbyanemployerplan—andSocialSecuritybenefits,whicharebothtiedtopastearnings.
Inaddition,higherearningsgenerallyenablepeopletosavemoreforretirementthrougha401(k)-typeretirementplanoranothervehicle.
Consequently,asearningsbecomemoreunequallydistributed,retirementincometendstobecomemoreunequal.
Thecumulativeadvantagetheoryemphasizesthatearlylifeadvantagescumulateoverthelifecourse,magnifyingeconomicinequalityinlaterlife(FerraroandShippee2009;O'Rand1996,2003).
However,SocialSecurity,whichaccountsforabouthalfofallincomereceivedatages65andolder(BeeandMitchell2017),replacesalargershareofpreretirementearningsforlow-earningworkersthanforhigh-earningWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES5workers,mitigatinginequalityatolderages.
Nonetheless,CrystalandShea(1990a,1990b)foundthatincomeinthe1980swasmoreunequallydistributedatolderagesthanyoungerages.
Inafollow-upstudy,Crystal,Shea,andReyes(2016)foundthatlater-lifeinequalitywasgreaterin2010thaninthe1980s,butthattheagedifferenceininequalityhadfallen.
Recentincreasesinwageinequalitywilllikelyraisefutureeconomicinequalityatolderages.
Althoughtheaveragerealhourlywagehasgrownoverthepast35years(BoardofTrustees2016),thegainshavebeenconcentratedamongworkersinthetoppartoftheearningsdistribution.
Figure2showsthepercentagechangeintherealhourlywagefrom1980to2012bythepercentileofthewagedistributionandbysex,basedondatafromtheCurrentPopulationSurvey(CPS).
Mostofthewagegrowthwasconcentratednearthetopofthewagedistributionforbothmenandwomen,whilewagesinthemiddleandnearthebottombarelychangedoverthe32-yearperiod,especiallyformen.
Aboutone-halfofemployedmenexperiencednorealwagegrowth,andrealwagesincreasedby20percentormoreonlyformeninthetop15percentofthewagedistribution.
Workingwomeninthemiddleofthewagedistributionfaredbetter,withabout7in10experiencingrealwagegrowthofatleast20percent.
Thewagegapbetweenwomenandmendeclinedasgendergapsineducation,occupationalrepresentation,andunioncoverageshrank(BlauandKahn2016).
Thefiguredoesnotreportgrowthratesforthetop1percentofthedistribution,however,becausetheCPStop-codesearningsandundersamplespeoplewiththehighestincomes.
Gainsinannualearnings,aslightlydifferentconcept,havebeenconcentratedattheverytopofthedistribution,perdatacompiledbyPikettyandSaez(2006)andtheir2012update.
Theyfoundthatbetween1980and2011,averageearningsincreased135percentinthetop1percent,254percentinthetop0.
1percent,and455percentinthetop0.
01percent.
6WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE2PercentageChangeintheLogRealWagebetween1980and2012byPercentileURBANINSTITUTESource:Authors'calculationsfromthe1981and2013CPS.
Notes:Thehorizontalaxisrepresentspercentilesofthewagedistribution,andtheverticalaxisrepresentsthepercentagedifferenceinthehourlywagebetween1980and2012.
Thesamplewasrestrictedtoworkersages16to64.
Wageswereweightedbythesurveysamplingweightmultipliedbythenumberofhoursworked.
ExplainingRisingInequalityMuchofthegrowthinwageinequalitycanbeattributedtotherisingearningsgapbetweenworkerswithandwithoutacollegeeducation(Autor2014),whichinturnhasbeendrivenbyskill-biasedtechnologicalchange.
Theskill-biasedtechnologicalchangehypothesis,firstproposedbyTinbergen(1974)andformalizedbyKatzandMurphy(1992),assumesthathigh-andlow-skilledlabor,eachusingadistincttypeoftechnology,areimperfectsubstitutesintheproductionfunction.
Askill-biasedtechnologicalchangeisatechnologyshiftthatraisestheproductivityofhigh-skilledlaborrelativetolow-skilledlabor,increasingtherelativedemandandwagesforhigh-skilledlaborandthusraisingtheeducationpremium.
Theskill-biasedtechnologicalchangehypothesistakesthesupplyofskilledandunskilledlaborasgiven.
Autor(2014)showedthattheshareofUSlaborsuppliedbycollege-educatedworkershasbeengrowingsteadily.
Ifotherfactorsdidnotchange,thisgrowthwouldhavereducedtheeducation-1001020304050607080102030405060708090Change(percent)PercentileofwagedistributionMenWomenWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES7premium.
However,Autorfoundthatthedemandforskilledlaborgrewevenfasterthanthesupply,raisingtheeducationpremiumovertime.
Becausethishypothesisassumesthatskilledandunskilledlaborarepartialcomplements,itpredictsthatskill-biasedtechnologicalchangeraisesproductivityandwagesforunskilledandskilledworkersalike.
Unskilledwagesriseevenifonlythetechnologyusedbyskilledlaborimproves.
However,skill-biasedtechnologicalchangedoesnotraisewagesasmuchforunskilledlaborasforskilledlabor,thusincreasingtheeducationpremium.
Acompetingexplanationfortheobservedincreaseintheeducationpremiumfocusesongrowthininternationaltrade.
Themanufacturingindustry,whichemployslow-skilledlaborrelativelyintensively,hasfacedgrowingcompetitionfromdevelopingcountries.
Standardtradetheorypredictsthatincreasedcompetitionfromimportsreducespricesformanufacturedgoods,loweringrelativewagesforlow-skilledlaborwhileincreasingwagesforhigh-skilledlaborbecauseoftherelativedominanceoflow-skilledlaborinmanufacturing(StolperandSamuelson1941).
Tradetheoryandtheskill-biasedtechnologicalchangetheorypredictsimilarimpactsontheeducationpremium,exceptthattheinternationaltradeexplanationimpliesthatwagesforlow-skilledworkersfall.
Thefirstwaveofresearchontheeffectsofinternationaltradeonwageinequality,inthe1990s,concludedthattheimpactwasmodest.
Leamer(1996)analyzedpricesofmanufacturinggoodsandfailedtofindasignificantdecreaseduringthe1980swhenwageinequalitywasrising.
Borjas,Freeman,andKatz(1997)estimatedthattradeaccountedforabout20percentoftheincreaseintheUSeducationpremiumbetween1980and1995.
Morerecenttradeliteraturefocusesonothermechanismsthroughwhichtradeaffectswages,includingintra-industryandlocalityeffects.
Helpman(2016)concludedthatthereisnoevidencethattradeistheprimarydriveroftheincreaseinwageinequality.
Trade-relatedwageimpactsappeartobestronginindustriesaffectedbytradeandinareasinwhichtheseindustriesareconcentrated(Autor,Dorn,andHanson2013;HakobyanandMcLaren2016).
Unionsalsoaffectthewagedistribution,andtheirpowerhasfallendramaticallysincethe1980s.
Becauseunionsmostlyprotectlow-skilledworkers,deunionizationtendstoreducetheirwages.
Unionizationhasdeclinedsteadilyoverthepastseveraldecades.
Between1973and2017,theshareoftheUSworkforcebelongingtolaborunionsfellfrom24to11percent(HirschandMacpherson2018).
Card(1996)estimatedthatthedeclineinunionizationduringthe1980saccountedforabout20percentoftheincreaseinmen'swageinequality.
8WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESImpactofMinimumWageIncreaseRaisingthefederalminimumwagecanincreaseworkers'wages,reducetheirchancesofbeingemployed,andreducetheirworkhours.
Thelikelyimpactvariesacrossthewagedistribution.
Wagesthatwerebelowtheoldminimumrisewhentheminimumincreasesbutgenerallyremainbelowthenewminimum.
Wagesforworkersearningasubminimumwagegenerallydonotrisetothenewminimumbecausetheminimumwagelawisnotperfectlyenforced,andsomeworkersarenotcoveredbyminimumwagelaws.
Neumark,Schweitzer,andWascher(2004)examinedtheimpactofchangesinstateandfederalminimumwagelawsfrom1979to1997andestimatedtheminimumwageelasticity,definedasthepercentchangeinaworker'swagedividedbythepercentchangeintheminimumwage.
Theyconcludedthat,inthefirstyearaftertheminimumwageincrease,theminimumwageelasticityforwagesbelowtheoldminimumexceeded1,meaningthatthepercentincreaseinthewagepaidexceededthepercentincreaseintheminimumwage.
Forwagesbetweentheoldandnewminimum,theyestimatedthattheelasticitywaslessthan1andfellaswagesrose.
Theyalsoestimatedanelasticityof0.
8forwagesclosetotheminimumwageandanelasticityof0.
26forwages30to50percenthigherthantheoldminimumwage.
Inadditiontodirectlyaffectingworkerswhosewagesarebelowthenewminimum,raisingtheminimumwagecanincreasewagesforsomeworkerspaidmorethanthenewminimumbecauseofspillovereffectstiedtoemployers'preferencetopreserveexistingwagedifferentials.
CBO(2014)estimatedthatthesespillovereffectsreachashighasthenewminimumwageplus50percentofthechangeintheminimumwage.
Neumarkandcolleagues(2004)estimatedthattheminimumwageelasticitywas0.
16forwagesbetweenoneandone-halfandtwotimestheoldminimum.
Raisingtheminimumwagecanalsoaffectemployment.
Tooffsettheincreaseinlaborcoststhatwouldotherwiseresultfromaminimumwageboost,employerscanchoosetoreducethenumberofworkerstheyemployorthenumberofhourseachemployeeworks.
Intheshortrun,thisresponsewouldlikelyreduceproduction.
Inthelongerrun,manyemployersmayreplacelow-wageworkersaffectedbytheminimumwagewithhigher-skilledworkersorcapital.
Severalstudieshaveinvestigatedhowemployersgenerallyrespondtoanincreaseintheminimumwage,butnoconsensushasemerged.
Estimatesoftheminimumwageelasticityofemployment—thepercentchangeinemploymentdividedbythepercentchangeintheminimumwage—rangefromnearly-1toasmallpositivenumber,withmoststudiesfindingelasticitiesofemploymentnear0.
2Thesestudiesexaminedrelativelysmallminimumwagehikesandthusmaynotoffermuchguidanceonhowwagesandemploymentwouldrespondtolargerincreases,suchastheonewesimulatedinthisstudy.
ArecentincreaseinSeattle'sminimumwage,whichrosefrom$9.
47to$13overWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES9ayear,iscomparabletoonewesimulated,buttwostudiesofitseffectsreacheddifferentconclusions.
Reich,Allegretto,andGodoey(2017)foundeffectsfortheSeattleminimumwagehikesimilartopreviousstudies,estimatingawageelasticitybetween0.
1and0.
2andanemploymentelasticitythatwasstatisticallyindistinguishablefrom0.
Ontheotherhand,Jardimandcolleagues(2017)estimatedasmallwageelasticitybutanemploymentelasticityofabout-3.
Employerscanalsorespondtoaminimumwageincreasebyraisingpricesofgoodsandservicestocompensateforhigherlaborcosts.
Whilethereissomeevidencethatfirmspassanincreaseinthecostoflaborontoconsumersinthisway,thiseffectissmalleveninindustrieswithahighshareoflow-wageworkers,suchasthefast-foodindustry(Aaronson2001;CardandKrueger1994).
Theestimatedeffectismuchsmallerintheretailindustry(GanapatiandWeaver2017),wheretheshareoflow-wageworkersapproximatestheeconomyaverage.
Pricesofgoodsandservicesmightalsorisebecausethedemandforgoodsandservicesrisesastheminimumwageboostincreasesfamilyincomes.
Ifanincreaseintheminimumwageraisesfamilyincome,mostlylow-incomefamiliesgain.
Becausethesefamilieshaveahigherpropensitytoconsumethanhigh-incomefamilies,theoveralldemandforconsumptionwouldrise,potentiallyboostinglabordemandandpartiallyoffsettingthenegativeimpactoftheminimumwageincrease.
Employersmayalsochoosetoabsorbtheincreaseinlaborcostsbytakingcutsintheirprofits.
CBO(2014)foundthatraisingtheminimumwagefrom$7.
25to$10.
10—a39percentincrease—wouldnotchangetotalfamilyincomemuch,butitwouldredistributeincomefromhigh-incomefamiliestolow-andmoderate-incomefamilies.
Familieswithincomesbelowsixtimesthefederalpovertylevelwouldexperienceanetgainof$19billionayear,whilefamilieswithhigherincomeswouldexperienceanetlossof$17billionayear,foranoverallincreaseof$2billionayear,lessthan0.
02percentoftotal2016laborcompensation.
Suchasmallchangeintotalfamilyincomeisunlikelytoaffectthelabormarketortheeconomymuch.
MethodsToassesstheimpactofgrowingwageinequalityandapotentialincreaseinthefederalminimumwageonfutureretirementincomes,wesimulatedincomeatages67to75overthenextsevendecadesunderthreescenarios,summarizedintable1.
Baselinescenario1maintainedtheexistingunderlyinglevelofwageinequality(asmeasuredbytheeducationpremium)andthecurrentfederalminimumwageof$7.
25throughoutthesimulationperiod;scenario2raisedwageinequalityandkeptthefederalminimumwageatitscurrentlevel;andscenario3raisedwageinequalityandincreasedtheminimum10WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESwageto$12perhourin2017,adjustingforinflationinsubsequentyears.
Eachsimulationheldstateminimumwagesattheirnominal2016levelthroughoutthesimulationperiod.
Wefocusedonpeoplewhowereatleastage67becausetheyhadreachedSocialSecurity'sfullretirementageandthuswerelikelytoberetiredandnotworking.
Weexcludedpeopleolderthan75,whenmortalityratesbeginrisingsharply.
TABLE1SimulationScenariosScenarioWageinequalityMinimumwage1.
BaselineFixedatcurrentlevel$7.
25(currentnominallevel)2.
RisinginequalityIncreasing$7.
25(nominal)3.
PolicyinterventionIncreasing$12.
00(adjustwithinflation)Ourprojectionsshowhowpotentialchangesinwageinequalityandthefederalminimumwagemightaffectfutureearningsand,inturn,retirementincome.
Toaccountforthetendencyofmarriedcouplestopooltheirearnings,wecomputedsharedlifetimeearningsbysummingoveralifetimeowninflation-adjustedearningsforeachyearthatapersonwassingleandone-halfofowninflation-adjustedearningsplusone-halfofaspouse'sinflation-adjustedearningsforeachyearthatapersonwasmarried.
Wemeasuredincomeinretirementasnetpercapitaannuityincome.
ItincludedSocialSecuritybenefits,incomefromemployer-sponsoredDBpensionplans,earnings,SupplementalSecurityincome(SSI),incomefromotherfederalassistanceprograms,andimputedrent,plusannuitizedassetincome.
3Thisincomewascalculatedastheannualactuariallyfairannuitypaymentafamilywouldreceiveifitannuitized80percentofitsretirementaccountsandotherfinancialassetsusinga3percentannualrealreturn.
4Excludingannuitizedassetincomewouldunderstatethefinancialsecurityprovidedbydefinedcontributionretirementplans.
Toobtainavaluenetoftaxesandpremiums,wesubtractedfromtotalincomefederalandstateincometaxes;payrolltaxesthatfundtheoldage,survivor,anddisabilityinsurance(OASDI)programandMedicare'shospitalinsurance(HI)program;theMedicaresurtaxonearningsandinvestmentincomeforhigh-incomepeople;andpremiumsforMedicarePartsBandD.
Weobtainedapercapitameasurebydividingmarriedpeople'sfamilyincomebytwo.
Theanalysisshowedhowthefuturedistributionofsharedlifetimeearningsandretirementincomewoulddifferundereachscenario.
WecomparedmeanearningsandretirementincomebyquintileofsharedlifetimeearningsandcomputedGinicoefficientsofretirementincomefrom2015to2087.
Tounderstandwhichportionoftheretirementincomedistributionshiftedinresponsetorisingwageinequalityandanincreaseinthefederalminimumwage,wecomputedratiosofthe90thto10thpercentilesofretirementincome,the90thto50thpercentilesofretirementincome,andthe50thtoWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES1110thpercentilesofretirementincome.
Wealsodecomposedchangesinretirementincomebyincomesourcetounderstandbetterhowchangesinlifetimeearningsfilterthroughtoretirementincome.
ChangesintheearningsdistributionwillaffectthesolvencyoftheOASDItrustfund,butweignoredthiseffectandassumedthatthetrustfundwouldpaybenefitsasscheduledundercurrentlaw.
Anincreaseinearningsinequalitygenerallyworsensthefund'ssolvencybecauseitreducestaxablepayroll—coveredearningsbelowtheSocialSecuritytaxablemaximum—andthusthetrustfund'srevenue.
AlthoughthereductioninSocialSecuritybenefitsassociatedwithrisingwageinequalitywouldpartiallyoffsetthisrevenueloss,benefitreductionswouldoccurmuchlaterthanrevenuelosses,andrevenuelosseswouldexceedbenefitcutsbecauseoftheprogressivityoftheSocialSecuritybenefitformula.
DYNASIM4WeprojectedretirementincomeundereachscenariousingDYNASIM4,adynamicmicrosimulationmodeldesignedtoanalyzethelong-rundistributionalconsequencesofretirementandagingissues.
Themodelstartswitharepresentativesampleofindividualsandfamiliesfromthe2004and2008SurveyofIncomeandProgramParticipationandagesthemyearbyyear,simulatingkeydemographic,economic,andhealthevents.
DYNASIM4projectsthat,eachyear,somepeopleinthesamplegetmarried,haveachild,orfindajob,andotherpeoplebecomedivorcedorwidowed,stopworking,begincollectingSocialSecurity,becomedisabled,ordie.
Thesetransitionsarebasedonprobabilitiesgeneratedbycarefullycalibratedequationsestimatedfromnationallyrepresentativehouseholdsurveydata.
Theequationsaccountforimportantdifferencesbysex,education,earnings,andothercharacteristicsinthelikelihoodofvariousexperiences.
OtherequationsinDYNASIM4projectannualearnings,savings,andhomevalues.
Themodelusesprogramrules—combinedwithprojectionsoflifetimeearnings,disabilitystatus,andhouseholdincomeandwealth—toprojectSocialSecurityretirementanddisabilitybenefitsandMedicaidcoverage.
ForconsistencywithSocialSecurity'sprojectionsaboutsystemfinances,wegenerallyfollowtheSocialSecurityandMedicaretrustees'assumptions.
FormoreinformationaboutDYNASIM,seeUrbanInstitute(2015)andFavreault,Smith,andJohnson(2015).
Projectedearningsdependonhowfastthesimulationsassumeaveragewagesgrow.
Between1966and2015,realwagegrowthaveraged0.
81percentperyear,butrealwagesgrewmuchmoreslowlyduringcertaindecadesandmuchmorerapidlyinothers(BoardofTrustees2016).
Annualrealwagegrowthaveraged-0.
01percentbetween1973and1979,0.
47percentbetween1979and1989,and0.
45percentbetween2007and2015.
Between1989and2000,bycontrast,realwagesgrew1.
5512WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESpercentperyear.
StandardDYNASIM4projectionsusetheintermediatewagegrowthassumptionsadoptedbytheSocialSecuritytrustees,whichassumethatrealaveragewageswillgrow1.
2percentperyearfrom2026to2091(BoardofTrustees2016).
Althoughtheseassumedgrowthratesarehigherthansomeacademic-styleprojections(Fernald2016;Gordon2014),theyareroughlyconsistentwithCBO(2016)'sassumptions.
Ourbaselinesimulationsgeneratedunderscenario1assumedthatrealwagesgrowatthesamerateoverthesimulationperiodforalleducationgroups,maintainingtheexistinglevelofwageinequality.
Underscenarios2and3,whichsimulatedearningswhenwageinequalitycontinuedtorise,projectedearningsgrewmorerapidlyforcollegegraduatesthanforworkerswithlesseducation.
DYNASIM4simulatesannualearningseveryyearforpeopleages16to80astheproductofthehourlywage,employmentstatus,andtheannualnumberofhoursworked,eachofwhichwassimulatedindependently.
Thehourlywageequationisbasedontheeducationpremiummodel,firstproposedandestimatedbyKatzandMurphy(1992).
WeusedearningsequationparametersfromAutor(2014),whoreestimatedtheeducationpremiummodelusingmorerecentdata.
EstimatingtheEducationPremiumModelTheeducationpremiummodelcaststhewagebonusassociatedwithacollegediplomaintermsofrelativelaborsupplyanddemand.
Bothrelativesupplyanddemandaredefinedasratiosoflaborunitsassociatedwithcollege-educatedworkersrelativetothoseassociatedwithhigh-school-educatedworkers.
Theeducationpremiumincreasesasrelativedemandrisesandrelativesupplyfalls.
Forexample,anincreaseincollegeattendanceandgraduationrateswouldraisetheamountoflaborunitssuppliedbycollege-educatedworkers,depressingtheirwagesrelativetowagesforworkerswithonlyahighschooleducationwhenotherfactorsareheldconstant.
Conversely,adoptinganewtechnologythatmakescollege-educatedworkersmoreproductivewouldincreaserelativedemandfortheseworkersandthusraisetheeducationpremium.
Theeducationpremium,,isthelogarithmoftheratioofcomposition-adjustedmeanweeklywagesforcollege-educatedworkerstowagesforhigh-school-educatedworkersandcanbemodeledas:=0+1+2+32+(1)whereSctandShtaresharesofefficiencyunits—laborhoursweightedbythehourlywage—suppliedbycollegeandhighschoolgraduates,respectively,inyeart;thelinearandquadratictimetermsproxyforWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES13relativelabordemand;andistheerrorterm.
Autorestimatedthisequationoncross-sectionalCPSdatafrom1964to2013.
Figure3showshisestimatesoftheeducationpremiumfrom1963to2012,andfigure4showshisestimatesofthelaborsupplyofcollegegraduatestohighschoolgraduates.
FIGURE3WageGapAssociatedwithHavingatLeastaCollegeDegreeURBANINSTITUTESource:Authors'calculationsfromAutor(2014).
Notes:Relativesupplywasmeasuredbyefficiencyunits.
Hoursoflaborsupplywereweightedbythewage.
Therelativelaborsupplywascalculatedasthelogarithmoftheratiooftheshareofefficiencyunitsprovidedbycollege-educatedworkerstotheshareofefficiencyunitsprovidedbyworkerswithnomorethanahighschooldiploma.
Theanalysiscountedone-halfofworkerswithsomecollegeascollege-educatedlaborandtheotherhalfashigh-school-educatedlabor.
40506070809010011019701980199020002010Collegepremium(percent)Year14WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE4LaborSupplyofCollegeGraduatesRelativetoHighSchoolGraduatesURBANINSTITUTESource:Authors'calculationsfromAutor(2014).
Notes:Relativesupplywasmeasuredbyefficiencyunits.
Hoursoflaborsupplywereweightedbythewage.
Therelativelaborsupplywascalculatedasthelogarithmoftheratiooftheshareofefficiencyunitsprovidedbycollege-educatedworkerstotheshareofefficiencyunitsprovidedbyworkerswithnomorethanahighschooldiploma.
Theanalysiscountedone-halfofworkerswithsomecollegeascollege-educatedlaborandtheotherhalfashigh-school-educatedlabor.
Table2reportsparameterestimatesforequation(1),theeducationpremium,fromAutor(2014).
Asexpected,theeducationpremiumwasnegativelycorrelatedwiththerelativesupplyofcollege-educatedlabor.
Thecoefficientsforthetimetrendindicatethattherelativedemandforskilledlaborincreasesovertime,butatadiminishingrate.
WeusedtheseparameterestimatestosimulatetheeducationpremiuminDYNASIM4through2087(figure5).
5ThepredictionsoftheeducationpremiumgeneratedbyAutor's(2014)wageequationalignedcloselywiththeeducationpremiumobservedinCPSdatafrom1963to2013,andtheeducationpremiumsimulatedbyDYNASIM4followedthesametrajectoryasthehistoricalone.
-1-0.
500.
519701980199020002010RelativelaborsupplyYearWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES15TABLE2EstimatedParametersfortheEducationPremiumEquationCoefficientTime0.
0306***(0.
00468)Timesquared/100-0.
0144***(0.
00421)Relativelaborsupply-0.
588***(0.
0922)Constant-0.
213**(0.
0972)Observations50AdjustedR-squared0.
947Source:Autor(2014).
FIGURE5WagePremiumAssociatedwithHavingatLeastaFour-YearCollegeDegreeActual,predicted,andsimulatedvaluesURBANINSTITUTESources:Autor(2014)andDYNASIM4ID942B.
020406080100120140160180197019801990200020102020203020402050206020702080Educationpremium(%)YearPredictedActualSimulated16WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESEstimatingtheWageModelAwell-knownchallengewhenestimatingawageequationissampleselectionbias.
Thedependentvariableistruncatedbecausewagescannotbeobservedfornonworkers,andselectionintotheobservedsampleisnonrandom,biasingordinaryleastsquaresestimatesoftheregressioncoefficients(Heckman1976).
Thesampleselectionbiasarisesfromthecorrelationbetweentheresidualandregressorsintheequation,bothofwhicharecorrelatedwiththesampleselectionprocess.
Togenerateunbiasedestimatesofourregressioncoefficients,weappliedHeckman's(1979)two-stageestimationprocedure.
Thefirst-stageemploymentequationestimatedanewvariable,theso-calledinverseMillsratio,whichenteredthesecond-stagewageequationasaregressor.
Equations(2)and(3)describeourtwoestimatingequations:=+(2)=++++(3)Ourfirst-stageestimationmodel(2)wasaprobitinwhichthedependentvariablewasanindicatorvariablethatequaled1ifpersonwasemployedintheyearofthesurvey(0otherwise),θwasavectorofcoefficients,Vitwasavectorofindividualcharacteristics,andμitwastheerrorterm.
Inthesecondstage,weusedanordinaryleastsquaresregressionoftherealhourlywageonwt,theaveragerealhourlywageinyeart;,thevectorofindividualcharacteristics;Zt,theeducationpremiuminyeart,whichwasmultipliedbyanindicatorvariableforafour-yearcollegedegree;and,theinverseMillsratio.
,.
wastheerrorterm.
6TheindividualcharacteristicsincludedinbothandVitwereindicatorvariablesformaritalstatus,Hispanicethnicity,threegeographicregions(South,Northeast,andWest,withMidwestasthereferencegroup),andfoureducationalattainmentlevels(highschoolgraduate,somecollegebutlessthanfouryears,four-yearcollegedegree,andgraduateeducation,withnotahighschoolgraduateasthereferencegroup).
Inaddition,weincludedagesplines(withkinksatages25,30,40,50,and60),andtheirinteractionswithboththecollegeandhighschoolgraduateindicators.
VectorVitalsoincludedindicatorvariablesforforeignorigin;healthstatus;andreceiptofSocialSecurityincome;andnonbinaryvariablesequaltothenumberofchildren(totalandthoseyoungerthanage5);spouse'slogrealincome;andafifth-degreepolynomialoftime,whichweexcludedfromvector.
Vectorincluded16five-yearbirthcohorts(with1901–05asthereferencecategory).
Theregressioncoefficientswereestimatedonapooledcross-sectionalsampleofthecivilianpopulationages17to81fromthe1976to2012releasesoftheCPS.
7Weexcludedself-employedworkersandthoseemployedintheprivatehouseholdsector.
OurdependentvariablewastheWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES17logarithmoftherealhourlywage,calculatedbydividingreportedannualearningsbytheproductofthenumberofweeksworkedintheprioryearandtheusualnumberofhoursworkedperweek,adjustingforthechangeintheconsumerpriceindex.
Becausethesevariablesreferredtotheyearpriortothesurvey,theeffectiveagesofsampleindividualsrangedfrom16to80.
Weestimatedequation(2)ontheentiresample,stratifiedbyraceandsex.
Appendixtable1reportsparameterestimates.
WeusedtheestimatedcoefficientstocalculatetheinverseMillsratio,=(θVit)/Φ(θVit),whereisthestandardnormaldensityandΦisthestandardnormalcumulativedistributionfunction.
Thisvariablewasincludedasaregressorinequation(3).
Theaverageannualhourlywage,calculatedasaweightedmeanofhourlywagesforallworkersintheCPS,stratifiedbysexandrace,increasedtheregression'sadjustedR2andimprovedthefitofpredictedwagesovertheentiresampleperiod,inthatourprojectedwagedistributionsfromthebeginningoftheprojectionperiod,from2006throughthemid-2010s,werequitesimilartotheobservedhistoricaldistributions.
Improvedpredictivepowerisanimportantconsiderationwhengeneratingprojections.
Appendixtable2reportsparameterestimates.
Tosimulatefuturehourlywages,wefirststatisticallymatchedrespondentsfromthePanelStudyofIncomeDynamicstoDYNASIM4'sSurveyofIncomeandProgramParticipationinputfile.
Wethenusedequation(3)topredictpastwages,computedthedifferencebetweenpredictedandactualwages,andusedafixedeffectsmodeltoregressthepredictederrortermonitslaggedvaluetodecomposetheerrorintoanindividual-specificpermanenttermandatransitoryterm.
Basedonthoseregressionresults,reportedinappendixtable3,andourestimatesofα,β,andfrom(3),weusedequation(4)tosimulatefuturehourlywages:=α+γCitZt+′++(4)whereisanindividual-specificpermanenterrortermthatrepresentsunobservableindividualcharacteristics,isatransitorydisturbancemodeledasafirstorderautoregressiveprocess,=,1+,,andtheotherparametersareasdescribedin(3).
Theautoregressivecoefficientinthetransitoryerrormodelwasestimatedasthecoefficientonthelaggederrorterminappendixtable3,andthestandarddeviationofthetransitoryerrorreportedintheappendixtablewasusedtogeneratetherandomcontemporaneousterm.
TosimulatehourlywagesinDYNASIM4foranindividualinaparticularyear,weretrievedtheappropriatevectoroftheindividual'scharacteristics,thepermanenterrorterm,andthepreviousyear'stransitoryerrortermfromtheindividual'srecord.
Thecontemporaneoustransitoryerrortermwascalculatedfromarandomlygeneratednumberandlastyear'stransitoryerror.
Thevectorofindividualcharacteristicswasmultipliedbythevectorofestimatedcoefficients.
Thisvalueandthepermanentand18WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMEStransitoryerrortermswereaddedtogethertoobtainthesimulatedvalue.
Inscenario1,wesettheeducationpremiuminequation(4)equaltoits2006valuethroughoutthesimulationperiod.
Inscenarios2and3,theeducationpremiumvariedwithtimeandtherelativesupplyofcollege-educationlabor.
DYNASIM'sEmploymentandHoursModelToprojectannualearnings,DYNASIM4combinessimulatedhourlyearningswithsimulatedemploymentstatusandsimulatedannualhoursofemployment.
Thesimulationequationsforlit,laboremploymentstatus,andhit,hoursofwork,resembletheearningsequationandaredepictedbelow:=′12+1+1(5)=X′23+2+2(6)karevectorsofobservableindividualcharacteristics,karevectorsofassociatedcoefficients,kareindividual-specificpermanenterrortermsthatrepresentunobservableindividualcharacteristics,andaretransitorydisturbancesthataremodeledasfirstorderautoregressiveprocesses,,=,,1+,,,with=12.
Aswiththeearningsequations,theseequationswereestimatedandsimulatedseparatelyfornon-blackwomen,blackwomen,non-blackmen,andblackmen.
Theemploymentequationcontrolledforage,education,age-educationinteractions,schoolenrollment,disability,healthstatus,maritalstatus,immigrantstatus,numberofminorchildren,spouse'slaborincome,thestateunemploymentrate,SocialSecuritybeneficiarystatus,region,five-yearbirthcohort,andindicatorsforheadofhouseholdorthehead'sspouse.
Thehoursequationincludedthehourlywage,age,education,age-educationinteractions,schoolenrollment,disability,divorcedstatus,numberofminorchildren,spouse'slaborincome,thestateunemploymentrate,region,andfive-yearbirthcohort.
Theemploymentstatusequationwasestimatedasarandomeffectsprobit,andthehoursworkedequationwasestimatedasatobit,usingpaneldatafromtheNationalSurveyofYouth1979and1997forpeopleages25andyounger,andthePanelStudyofIncomeDynamicsfrom1981to2011forpeopleages26andolder.
Weusedtheseequationstoprojectemploymentstatusandannualhours,followingthetechniquesweusedtoprojectthehourlywage.
WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES19SimulatingtheMinimumWageScenario3simulatedearningsundertheassumptionthatthefederalminimumwageincreasedin2017from$7.
25perhourto$12,a66percentjump,andsubsequentlyadjustedwithinflation.
Asnotedearlier,thisincreasewouldaffectwagesandemploymentforcoveredworkers,dependingonhowmuchaworkerearned.
Thefirststepwastoidentifyworkersexemptfromthefederalminimumwage.
9Becausenonationallyrepresentativedataset,toourknowledge,asksrespondentswhethertheyarecoveredbyminimumwagelegislation,wehadtoinferexemptstatus.
Weclassifiedwageandsalaryworkersasexemptfromfederalandstateminimumwagelawsiftheywerepaidhourlyandreceivedawagethatwaslessthantheeffectiveminimuminthestatewheretheylived.
Weassumedthattheeffectiveminimumwagewasthehigherofthestateandfederalminimumwageandthatthefederalminimum-wageexemptionrulesappliedforallstateminimums.
Ashortcomingofthisapproachisthatitleavesoutexemptworkerswhowerepaidanhourlyratethatexceededtheminimumwage.
ToidentifyexemptworkersinDYNASIM4simulations,weestimatedalogisticmodelofthelikelihoodthataworkerwasexempt,inwhichthedependentvariablewasanindicatorthatequaled1ifaworkerwasexempt,and0otherwise.
Ourindependentvariableswerefourdummyvariablesforeducationalattainment(highschooldiploma,somecollege,four-yearcollegediploma,andagraduateschooldiploma);adummyformaritalstatus;adummyforHispanicethnicity;sixagesplines(at16,19,25,50,60,and65)andtheirinteractionswithdummiesforhighschooldiplomaandcollegediploma(forcollegegraduates,interactionswiththe16-and19-yearsplinesarereplacedbyaninteractionwiththe22spline);16dummiesforbirthcohorts(five-yearcohortsbetween1906and1996);and50dummiesforstatesofresidence.
Weestimatedthemodelseparatelyforblackandnon-blackwomen,andblackandnon-blackmen.
WeestimatedthemodelontheCPSMergedOutgoingRotationGroup(CPS-MORG)dataset,whichprovidesmorereliableinformationonthehourlywagethanmostotherdatasets.
CPS-MORGaskedemployedrespondentswhowerepaidhourlyabouttheirhourlywage,whereastheCPS'sAnnualSocialandEconomicSupplementaskedrespondentsonlyabouttheirannualearnings.
Forotherworkers,wecalculatedthehourlywagebydividingtheweeklywagebytheusualnumberofhoursworkedperweek.
Appendixtable4reportsestimatedparameters.
WecalibratedtheshareofworkersinDYNASIM4assumedtobeexemptfromtheminimumwagetotheshareofworkersinCPS-MORGwhoearnedlessthanthefederalminimumwagein2016,whichwas2.
4percent.
ForeachworkerinDYNASIM4,weindependentlysimulatedthehourlywageandthelikelihoodthataworkerwasexemptfromthe20WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESminimumwagelaw.
Wethenincreasedthesimulatedwageforcoveredworkersiftheirinitialsimulatedwagefellbelowtheeffectiveminimumwage.
Whentheminimumwageincreasesfrom$7.
25to$12perhour,weassumedthatallwagesbelowthenewminimumwouldrise.
Additionally,basedonCBO(2014),weassumedthatspillovereffectswouldaffectwagesonlyupto$14.
37,thenewminimumwageplusone-halftheincreaseintheminimumwage.
Weconstructedafunctionthatmapspreincreasewagesintopostincreasewagesconsistentwithestimatesintheliteratureofwageelasticitieswithrespecttotheminimumwage,asreportedintable3forselectedwages.
10Weassumedthatinitialwagesforcoveredworkersthatfellbelowtheoldminimumincreasedbythefulldollarchangeintheminimum,whichimpliesanelasticitygreaterthan1.
Wagesthatfellbetweentheoldminimumwageandtheupperboundonthespillovereffectwouldincreasebyanamountthatfallslinearlyasthewagerisesuntilitreaches0attheupperbound.
Theminimumwageincreasewouldnotchangehighersimulatedwages.
TABLE3AssumptionsabouttheWageImpactofIncreasingtheMinimumWagefrom$7.
25to$12Source:Authors'calculations.
Raisingtheminimumwagecanalsoaffectemployment.
WeusedestimatesfromCBO(2014)tosimulatetheimpactofaminimumwageincreaseonemploymentratesandestimatesfromNeumarkandcolleagues(2004)tosimulatetheimpactonworkhours(table4).
Theminimumwageelasticitiesinthetablerepresentthetotaleffectofaminimumwageincreaseonemploymentforallworkers.
However,becausetheincreaseaffectsonlyafractionofworkersandmanyofthemreceiveawageincreasethatissmallerthantheincreaseintheminimumwagethatcausedit,elasticitieswithrespecttotheminimumwagemustbeconvertedtoelasticitieswithrespecttothewage.
Todothat,weusedOldwageNewwageChangeElasticity5.
009.
754.
751.
4506.
0010.
754.
751.
2087.
0011.
754.
751.
0367.
2512.
004.
751.
0008.
0012.
254.
250.
8119.
0012.
583.
580.
60810.
0012.
922.
920.
44511.
0013.
252.
250.
31212.
0013.
581.
580.
20113.
0013.
920.
920.
10814.
0014.
250.
250.
02714.
3814.
380.
000.
000WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES21theconversionmultiplierof4.
5estimatedbyCBO(2014)usingCPSdatafrom1979to2009.
Forexample,ifanadultworkerwhowaspaid$10perhourreceivedanincreaseof$2—or20percent—whentheminimumwageincreased,thelikelihoodofemploymentforthisworkerwoulddecreaseby0.
2x4.
5x0.
025=0.
0225or2.
25percent,and,iftheworkerremainedemployed,heorshewouldwork0.
2x4.
5x0.
3=0.
27or27percentfewerhours.
TABLE4AssumptionsabouttheMinimumWageElasticitiesofEmploymentandHoursWorked2017inflation-adjusteddollarsWagerangeElasticityofemployment(teenagers)Elasticityofemployment(adults)ElasticityofhoursworkedLessthanorequalto$12.
00-0.
075-0.
025-0.
3Morethan$12.
000.
00.
00.
0Sources:CBO(2014);Neumarketal.
(2004).
Thesimulatedminimumwagehikein2017initiallyreducedemploymentlevels,especiallyforadultswithlimitededucation,buttheeffectsdissipatedovertime(figure6).
Comparedwithbaselinescenario1,meanannualhoursworkedin2017foradultsages25to69inscenario3was9percentlowerforthosewhodidnotcompletehighschool,7percentlowerforthosewithahighschooldiplomawhodidnotattendcollege,6percentlowerforthosewhoattendedcollegebutdidnotobtainafour-yeardegree,and3percentlowerforthosewithatleastafour-yearcollegedegree.
Theseobserveddeclinesreflectbothadropintheshareofadultsemployedandinhoursworkedforthoseemployed.
Employmentquicklyreturnedtothebaselinelevelforcollegegraduatesandremainedabitbelowthebaselineforothers.
22WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE6TheEffectoftheIncreaseinMinimumWageonEmploymentbyEducationURBANINSTITUTESource:Authors'estimatesfromDYNASIMID942andID942C.
Note:Thefigureshowsthepercentagedifferenceinmeanannualhoursworkedbetweenscenario3andthebaselinescenario1foradultsages25to69.
Theminimumwagehikeboosted2017earningsforadultswithlimitededucation,morethanoffsettingtheemploymentdrop(figure7).
In2017,annualearningsforadultsages25to69whodidnotcompletehighschoolwere4.
6percenthigherthanin2016underscenario3,butonly2.
7percenthigherunderbaselinescenario1,whichkepttheminimumwageatitscurrentlevel.
Averageearningsalsoincreasedmoreundertheminimumwagehikescenariothanthebaselineforhighschoolgraduates,andtheydidnotchangemuchforadultswhoattendedcollege.
-10%-8%-6%-4%-2%0%2%4%20102020203020402050206020702080DidnotcompletehighschoolHighschoolgraduateSomecollegeCollegegraduateWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES23FIGURE7AnnualEarningsGrowthbySimulationScenarioandEducationalAttainment,2017(%)Adultsages25to69URBANINSTITUTESource:Authors'estimatesfromDYNASIMID942andID942C.
Notes:Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andincreasedwageinequality,asdescribedinthetext.
Morethanathirdofworkerscouldbeaffectedbyraisingthefederalminimumwageto$12perhour.
Figure8showsthe2016distributionofhourlywagesforcivilianworkersages16to64.
Theshadedareasrepresentwagesthatwouldriseiftheminimumwagewereincreasedfrom$7.
25to$12.
Thebluearearepresentsthe2.
4percentofworkerswhowerepaidlessthan$7.
25perhour.
Thegrayarearepresentsthe25.
5percentofworkerspaidbetween$7.
25and$12perhour,andtheyellowarearepresentsthe9percentofworkersearningbetween$12and$14.
38perhourwhowouldbeaffectedbyspillovereffects.
012345DidnotcompletehighschoolHighschoolgraduateSomecollegeCollegegraduateScenario1(baseline)Scenario324WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE8DistributionofHourlyWagesin2016URBANINSTITUTESource:Authors'estimatesfromCPS-MORG2016.
Notes:Thehorizontalaxisrepresentshourlywagein2016dollars,andtheverticalaxisrepresentsprobabilitydensity.
Thesamplewasrestrictedtoworkersages16to64.
Wageswereweightedbytheearningsweight.
Shadedareasrepresenttheshareofworkerswhowouldbeaffectedbyanincreaseinthefederalminimumwagefrom$7.
25to$12,directlyorthroughspillovereffects.
Oursimulationsdidnotexplicitlyincorporatepriceorprofitresponsestoanincreaseintheminimumwage,whichcouldreducerealincomesforhigh-earningfamilies.
Nonetheless,ourmethodsledtolowerrealwagesforhighearnerswhentheminimumwageincreased.
DYNASIM4'searningssimulationswerecalibratedtotheSocialSecuritytrustees'projectionsoffutureannualearnings.
Sinceweusedthesameprojectionsinallthreescenarios,theminimumwageincreasedidnotchangetotaleconomy-wideearnings.
Whenwagesincreasedforlow-wageworkers,highearnersreceivedapaycutsothattotalearningspaiddidnotchange.
Thisprocessemulatedthedecreaseinrealwagesthatwouldoccurwhenpricelevelsriseinresponsetoaminimumwageincrease.
Becausethelow-wageworkerswhosewagesincreasewiththeminimumwagemakeupasmallshareoftheworkforce,thepaycutdistributedamonghigher-wageworkerswasrelativelysmall.
00.
010.
020.
030.
040.
050.
060.
07Wage<=$7.
25$7.
25
Figure9showsinversecumulativedistributionfunctionsforhourlywagesin2016.
ThebluelinerepresentsthedistributionestimatedfromCPS-MORG2016data,andtheblacklinerepresentsthedistributionsimulatedbyDYNASIM4underscenario2in2016,thetenthyearofthesimulation.
Thesimulateddistributionexhibitsagoodfit,especiallyatthebottomofthewagedistributionneartheminimumwage,whichisimportantbecauseouranalysisfocusesonhowboostingtheminimumwagemighthelplow-incomeworkers.
Thetwodistributionsdivergesomewhatintheupperdecile,wherethesimulationovershootsthehistoricdata.
FIGURE9CumulativeDistributionofSimulatedandHistoricalHourlyWages,2016URBANINSTITUTESources:Authors'estimatesfromCPS-MORG(2016)andDYNASIMID942B.
Notes:FortheCPS-MORG,wesethourlywagesequaltoreportedwagesforworkerspaidhourly;otherwise,hourlywageswereestimatedasweeklywagesdividedbyhoursworkedperweek.
Distributionsweretruncatedat$100perhour.
Themostchallengingtestfortheminimumwagemodelishowitperformsovertime.
Astheminimumwagechanges—throughlegislationorbecauseinflationreducesitsrealvalue—theshareofwagesatorbelowtheminimumwagealsochanges.
Figure10showstheshareofhourlywagesthatareatorbelowthefederalminimumwageovertimeunderthethreesimulationscenariosandintheCPS-MORG.
Theblue,black,andyellowlinesrepresentvaluessimulatedbyDYNASIM4underscenarios1,-20020406080100MORGDYNASIMPercentileHourlywage($)26WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES2,and3,respectively;themagentalinerepresentsvaluesestimatedfromCPS-MORGdatafrom2007to2016.
Thethreeconsecutiveincreasesinthefederalminimumwagein2007(from$5.
15to$5.
85),2008(from$5.
85to$6.
55),and2009(from$6.
55to$7.
25)sharplyincreasedtheshareofwagesbelowtheminimumwageinthoseyears.
Theminimumwagehassinceremainedfixed,andtheshareofwagesatorbelowtheminimumfellasinflationerodeditsrealvalue.
FIGURE10PercentageofTotalWagesatorbelowtheFederalMinimumWagebySimulationScenario,2006–85URBANINSTITUTESources:Authors'estimatesfromCPS-MORG(2007–16)andDYNASIMID942,ID942B,andID942C.
Notes:Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Intheprojectionyears,theshareofwagesatorbelowtheminimumwagecontinuedtofallunderscenarios1and2,asthenominalvalueoftheminimumwageremainedfixedat$7.
25andinflationerodeditsrealvalue.
Underscenario3,however,theminimumwageincreasedto$12in2017,andtheshareofwagesbelowtheminimumwagesurgedtomorethan7percent.
A66percentincreaseintheminimumwagecauseda156percentincreaseintheshareofwagesatorbelowtheminimumwage.
01234567820102020203020402050206020702080Scenario1Scenario2Scenario3MORGYearPercentWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES27Theseeffectsarecomparabletothoseobservedinthe2000s,whentheminimumwageincreasedby41percentoverthreeyearsandraisedtheshareofwagesatorbelowtheminimumby124percent.
After2017,theminimumwagewasindexedforinflation,whichslowedbutdidnoteliminatethedeclineoftheshareofwagesatorbelowtheminimum.
ResultsThethreesimulationscenariosgeneratedthreedistincttrajectoriesoftheearningsdistribution.
Underthebaseline,theearningsratioofcollegegraduatestohighschoolgraduatesdeclinedovertime(figure11).
Eventhoughthebaselinescenarioheldtheeducationpremiuminthewagemodelconstant,otherfactors,suchasanincreaseintheshareofcollegegraduatesandchangesintheirdemographics,reducedtheearningsgapbetweencollegeandhighschoolgraduates.
Underscenario2,therisingwageinequalityscenario,theearningsratiogrewsteadily,from2.
0in2017to2.
1in2046,2.
2in2058,and2.
3in2077.
Bytheendofthesimulationperiodin2087,theearningsratiowas27percenthigherunderscenario2thanunderthebaseline.
Raisingtheminimumwagein2017reducedthepayadvantageforcollegegraduates,astheearningsratiofellto1.
9in2019.
Thehigherminimumwage,whichwastiedtopriceinflation,kepttheearningsratiolowerthanitotherwisewouldhavebeenforabouttwodecades.
However,theminimumwagebecamelesseffectiveovertime,diminishingitsimpact,becausewagesgrewfasterthanprices.
28WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESFIGURE11EarningsRatio,CollegeGraduatestoHighSchoolGraduatesbySimulationScenario,2006–88URBANINSTITUTESource:Authors'estimatesfromDYNASIMID942,ID942B,andID942C.
Notes:Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
SharedLifetimeEarningsProjectionsEachscenarioprojectedthatsharedlifetimeearningswillincreasesubstantiallyovertime,aswagesarepredictedtogrowfasterthanprices(table5).
Foradultsages67to75,DYNASIM4projectedthatunderbaselinescenario1,meaninflation-adjustedsharedlifetimeearningsinthemiddlequintileofthesharedlifetimeearningsdistributionwillgrowfrom$1.
5millionin2025to$2.
0millionin2055,a33percentincrease.
By2085,themeanvalueforthemiddlefifthofthedistributionwillreach$3.
0millionunderthebaselinescenario,50percentmorethanin2055andtwiceasmuchasin2025.
Thebaselinescenario,whichholdstheeducationpremiumconstantovertime,projectedthatlifetimeearningswillgrowatsimilarratesthroughoutthedistribution,althoughlifetimeearningslevelsdifferedsharply.
In2055,forexample,projectedmeansharedlifetimeearningsforadultsages67to75wasonlyabout00.
511.
522.
520102020203020402050206020702080EarningsratioYearBaselineScenario2Scenario3WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES29one-thirdasmuchforthebottomquintileofthedistributionasforthemiddlequintile,whilemeanlifetimeearningsweremorethantwiceashighinthetopquintileasinthemiddlequintile.
TABLE5ProjectedMeanSharedLifetimeEarningsforAdultsAges67to75byScenario,Year,andQuintileThousandsof2015inflation-adjusteddollarsYearQuintileScenario2025203520452055206520752085BottomBaseline4995346206607368691,03724995276076327018019443500536625670751870970SecondBaseline1,0791,1261,2841,4191,5611,8112,12821,0771,1191,2611,3671,4881,6921,94331,0791,1281,2761,3861,5271,7441,958ThirdBaseline1,5341,6051,7942,0212,2432,5882,95321,5321,6001,7751,9852,1782,5102,78931,5331,5991,7681,9842,1812,4852,776FourthBaseline2,0672,1782,4522,7643,0433,5403,90422,0682,1812,4742,7703,1063,5693,92432,0632,1712,4462,7583,0483,5233,896TopBaseline3,2823,8924,3004,8425,4226,2556,42123,3033,9324,3784,9775,6736,5876,90433,2923,9134,3234,9375,5836,4566,974Source:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:Sharedlifetimeearningswerecalculatedbysummingoveralifetimeownearningsforeachyearthatapersonwassingleandone-halfofownearningsplusone-halfofspouse'searningsforeachyearthatapersonwasmarried.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Asexpected,oursimulationsshowthatrisingwageinequalitywouldboostsharedlifetimeearningsinthetopportionofthedistribution,especiallyinthetopfifth,andwouldreducesharedlifetimeearningsinthebottomportionofthedistribution(figure12).
Averagewagesgrewatthesamerateundertherisingwageinequalityscenario—scenario2—andbaselinescenario1,butwagesforcollegegraduatesgrewfasterinscenario2thanthebaseline,reducingwagegrowthandlifetimeearningsforless-educatedworkersinscenario2.
Becausescenario2didnotbegintodivergefromscenario1until2006andtheeffectscumulatedovertime,projecteddifferencesinsharedlifetimeearningsforadultsages67to75didnotbecomesubstantialforseveraldecades.
Relativetothebaselinescenario,meansharedlifetimeearningsforthetopquintileunderscenario2were2percenthigherin2045,5percenthigherin2065,and8percenthigherin2085.
Inthebottomquintile,meansharedlifetimeearningsunderscenario2fellshortofthebaselinemeanby2percentin2045,5percentin2065,and9percentin2085.
Sharedlifetimeearningswerealsolowerinscenario2thaninthebaselineforthesecondand30WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESthirdquintilesofthesharedlifetimeearningsdistribution.
In2085,projectedmeansharedlifetimeearningsforthesecondquintilewere9percentlowerunderscenario2thanscenario1.
FIGURE12EffectsofRisingWageInequalityandAnIncreaseintheMinimumWageonSharedLifetimeEarningsbyQuintile,2045,2065,and2085Adultsages67to75URBANINSTITUTESource:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:Thefigureshowsthepercentagechangeinmeansharedlifetimeearningsbyquintileofsharedlifetimeearnings,relativetobaselinescenario1.
Sharedlifetimeearningswerecalculatedbysummingoveralifetimeownearningsforeachyearthatapersonwassingleandone-halfofownearningsplusone-halfofspouse'searningsforeachyearthatapersonwasmarried.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Raisingtheminimumwageoffsetsthelifetimeearningslostthroughrisingwageinequalityforadultsnearthebottomoftheearningsdistribution.
Through2075,meansharedlifetimeearningsforadultsages67to75inthebottomquintilewerehigherunderscenario3—whichcombinedrisingwage-10%-8%-6%-4%-2%0%2%4%6%8%10%Bottom2nd3rd4thTopBottom2nd3rd4thTopBottom2nd3rd4thTop204520652085Scenario2Scenario3WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES31inequalitywithanincreaseintheminimumwage—thanunderbaselinescenario1.
Foradultsinthesecondquintile,theminimumwageincreasepartlyoffsettheearningslossesassociatedwithrisingwageinequality,butmeansharedlifetimeearningsremainedlowerunderscenario3thanthebaselinescenario.
Moreover,inbothofthebottomtwoquintiles,theprotectionsaffordedbytheminimumwagehikediminishedovertime,asrisingrealwageserodedthevalueoftheminimumwage.
Raisingtheminimumwagealsoreducedsomeofthelifetimeearningsgainsthatthetoptwoquintilesexperiencedaswageinequalitygrew,becausesomeofthewagegainsshiftedtothebottompartofthedistribution.
NetPerCapitaAnnuityIncomeatAges67to75DYNASIM4projectedthatretirementincomeswillgrowsignificantlyovertimeaslifetimeearningsgrow,butmeanretirementincomeswillgrowmoreslowlythanmeanlifetimeearnings.
Underthebaselinescenario,meaninflation-adjustednetpercapitaannuityincomeatages67to75forpeopleinthemiddlequintileofthesharedlifetimeearningsdistributionwillincreasefrom$39,500in2025to$44,700in2055,a13percentincrease(table6).
In2085,themeanvalueforthemiddlefifthofthedistributionwillreach$57,500,29percentmorethanin2055and46percentmorethanin2025.
Thebaselinescenarioprojectedthatretirementincomeswillgrowthroughoutthelifetimeearningsdistribution,althoughtheywillincreasesomewhatmoreslowlyforthosenearthetopofthedistributionthanforthosenearthemiddleorbottom.
Moreover,retirementincomeunderthebaselinewasprojectedtogrowonlyabouthalfasfastaslifetimeearnings.
32WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESTABLE6ProjectedMeanNetAnnuityIncomeforAdultsAges67to75byYearandQuintileofSharedLifetimeEarnings2015inflation-adjusteddollarsYear2025203520452055206520752085BottomquintileBaseline10,40010,10010,70011,40011,70013,30015,200Scenario210,40010,20010,40010,60011,00012,00013,300Scenario310,40010,30010,70011,40011,40012,90014,000SecondquintileBaseline25,00024,50025,50028,30030,20033,90037,700Scenario224,70024,30025,20027,70029,10032,50035,100Scenario324,90024,50025,40028,00029,40032,80035,700ThirdquintileBaseline39,50039,30040,20044,70047,60053,10057,500Scenario239,20038,80039,80044,20047,10051,40054,900Scenario339,40039,00039,80044,40047,10051,40056,000FourthquintileBaseline60,20059,70062,20068,40072,10080,90085,300Scenario259,90059,50063,00068,10072,90079,20085,300Scenario360,20059,10062,70068,10072,30080,20085,000TopquintileBaseline119,900127,000136,300138,700146,100162,500162,800Scenario2120,400128,000139,900143,300152,900167,100173,900Scenario3119,600126,100136,700141,300149,900165,100173,800Source:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:NetpercapitaannuityincomeincludedSocialSecurity,DBpension,annuitizedassetincome,SSI,earnings,imputedrent,andfederalassistanceprograms,lessfederalandstateincometaxes,payrolltaxes,andMedicarePartsBandDpremiums.
Sharedlifetimeearningswerecalculatedbysummingoveralifetimeownearningsforeachyearthatapersonwassingleandone-halfofownearningsplusone-halfofspouse'searningsforeachyearthatapersonwasmarried.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Theprojectionsshowthatrisingwageinequalitywillreducesimulatednetpercapitaannuityincomeatages67to75forpeoplenearthebottomofthesharedlifetimeearningsdistributionandincreaseincomeforpeoplenearthetop(figure13).
Aswithlifetimeearnings,theeffectsonincomeatages67to75grewovertime.
Comparedwiththebaselinescenario,meanincomeatages67to75forpeopleinthebottomquintileofthelifetimeearningsdistributionwas3percentlowerin2045,6percentlowerin2065,and13percentlowerin2085.
Thesepercentagedeclinesexceededthoseobservedforsharedlifetimeearnings.
Meanincomesinthesecondandthirdsharedlifetimeearningsquintileswerealsolowerunderscenario2,whichraisedwageinequality,thanthebaseline,buttheWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES33reductionsweresmallerthanforthebottomquintile.
Inthetopsharedlifetimeearningsquintile,meanincomeatages67to75underscenario2exceededthebaselinemeanby3percentin2045,5percentin2065,and7percentin2085,similartotheincreaseswesimulatedforsharedlifetimeearnings.
FIGURE13EffectsofRisingWageInequalityandanIncreaseintheMinimumWageonNetPerCapitaAnnuityIncomebyQuintileofSharedLifetimeEarningsAdultsages67to75URBANINSTITUTESource:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:Thefigureshowsthepercentagechangeinmeanpercapitaannuityincomebyquintileofsharedlifetimeearningsrelativetobaselinescenario1.
NetpercapitaannuityincomeincludedSocialSecuritybenefits,DBpensionbenefits,SSI,earnings,annuitizedassetincome,imputedrent,andfederalassistanceprogramsnetoftaxesandMedicarepremiums.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Raisingtheminimumwageoffsetmuchoftheretirementincomelossesinthebottomtwolifetimeearningsquintilesthatresultedfromgrowingwageinequality.
Comparedwiththebaselinescenariosimulations,2045meannetpercapitaannuityincomeatages67to75underscenario3wasonly0.
1percentlowerforthebottomquintileand0.
4percentlowerforthesecondquintile.
Raisingtheminimumwagedidnotaffectretirementincomesmuchinthethirdorfourthlifetimeearningsquintiles,butitreducedretirementincomesinthetopincomequintile,offsettingmuchofthegainsthequintile-15%-10%-5%0%5%10%Bottom2nd3rd4thTopBottom2nd3rd4thTopBottom2nd3rd4thTop204520652085Scenario2Scenario334WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESexperiencedaswageinequalityincreased.
Theimpactoftheminimumwageonretirementincomesdiminishedoverthedecades,asrealwagegrowtherodedthevalueofthesimulatedminimum.
Amongretireesinthebottomfifthofthelifetimeearningsdistribution,ahigherminimumwagewouldoffsetabout57percentoftheincomeatages67to75losttorisingwageinequalityin2065and37percentoflostretirementincomein2085.
Changesinthewagedistributionunderscenarios2and3slowlyraisedoverallretirementincomeinequality.
Underscenario2thatraisedwageinequality,theGinicoefficientofnetpercapitaannuityincomeatages67to75begantoexceedthebaselineGiniaroundyear2035(figure14).
ThegapintheGinigrewoverthesimulationperiod,reaching0.
03attheendofthesimulation.
Theminimumwageincreaseunderscenario3partiallyoffsettheeffectofwageinequality,butitseffectdiminishedoverthesimulationperiod.
Bytheendofthesimulation,theGinicoefficientsunderscenarios2and3werevirtuallyindistinguishable.
WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES35FIGURE14GiniCoefficientofNetPerCapitaAnnuityIncomeAdultsages67to75URBANINSTITUTESource:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:NetpercapitaannuityincomeincludedSocialSecuritybenefits,DBpensionbenefits,SSI,earnings,annuitizedassetincome,imputedrent,andfederalassistanceprogramsnetoftaxesandMedicarepremiums.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhour,adjustedforinflationthereafter,andkeptwageinequalityatthesamelevelasinscenario2.
Comparingincomesatvariouspointsofthedistributionisanotherwaytomeasureincomeinequalityandidentifiestheportionofthedistributioninwhichincomesdiffermost.
Theratioofthe90thtothe10thpercentilesofthenetpercapitaannuityincomedistributionatages67to75followsasimilartimepatternunderthethreescenarios—itrisesuntil2065andthenfallstowardstheendofthesimulation(table7).
Itsvalues,however,differundereachscenario,reflectingdifferentlevelsofincomeinequality.
Theratioofthe90thtothe10thpercentileishighestunderscenario2,whichraisestheeducationpremium;theratioincreasesfrom8.
60in2025to11.
07in2065.
Scenario3,whichraisesthefederalminimumwage,reducedtheratiototherangeithadunderthebaselineuntil2065,offsettingtheeffectsoftheincreaseintheeducationpremium.
Insubsequentyears,theratiogrewfasterunderscenario3thanunderthebaseline,asinflationerodedthevalueoftheminimumwage.
0.
410.
420.
430.
440.
450.
460.
470.
4820152025203520452055206520752085GinicoefficientYearBaselineScenario2Scenario336WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESTABLE7IncomeRatiosforNetPerCapitaAnnuityIncomebyYearandMeasureAdultsAges67to75Year2025204520652085Ratioof90thto10thpercentilesBaseline8.
489.
0710.
308.
99Scenario28.
609.
6211.
0710.
56Scenario38.
389.
2210.
3810.
00Ratioof90thto50thpercentilesBaseline2.
502.
772.
772.
65Scenario22.
542.
872.
912.
86Scenario32.
502.
802.
822.
81Ratioof50thto10thpercentilesBaseline3.
393.
273.
723.
39Scenario23.
393.
363.
803.
70Scenario33.
353.
303.
683.
57Source:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Notes:NetpercapitaannuityincomeincludedSocialSecuritybenefits,DBpensionbenefits,SSI,earnings,annuitizedassetincome,imputedrent,andfederalassistanceprogramsnetoftaxesandMedicarepremiums.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhourandkeptwageinequalityatthesamelevelasinscenario2.
WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES37Undereachscenario,theratioofthe90thtothe50thpercentilesofthenetpercapitaannuityincomedistributionatages67to75wasabout2.
5in2025.
Itincreasedovertimeasstronglifetimeearningsgrowthforhigh-wageworkersraisedfutureretirementincomesnearthetopofthedistribution.
Raisingthefederalminimumwageinscenario3didnotchangetheratiomuch,becauseitdidnotraiselifetimeearningsmuchforpeopleinthemiddleoftheincomedistribution.
Theratioofthe50thtothe10thpercentilesofnetpercapitaincomewassubstantiallylowerunderscenario3thanunderscenario2,however,astheincreaseinthefederalminimumwageraisedlifetimeearningsandretirementincomesforlow-wageworkers.
Onemeasureofwellbeingbywhichtheminimumwageincreasecouldsuccessfullyoffsetrisingwageinequalityisthepovertyrate.
Risingwageinequalityledtoamodestincreaseintheprojectedpovertyrateamongadultsages67to75relativetothebaseline(table8).
Insomeyears,theminimumwageincreasenotonlycanceledtheeffectofwageinequality,butitreducedthepovertyratebelowthebaselinepovertyrate,probablybecausetheminimumwageincreasewasmosteffectivenearthebottomofthewagedistributionandconsequentlythebottomoftheretirementincomedistribution,wherepeoplefacethehighestriskofpoverty.
Theeffectofrisingwageinequalitywasweakerthantheminimumwageeffectattheverybottomofthewageandretirementincomedistribution,eventhoughitwasstrongeroverall.
TABLE8PovertyRatesatAges67to75byYearandProjectionScenario2025204520652085Baseline6.
44.
64.
23.
1Scenario26.
74.
84.
33.
2Scenario36.
54.
53.
83.
1Source:Authors'estimatesfromDYNASIM4ID942,ID942B,andID942C.
Risingwageinequalityandanincreaseintheminimumwageaffectednetretirementincomethroughvariousincomesourcesandtaxes.
Tounderstandhowthesechangesaffectdifferenttypesofincomeandtaxes,wedecomposednetpercapitaannuityincomein2065bytypeofincomeandtax,foreachsharedlifetimeearningsquintile(table9).
Thetoppanelshowstheeffectsofrisingwageinequalityoneachtypeofincomeandtaxasthedifferencebetweenmeanoutcomesunderscenario2andthebaseline.
Thebottompanelshowstheeffectsoftheincreaseintheminimumwageasthedifferencebetweenoutcomesunderscenarios3and2.
Thedifferenceforeachcategoryisexpressedrelativetonetincomeunderthebaseline,sothatthedecompositiontermssumtothetotaleffect.
38WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESTABLE9EffectsofGrowingWageInequalityandRaisingMinimumWageonIncomeamongPeopleAges67to75in2065byIncomeSource,TaxCategory,andSharedLifetimeEarningsQuintilePercentBottomquintileSecondquintileThirdquintileFourthquintileTopquintileEffectsofgrowingwageinequality(scenario2–scenario1)Totalpre-taxincome-6.
3-4.
2-1.
30.
810.
7SocialSecuritybenefits-3.
4-2.
7-0.
30.
20.
6SSI-0.
10.
00.
00.
00.
0Annuitizedfinancialincome-1.
8-1.
0-1.
20.
62.
2DBpensionincome0.
1-0.
2-0.
1-0.
60.
0Earnedincome0.
10.
30.
1-0.
37.
0Imputedrentalincome-1.
3-0.
40.
00.
90.
9Federalassistanceprograms0.
0-0.
20.
10.
1-0.
1Totaltax0.
0-0.
5-0.
2-0.
36.
1Federalincometax0.
0-0.
3-0.
1-0.
24.
3Stateincometax0.
0-0.
1-0.
10.
01.
0OASDItax0.
00.
00.
00.
0%0.
2HItax0.
00.
00.
00.
00.
1Medicaresurtax0.
00.
00.
00.
00.
5MedicarePartBpremium0.
0-0.
10.
00.
0%0.
0MedicarePartDpremium0.
1-0.
10.
00.
00.
0Netincome-6.
3-3.
7-1.
11.
24.
6Effectsofraisingtheminimumwage(scenario3–scenario2)Totalpretaxincome2.
61.
30.
3-0.
8-6.
2SocialSecuritybenefits2.
01.
1-0.
8-0.
6-0.
3SSI-0.
10.
00.
0000.
0Annuitizedfinancialincome-0.
70.
10.
1-0.
30.
6DBpensionincome-0.
20.
30.
00.
30.
0Earnedincome0.
30.
21.
40.
8-5.
6Imputedrentalincome1.
3-0.
5-0.
2-0.
9-1.
0Federalassistanceprograms0.
00.
0-0.
3-0.
10.
1Totaltax-0.
40.
30.
20.
0-4.
2Federalincometax-0.
30.
10.
10.
0-2.
8Stateincometax0.
00.
00.
00.
1-0.
8OASDItax0.
00.
00.
10.
1-0.
1HItax0.
00.
00.
00.
0-0.
1Medicaresurtax0.
00.
00.
00.
0-0.
4MedicarePartBpremium0.
00.
10.
0-0.
10.
0MedicarePartDpremium-0.
10.
00.
00.
00.
0Netincome3.
01.
00.
1-0.
8-2.
0Source:DYNASIM4ID942,ID942B,andID942CNotes:Thetoppanelreportsmeanincomeandtaxesunderscenario2minusmeanincomeandtaxesunderbaselinescenario1,andthebottompanelreportsmeanincomeandtaxesunderscenario3minusmeanincomeandtaxesunderscenario2,bothexpressedasapercentageofthebaseline.
Scenario1keptwageinequalityatthe2006levelandthefederalminimumwageatthecurrentlevelof$7.
25perhour.
Scenario2raisedwageinequality,asdescribedinthetext,andmaintainedthefederalminimumwageat$7.
25perhour.
Scenario3raisedthefederalminimumwageto$12.
00perhourandkeptwageinequalityatthesamelevelasinscenario2.
WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES39Themainmechanismthroughwhichrisingwageinequalityaffectedincomeatages67to75inthebottomtwoquintilesofthesharedlifetimeearningsdistributionwasSocialSecuritybenefits,whichaccountedforslightlymorethanone-halfofthereductioninmeanincomesineachquintile.
Totalpretaxincomefell6.
3percentinthebottomquintile,andlowerSocialSecuritybenefitsreducedpretaxincomeby3.
4percent.
Inthesecondquintile,totalpretaxincomefellby4.
2percent,andSocialSecuritybenefitcutsreducedpretaxincomeby2.
7percent.
Theremainingincomelosscamemostlyfromreductionsinannuitizedfinancialincomeandimputedrentalincome.
Taxesplayedvirtuallynoroleinthebottomquintile,andreducedtaxesfromrisingwageinequalityincreasednetincomebyonly0.
5percentinthesecondquintile.
IncreasedSocialSecuritybenefitsalsodrovetheretirementincomegrowthforthesetwoquintilesthatresultedfromtheincreaseintheminimumwage.
SocialSecuritybenefitscontributedtwo-thirdsoftheoverallgrowthinnetincomeinthebottomquintileandmorethan100percentofthenetincomegrowthinthesecondquintile.
Inthethirdandfourthquintiles,individualincomesourcesandtaxeschangedincomebysmallamounts—inmostcasesbylessthan1percent—andthereforeweredifficulttointerpretconfidently.
Inthetopquintileofthesharedlifetimeearningsdistribution,theimpactofbothrisingwageinequalityandtheminimumwageincreaseworkedmostlythroughchangesinearnings.
Risingwageinequalityincreasedtotalpretaxincomeinthetopquintileby10.
7percent,ashigherearningsraisedpretaxincomeby7.
0percentandhigherannuitizedassetincomeraisedpretaxincomeby2.
2percent.
Highertaxesoffsetmanyofthesegains,however,reducingincomeby6.
1percent,foranetgainof4.
6percent.
Mostoftheadditionaltaxobligationcamefromfederalincometaxes,althoughstateincometaxandtheMedicaresurtaxalsoplayedarole.
Thesimulatedminimumwagehikereducedearningsinthisquintile,causingtotalpretaxincometofallby6.
2percent.
However,taxreductions,mostlyfromlowerfederalandstateincometaxes,offsetmuchoftop-quintileincomelossesfromtheminimumwageincrease,resultinginanetincomereductionof2.
0percent.
Thecombinedimpactofrisingwageinequalityandaminimumwagehikeincreasednetincomeforthetopquintileby2.
6percent.
ConclusionsWageinequality,whichhasbeenrisingintheUnitedStatesfordecades,affectsincomethroughoutthelifecourse.
Ifthehourlywagegapbetweencollegeandhighschoolgraduatescontinuestogrowatitscurrentpace,thetopfifthoflifetimeearningsatages67to75wouldincrease2percentin2045,5percentin2065,and8percentin2085.
Atthesametime,lifetimeearningsinthebottomfifthofthedistributionwouldfall2percentin2045,5percentin2065,and9percentin2085.
Theimpactwould40WAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMESreverberateintoretirement,becauseSocialSecuritybenefitsaretiedtolifetimeearnings,andhowmuchpeoplecansaveforretirementdependsonhowmuchtheyearn.
Ourprojectionsshowthatmeanincomeatages67to75,measuredbroadlytoaccountforthevalueofowner-occupiedhousingandretirementsavingsandexcludingtaxandMedicarepremiumpayments,wouldincreaseinthetopfifthofthedistributionby3percentin2045,5percentin2065,and7percentin2085.
Inthebottomfifthofthedistribution,itwouldfallby3percentin2045,6percentin2065,and13percentin2085.
Theselossesexceedthepercentagedeclineinlifetimeearnings,despiteSocialSecurity'sprogressivebenefitformulathatreplacesalargershareofpreretirementearningsforpeoplewithlimitedlifetimeearningsthanforthosewithmorelifetimeearnings.
Ahigherminimumwagecouldmitigatebutnoteliminatetheeffectsofgrowingwageinequalityonfutureretirementincomes.
Immediatelyraisingthefederalminimumwagefrom$7.
25to$12perhourandsubsequentlyadjustingitforinflationwouldraiselifetimeearningsandfutureretirementincomeforpeopleinthebottomtwo-fifthsofthelifetimeearningsdistribution,buttheimpactwoulddissipateovertime,despitethebuilt-ininflationadjustment.
Asaveragewagesgrowfasterthanprices,theshareofworkerswhowouldotherwiseearnlessthantheminimumwagewouldfalleachdecade.
Forretireesinthebottomfifthofthelifetimeearningsdistribution,thesimulatedincreaseintheminimumwagewouldoffset57percentoftheretirementincomelosttorisingwageinequalityin2065and37percentoflostretirementincomein2085.
Ourprojectionsofrisingwageinequalityfocussolelyonchangesovertimeintheeducationpremium—thewagegapbetweencollegegraduatesandworkerswithlesseducation.
Theincreaseintheeducationpremiumcanaccountformuchofthepastgrowthinwageinequality,anditiscloselyassociatedwithothersourcesofwageinequality,suchastheincreaseininternationaltradeanddeclineinlaborunions.
However,ourfocusontheeducationpremiumexcludessomeimportantdriversofgrowingwageinequality,particularlytherecentsurgeinwagesinthetop1percentofthedistribution.
Althoughonlyaverysmallshareofthepopulationcollectstheseveryhighearnings,theirsizeandrapidgrowthlimitwagegainsforotherworkers.
Becauseourprojectionsdonotaccountfortheseincreasesattheverytopofthewagedistribution,ourestimatesmayunderstatethefuturegrowthinwageinequality.
Howearningsandretirementincomesactuallyevolveoverthecomingdecadeswilldependonmanyfactors.
Akeydeterminantwillbethegrowthinaveragewages.
OuranalysisreliesonassumptionsdevelopedbySocialSecurity'sBoardofTrustees(2016),whoanticipate1.
2percentannualinflation-adjustedgrowthinwages.
Thisassumedgrowthrateissignificantlyhigherthanthehistoricaverageannualrateof0.
81percentfrom1963to2015.
DeviationsfromourassumedwageWAGEINEQUALITY,MINIMUMWAGE,ANDRETIREMENTINCOMES41growthratecouldsignificantlyaffectfutureearningsandretirementincome.
Newpatternsinmarriage,employment,earnings,health,andsavingsthatemergeoverthecomingdecadeswillalsodetermineoutcomes.
Policychoiceswillshapethosepatternsandhelpsetfutureretirementincomes.
AkeydeterminantwillbehowpolicymakersrespondtoSocialSecurity'slong-runfinancingshortfall.
Ourprojectionsassumethatfutureretireeswillreceivefullbenefitsscheduledundercurrentlaw.
TheSocialSecurity'sBoardofTrustees(2016),however,projectthatthesystemwillonlyhaveenoughfundstopayaboutthree-fourthsofscheduledbenefitsafterthemid-2030s.
CutstoSocialSecuritycouldreshapethedistributionoffutureretirementincome.
Boostingtheminimumwageandmaintainingitsvalueovertimeasaveragewagesrisewouldhelplow-wageworkersandimprovetheirretirementprospects,butabroaderpolicyagendaisneededtoensurefinancialsecurityacrossthelifespan(EllwoodandPatel2018).
Workertrainingandapprenticeshipprogramscouldboostlifetimeearningsforless-skilledworkers.
SocialSecurityreforms,suchascreatingameaningfulminimumbenefitorchangingthebenefitformulatoimproveprogressivity,couldraisebenefitsforretireeswithlowlifetimeearnings.
Policyoptionstoboostretirementsavingsincludefinancialliteracytrainingandfederalorstatemandatesrequiringemployersthatdonotofferapensionplantoallowworkerstoautomaticallydivertaportionoftheirpaychecktoaretirementaccount.
Taxreformscouldraisesavingsincentivesforlow-wageworkerswhodonotbenefitmuchfromincometaxdeductions,theprimarysavingsincentiveinthecurrenttaxcode.
Together,theseprogramscouldprovideallworkerswithapathtoretirementsecurityevenaswageinequalitygrows.
42APPENDIXAppendixAPPENDIXTABLE1EstimatedCoefficientsandStandardErrorsintheLaborForceParticipationEquationbySexandRaceBlackmenNon-blackmenBlackwomenNon-blackwomenMarried0.
780***0.
654***0.
0568***-0.
192***(0.
0167)(0.
00781)(0.
0117)(0.
00579)Divorced0.
237***0.
192***0.
234***0.
227***(0.
0237)(0.
0120)(0.
0155)(0.
00854)Disabilitylimitswork-1.
301***-1.
153***-1.
225***-0.
944***(0.
0200)(0.
00753)(0.
0179)(0.
00697)Hispanic0.
0239-0.
0256***-0.
207***-0.
0696***(0.
0448)(0.
00737)(0.
0323)(0.
00481)RegionSouth0.
162***-0.
124***0.
103***-0.
166***(0.
0152)(0.
00565)(0.
0121)(0.
00373)Northeast0.
0295-0.
0736***0.
0399***-0.
0823***(0.
0191)(0.
00600)(0.
0150)(0.
00396)West-0.
0181-0.
142***-0.
0288-0.
134***(0.
0234)(0.
00605)(0.
0187)(0.
00399)EducationalattainmentHighschoolgraduate0.
325***0.
358***0.
559***0.
542***(0.
0594)(0.
0269)(0.
0581)(0.
0179)Somecollege0.
465***0.
418***0.
780***0.
719***(0.
0601)(0.
0270)(0.
0584)(0.
0180)Collegegraduate0.
555***0.
311***0.
961***0.
994***(0.
0894)(0.
0331)(0.
0746)(0.
0222)Graduateschool0.
645***0.
409***1.
090***1.
283***(0.
0950)(0.
0340)(0.
0779)(0.
0230)Numberofchildren0.
0387***0.
0737***-0.
0212***-0.
0830***(0.
00539)(0.
00233)(0.
00370)(0.
00134)Numberofchildrenyoungerthan5-0.
00452-0.
0415***-0.
225***-0.
378***(0.
0167)(0.
00647)(0.
0114)(0.
00316)Spouse'slogrealincome0.
0187***-0.
0104***0.
00211-0.
101***(0.
00453)(0.
00139)(0.
00397)(0.
00147)Foreignborn0.
0414*-0.
0620***-0.
146***-0.
329***(0.
0235)(0.
00862)(0.
0177)(0.
00552)Foreignbornmissing0.
278***0.
141***0.
0821***0.
0157*(0.
0386)(0.
0131)(0.
0305)(0.
00903)HealthExcellent0.
674***0.
544***0.
474***0.
275***(0.
0322)(0.
0116)(0.
0262)(0.
00829)Verygood0.
622***0.
536***0.
492***0.
340***(0.
0298)(0.
0110)(0.
0245)(0.
00811)Good0.
439***0.
396***0.
370***0.
262***(0.
0285)(0.
0108)(0.
0236)(0.
00818)APPENDIX43BlackmenNon-blackmenBlackwomenNon-blackwomenFair0.
115***0.
0932***0.
104***0.
0187*(0.
0308)(0.
0122)(0.
0255)(0.
00967)ReceivesSocialSecurity-1.
233***-1.
395***-0.
801***-0.
781***(0.
0199)(0.
00695)(0.
0162)(0.
00617)AgesplinesAge25-0.
0237-0.
006020.
0264*0.
00292(0.
0146)(0.
00635)(0.
0138)(0.
00429)Age300.
0307*-0.
00626-0.
0226-0.
00111(0.
0187)(0.
00807)(0.
0169)(0.
00537)Age40-0.
004140.
000523-0.
0218***-0.
0310***(0.
00974)(0.
00418)(0.
00764)(0.
00274)Age50-0.
0421***-0.
0269***-0.
0136*-0.
0162***(0.
00900)(0.
00375)(0.
00701)(0.
00269)Age600.
01010.
00435-0.
0414***0.
00762(0.
0133)(0.
00522)(0.
0120)(0.
00472)Age65-0.
0321**-0.
0361***0.
0123-0.
0425***(0.
0127)(0.
00482)(0.
0130)(0.
00520)InteractionsofagesplineseducationalattainmentAge25Xcollegegraduate0.
02650.
0503***0.
0128-0.
0466***(0.
0247)(0.
00918)(0.
0198)(0.
00572)Age30Xcollegegraduate-0.
0279-0.
0488***-0.
02200.
0185***(0.
0318)(0.
0118)(0.
0247)(0.
00706)Age40Xcollegegraduate-0.
00613-0.
003880.
0204*0.
0625***(0.
0174)(0.
00616)(0.
0122)(0.
00353)Age50Xcollegegraduate-0.
0316*-0.
0479***-0.
0538***-0.
0632***(0.
0171)(0.
00553)(0.
0124)(0.
00376)Age60Xcollegegraduate0.
0555**0.
0794***0.
0127-0.
00573(0.
0263)(0.
00754)(0.
0223)(0.
00711)Age65Xcollegegraduate-0.
0289-0.
0282***0.
008480.
0370***(0.
0264)(0.
00709)(0.
0246)(0.
00810)Age25Xhighschoolgraduate0.
0221-0.
00315-0.
0113-0.
0317***(0.
0166)(0.
00736)(0.
0152)(0.
00473)Age30Xhighschoolgraduate-0.
03060.
003270.
010000.
0322***(0.
0213)(0.
00938)(0.
0188)(0.
00593)Age40Xhighschoolgraduate-0.
001080.
001460.
004780.
00947***(0.
0114)(0.
00489)(0.
00879)(0.
00305)Age50Xhighschoolgraduate-0.
00846-0.
0323***-0.
0165*-0.
0278***(0.
0110)(0.
00447)(0.
00846)(0.
00305)Age60Xhighschoolgraduate0.
01080.
0440***0.
004100.
0131**(0.
0174)(0.
00626)(0.
0152)(0.
00538)Age65Xhighschoolgraduate0.
0167-0.
0121**-0.
00414-0.
00300(0.
0175)(0.
00587)(0.
0169)(0.
00600)TimetermsTime-2.
910***-1.
977***-1.
990***-1.
333***(0.
338)(0.
124)(0.
263)(0.
0817)Time^2/10019.
21***13.
34***13.
99***9.
626***(2.
389)(0.
882)(1.
872)(0.
585)Time^3/1000-5.
997***-4.
278***-4.
594***-3.
208***(0.
809)(0.
300)(0.
636)(0.
200)Time^4/100000.
892***0.
654***0.
715***0.
503***(0.
131)(0.
0488)(0.
103)(0.
0327)Time^5/100000-0.
0511***-0.
0385***-0.
0427***-0.
0302***(0.
00817)(0.
00305)(0.
00647)(0.
00205)44APPENDIXBlackmenNon-blackmenBlackwomenNon-blackwomenConstant16.
45***12.
07***10.
44***8.
395***(1.
845)(0.
675)(1.
428)(0.
440)Observations919699456501178461147207Source:Authors'estimatesfromCPS(1976–2012).
Notes:Standarderrorsareinparentheses.
*p<0.
1,**p<0.
05,***p<0.
01.
APPENDIX45APPENDIXTABLE2EstimatedCoefficientsandStandardErrorsintheSecond-StageOrdinaryLeastSquaresRegressionofLogHourlyWagebySexandRaceBlackmenNon-blackmenBlackwomenNon-blackwomenMeanlogrealhourlywage0.
577***0.
832***0.
621***0.
564***(0.
0550)(0.
0202)(0.
0626)(0.
0236)Collegepremium0.
751***0.
809***0.
428***0.
805***(0.
0824)(0.
0193)(0.
0658)(0.
0223)Married0.
191***0.
221***0.
0568***0.
0568***(0.
00794)(0.
00282)(0.
00452)(0.
00233)Hispanic-0.
0123-0.
144***-0.
0160-0.
0449***(0.
0180)(0.
00230)(0.
0158)(0.
00249)RegionSouth-0.
117***-0.
0208***-0.
0814***0.
00408**(0.
00639)(0.
00177)(0.
00588)(0.
00195)Northeast0.
01180.
0647***0.
0778***0.
0996***(0.
00790)(0.
00182)(0.
00705)(0.
00202)West0.
0251***0.
0407***0.
0565***0.
0563***(0.
00960)(0.
00190)(0.
00900)(0.
00212)EducationalattainmentHighschoolgraduate0.
134***0.
151***0.
103***0.
188***(0.
0312)(0.
00930)(0.
0348)(0.
0118)Somecollege0.
272***0.
270***0.
279***0.
358***(0.
0314)(0.
00934)(0.
0351)(0.
0119)Collegegraduate0.
105*-0.
0248*0.
272***0.
130***(0.
0536)(0.
0136)(0.
0488)(0.
0162)Graduateschool0.
262***0.
0794***0.
486***0.
348***(0.
0542)(0.
0138)(0.
0500)(0.
0166)AgesplinesAge25-0.
002610.
00711***0.
0003860.
000949(0.
00789)(0.
00229)(0.
00846)(0.
00291)Age300.
00643-0.
00179-0.
004880.
0000290(0.
00987)(0.
00286)(0.
0103)(0.
00357)Age40-0.
00627-0.
00575***0.
00774*-0.
000633(0.
00470)(0.
00145)(0.
00462)(0.
00175)Age500.
00448-0.
00123-0.
00870*-0.
000910(0.
00455)(0.
00151)(0.
00450)(0.
00191)Age60-0.
007470.
0126***0.
0278***0.
00357(0.
00937)(0.
00348)(0.
0103)(0.
00459)Age650.
00757-0.
00422-0.
0271*-0.
0104*(0.
0121)(0.
00448)(0.
0148)(0.
00633)InteractionsofagesplineseducationalattainmentAge25Xcollegegraduate0.
01460.
0323***0.
0209**0.
0314***(0.
00990)(0.
00268)(0.
00957)(0.
00320)Age30Xcollegegraduate-0.
0125-0.
0229***-0.
0113-0.
0355***(0.
0123)(0.
00333)(0.
0118)(0.
00396)Age40Xcollegegraduate-0.
00292-0.
0161***-0.
0159***-0.
00367*(0.
00600)(0.
00168)(0.
00554)(0.
00200)Age50Xcollegegraduate-0.
0128*0.
001370.
008060.
00488**(0.
00668)(0.
00185)(0.
00606)(0.
00228)Age60Xcollegegraduate0.
0120-0.
0131***-0.
0144-0.
0182***(0.
0160)(0.
00466)(0.
0155)(0.
00615)46APPENDIXBlackmenNon-blackmenBlackwomenNon-blackwomenAge65Xcollegegraduate-0.
003500.
0149**0.
008260.
0193**(0.
0223)(0.
00652)(0.
0221)(0.
00900)Age25Xhighschoolgraduate0.
01300.
0173***0.
01140.
00720**(0.
00841)(0.
00244)(0.
00888)(0.
00306)Age30Xhighschoolgraduate-0.
00963-0.
0149***-0.
00435-0.
00856**(0.
0105)(0.
00306)(0.
0109)(0.
00378)Age40Xhighschoolgraduate-0.
00510-0.
00561***-0.
0147***0.
000335(0.
00509)(0.
00156)(0.
00496)(0.
00186)Age50Xhighschoolgraduate-0.
00417-0.
002510.
0104**-0.
000659(0.
00521)(0.
00166)(0.
00500)(0.
00204)Age60Xhighschoolgraduate0.
00111-0.
00746*-0.
0244**-0.
0110**(0.
0116)(0.
00393)(0.
0120)(0.
00501)Age65Xhighschoolgraduate0.
003440.
0121**0.
02770.
0139**(0.
0158)(0.
00537)(0.
0179)(0.
00704)Millsratio-0.
267***-0.
406***-0.
230***-0.
140***(0.
0160)(0.
00658)(0.
0147)(0.
00492)Birthcohorts1906–100.
151-0.
0443-0.
0355-0.
110*(0.
106)(0.
0400)(0.
123)(0.
0616)1911–150.
1610.
0243-0.
0567-0.
0673(0.
0990)(0.
0384)(0.
102)(0.
0553)1916–200.
177*0.
0629*-0.
0305-0.
115**(0.
0973)(0.
0376)(0.
0967)(0.
0547)1921–250.
174*0.
0734**-0.
0709-0.
139**(0.
0956)(0.
0373)(0.
0984)(0.
0543)1926–300.
170*0.
0439-0.
0871-0.
177***(0.
0954)(0.
0373)(0.
0967)(0.
0544)1931–350.
161*-0.
00738-0.
110-0.
205***(0.
0955)(0.
0373)(0.
0965)(0.
0545)1936–400.
161*-0.
0486-0.
135-0.
241***(0.
0955)(0.
0373)(0.
0973)(0.
0547)1941–450.
118-0.
0911**-0.
159-0.
252***(0.
0958)(0.
0373)(0.
0980)(0.
0550)1946–500.
0827-0.
154***-0.
187*-0.
273***(0.
0960)(0.
0373)(0.
0988)(0.
0552)1951–55-0.
00625-0.
198***-0.
237**-0.
294***(0.
0962)(0.
0373)(0.
0996)(0.
0555)1956–60-0.
0730-0.
230***-0.
300***-0.
322***(0.
0965)(0.
0373)(0.
101)(0.
0558)1961–65-0.
105-0.
274***-0.
336***-0.
349***(0.
0971)(0.
0374)(0.
102)(0.
0562)1966–70-0.
123-0.
312***-0.
360***-0.
381***(0.
0976)(0.
0374)(0.
103)(0.
0566)1971–75-0.
153-0.
337***-0.
370***-0.
406***(0.
0984)(0.
0375)(0.
104)(0.
0570)1976–85-0.
213**-0.
376***-0.
432***-0.
460***(0.
0992)(0.
0376)(0.
106)(0.
0575)1986–96-0.
387***-0.
401***-0.
529***-0.
531***(0.
114)(0.
0414)(0.
118)(0.
0600)Constant0.
826***0.
07910.
970***1.
017***(0.
177)(0.
0727)(0.
171)(0.
0754)Observations6424573250372274681046APPENDIX47BlackmenNon-blackmenBlackwomenNon-blackwomenAdjustedR20.
1810.
2180.
2130.
175Source:Authors'estimatesfromCPS(1976–2012).
Notes:Standarderrorsinparentheses.
*p<0.
1,**p<0.
05,***p<0.
01.
APPENDIXTABLE3EstimatesofFixedEffectsRegressionoftheWageEquationErrorTermonItsLaggedValueMenWomenBlackNon-blackBlackNon-blackLaggederrorterm0.
218***0.
262***0.
113***0.
144***(0.
006)(0.
006)(0.
009)(0.
009)Constant0.
0164***0.
0443***-0.
0168***0.
0980***(0.
002)(0.
002)(0.
003)(0.
003)Observations31257378131508713410(standarddeviationofpermanenterrorterm)0.
4130.
4030.
4260.
436(standarddeviationoftransitoryerrorterm)0.
3820.
3620.
3750.
380Source:Authors'estimatesfromthePanelStudyofIncomeDynamics(1981–2011).
Notes:Theerrortermwascalculatedasthedifferencebetweenthehourlywagerecordedbythesurveyandthevaluepredictedbythemodelofhourlywagedescribedabove.
Standarderrorsareinparentheses.
*p<0.
1,**p<0.
05,***p<0.
01.
48APPENDIXAPPENDIXTABLE4EstimatedCoefficientsandStandardErrorsintheLogisticRegressionofBeingExemptfromtheMinimumWageLawbySexandRaceBlackmenNon-blackmenBlackwomenNon-blackwomenMarried0.
246***0.
322***0.
171***-0.
0453***(0.
0155)(0.
00464)(0.
0140)(0.
00445)Hispanic0.
127***-0.
305***0.
0759**-0.
272***(0.
0393)(0.
00741)(0.
0373)(0.
00799)EducationalattainmentHighschoolgraduate-0.
0172-0.
312***-0.
148-0.
220***(0.
101)(0.
0299)(0.
0993)(0.
0303)Somecollege0.
226**0.
0714**0.
1150.
00680(0.
102)(0.
0300)(0.
0996)(0.
0304)Collegegraduate1.
414***1.
359***1.
116***0.
769***(0.
135)(0.
0319)(0.
103)(0.
0294)Graduateschool1.
451***1.
219***1.
024***0.
646***(0.
138)(0.
0328)(0.
106)(0.
0304)Birthcohort1906–1000.
239-1.
3200.
145(.
)(0.
322)(0.
964)(0.
346)1911–15-1.
742***-0.
620***-1.
244***-0.
497***(0.
380)(0.
106)(0.
389)(0.
110)1916–20-1.
781***-0.
857***-1.
156***-0.
751***(0.
248)(0.
0693)(0.
253)(0.
0729)1921–25-1.
730***-0.
917***-1.
218***-0.
714***(0.
204)(0.
0578)(0.
200)(0.
0589)1926–30-1.
848***-0.
883***-1.
222***-0.
724***(0.
180)(0.
0525)(0.
180)(0.
0530)1931–35-1.
443***-0.
791***-1.
140***-0.
648***(0.
168)(0.
0502)(0.
170)(0.
0502)1936–40-1.
431***-0.
769***-1.
070***-0.
552***(0.
163)(0.
0492)(0.
166)(0.
0489)1941–45-1.
244***-0.
766***-1.
003***-0.
525***(0.
161)(0.
0486)(0.
164)(0.
0483)1946–50-1.
179***-0.
667***-0.
896***-0.
456***(0.
159)(0.
0483)(0.
163)(0.
0479)1951–55-1.
007***-0.
512***-0.
737***-0.
344***(0.
158)(0.
0481)(0.
162)(0.
0477)1956–60-0.
845***-0.
410***-0.
591***-0.
229***(0.
158)(0.
0479)(0.
161)(0.
0475)1961–65-0.
714***-0.
385***-0.
512***-0.
171***(0.
157)(0.
0478)(0.
161)(0.
0473)1966–70-0.
614***-0.
325***-0.
447***-0.
175***(0.
157)(0.
0477)(0.
161)(0.
0472)1971–75-0.
551***-0.
338***-0.
461***-0.
179***(0.
155)(0.
0472)(0.
159)(0.
0466)1976–85-0.
572***-0.
376***-0.
445***-0.
231***(0.
156)(0.
0474)(0.
160)(0.
0468)AgesplinesAge16-0.
0691-0.
0686***-0.
0508-0.
0389***(0.
0474)(0.
0137)(0.
0474)(0.
0141)Age190.
139**0.
159***0.
04880.
102***(0.
0632)(0.
0184)(0.
0640)(0.
0195)APPENDIX49BlackmenNon-blackmenBlackwomenNon-blackwomenAge25-0.
0411*-0.
0784***0.
0290-0.
0620***(0.
0234)(0.
00642)(0.
0242)(0.
00791)Age50-0.
00834-0.
0009200.
003460.
0201***(0.
0139)(0.
00454)(0.
0149)(0.
00604)Age600.
02620.
0288**0.
0613-0.
00182(0.
0379)(0.
0140)(0.
0398)(0.
0179)Age650.
0276-0.
0387**-0.
0451-0.
0152(0.
0446)(0.
0167)(0.
0455)(0.
0209)InteractionsofagesplineseducationalattainmentAge22Xcollegegraduate0.
04550.
129***0.
140***0.
148***(0.
0485)(0.
0116)(0.
0380)(0.
0108)Age25Xcollegegraduate-0.
0698-0.
130***-0.
166***-0.
152***(0.
0502)(0.
0121)(0.
0401)(0.
0116)Age50Xcollegegraduate-0.
00125-0.
0284***-0.
0116-0.
0169***(0.
0161)(0.
00486)(0.
0164)(0.
00630)Age60Xcollegegraduate0.
0130-0.
00773-0.
03160.
0345*(0.
0472)(0.
0151)(0.
0476)(0.
0192)Age65Xcollegegraduate-0.
002590.
0326*0.
0922-0.
0124(0.
0585)(0.
0184)(0.
0576)(0.
0230)Age19Xhighschoolgraduate0.
009430.
0340***0.
0582*0.
0451***(0.
0319)(0.
00944)(0.
0300)(0.
00944)Age22Xhighschoolgraduate0.
04970.
0604***0.
03360.
0275**(0.
0421)(0.
0129)(0.
0385)(0.
0125)Age25Xhighschoolgraduate-0.
0602**-0.
0795***-0.
0930***-0.
0562***(0.
0303)(0.
00870)(0.
0300)(0.
00997)Age50Xhighschoolgraduate0.
00200-0.
0175***-0.
0191-0.
0309***(0.
0148)(0.
00474)(0.
0156)(0.
00619)Age60Xhighschoolgraduate-0.
00563-0.
0370**-0.
0509-0.
0175(0.
0426)(0.
0149)(0.
0435)(0.
0186)Age65Xhighschoolgraduate-0.
05230.
02480.
04060.
0300(0.
0517)(0.
0180)(0.
0513)(0.
0219)StateAK-0.
434***-0.
911***-0.
340**-0.
770***(0.
130)(0.
0282)(0.
134)(0.
0298)AZ0.
307***0.
0468*0.
439***0.
0104(0.
0978)(0.
0252)(0.
0985)(0.
0267)AR-0.
107-0.
0830***-0.
187***-0.
198***(0.
0735)(0.
0266)(0.
0698)(0.
0281)CA0.
204***-0.
01710.
347***0.
0159(0.
0549)(0.
0205)(0.
0499)(0.
0219)CO0.
291***0.
007090.
472***0.
0411*(0.
0857)(0.
0231)(0.
0873)(0.
0244)CT0.
132*0.
166***0.
413***0.
161***(0.
0757)(0.
0240)(0.
0660)(0.
0250)DE0.
444***0.
146***0.
640***0.
288***(0.
0630)(0.
0266)(0.
0558)(0.
0273)DC0.
550***0.
136***0.
782***0.
693***(0.
0534)(0.
0288)(0.
0470)(0.
0290)FL0.
274***0.
307***0.
315***0.
325***(0.
0539)(0.
0217)(0.
0486)(0.
0227)GA0.
548***0.
337***0.
601***0.
279***(0.
0533)(0.
0247)(0.
0474)(0.
0262)HI-0.
356*-0.
292***-0.
226-0.
0062350APPENDIXBlackmenNon-blackmenBlackwomenNon-blackwomen(0.
185)(0.
0398)(0.
228)(0.
0404)ID0.
237-0.
409***-0.
407-0.
554***(0.
216)(0.
0256)(0.
395)(0.
0280)IL0.
351***0.
136***0.
651***0.
172***(0.
0565)(0.
0216)(0.
0490)(0.
0228)IN0.
0891-0.
228***0.
218***-0.
218***(0.
0840)(0.
0243)(0.
0748)(0.
0260)IA-0.
357***-0.
378***-0.
203-0.
451***(0.
132)(0.
0239)(0.
135)(0.
0253)KS0.
000388-0.
336***-0.
113-0.
366***(0.
0850)(0.
0243)(0.
0865)(0.
0260)KY-0.
215**-0.
306***-0.
259***-0.
360***(0.
0900)(0.
0255)(0.
0901)(0.
0272)LA0.
239***0.
106***0.
07390.
0524*(0.
0618)(0.
0282)(0.
0555)(0.
0298)ME0.
0688-0.
530***0.
518**-0.
509***(0.
186)(0.
0246)(0.
218)(0.
0259)MD0.
332***-0.
002450.
473***0.
178***(0.
0545)(0.
0249)(0.
0485)(0.
0257)MA0.
141*0.
138***0.
418***0.
154***(0.
0828)(0.
0230)(0.
0725)(0.
0239)MI0.
171***-0.
191***0.
139**-0.
197***(0.
0617)(0.
0221)(0.
0561)(0.
0238)MN0.
0180-0.
212***-0.
0778-0.
259***(0.
0877)(0.
0227)(0.
0912)(0.
0240)MS-0.
0977-0.
0483*-0.
247***-0.
112***(0.
0612)(0.
0293)(0.
0559)(0.
0310)MO0.
214***-0.
161***0.
545***-0.
114***(0.
0760)(0.
0246)(0.
0649)(0.
0258)MT-0.
743**-0.
626***-0.
540-0.
849***(0.
340)(0.
0268)(0.
486)(0.
0288)NE-0.
131-0.
211***-0.
0796-0.
442***(0.
103)(0.
0240)(0.
105)(0.
0257)NV0.
270***-0.
0955***0.
234***-0.
0788***(0.
0792)(0.
0249)(0.
0798)(0.
0266)NH0.
238*-0.
0705***0.
536***-0.
159***(0.
142)(0.
0232)(0.
162)(0.
0246)NJ0.
729***0.
391***0.
904***0.
462***(0.
0591)(0.
0227)(0.
0526)(0.
0236)NM-0.
00627-0.
307***0.
350**-0.
323***(0.
150)(0.
0278)(0.
155)(0.
0291)NY0.
890***0.
281***0.
962***0.
330***(0.
0523)(0.
0214)(0.
0460)(0.
0223)NC0.
114**0.
0716***0.
112**0.
0413*(0.
0560)(0.
0231)(0.
0505)(0.
0245)ND-0.
620***-0.
448***-0.
157-0.
628***(0.
187)(0.
0252)(0.
255)(0.
0266)OH-0.
0227-0.
202***-0.
0499-0.
205***(0.
0619)(0.
0217)(0.
0564)(0.
0232)OK-0.
0812-0.
128***0.
198**-0.
128***(0.
0885)(0.
0263)(0.
0784)(0.
0278)OR0.
0106-0.
384***-0.
358*-0.
369***(0.
150)(0.
0251)(0.
189)(0.
0268)PA0.
336***-0.
0786***0.
615***-0.
0111APPENDIX51BlackmenNon-blackmenBlackwomenNon-blackwomen(0.
0630)(0.
0214)(0.
0553)(0.
0227)RI0.
0340-0.
179***0.
0588-0.
0953***(0.
0899)(0.
0251)(0.
0890)(0.
0257)SC-0.
02870.
0172-0.
0963*-0.
0145(0.
0614)(0.
0266)(0.
0558)(0.
0281)SD-0.
700***-0.
430***-0.
783***-0.
635***(0.
175)(0.
0249)(0.
285)(0.
0264)TN0.
0285-0.
0433*0.
0766-0.
0531*(0.
0708)(0.
0258)(0.
0642)(0.
0273)TX0.
262***0.
224***0.
340***0.
126***(0.
0535)(0.
0210)(0.
0481)(0.
0224)UT-0.
0635-0.
211***-0.
0605-0.
411***(0.
179)(0.
0244)(0.
228)(0.
0271)VT-0.
168-0.
260***0.
372-0.
271***(0.
238)(0.
0250)(0.
299)(0.
0258)VA0.
322***0.
0676***0.
433***0.
159***(0.
0580)(0.
0240)(0.
0516)(0.
0250)WA-0.
0256-0.
406***-0.
0514-0.
351***(0.
104)(0.
0243)(0.
118)(0.
0259)WV0.
0851-0.
535***0.
0116-0.
436***(0.
133)(0.
0270)(0.
133)(0.
0285)WI-0.
294***-0.
365***-0.
206**-0.
419***(0.
105)(0.
0232)(0.
0959)(0.
0248)WY-0.
445**-0.
556***-0.
159-0.
660***(0.
225)(0.
0254)(0.
242)(0.
0276)Constant-1.
468***-1.
271***-1.
671***-1.
278***(0.
166)(0.
0510)(0.
166)(0.
0506)Observations10093512300821350191180499Source:Authors'estimatesfromCPS-MORG(1994–2016).
Notes:Standarderrorsinparentheses.
*p<0.
1,**p<0.
05,***p<0.
01.
52APPENDIXAPPENDIXTABLE5ProjectedMeanNetAnnuityIncomeforPeopleAges67to75byEducationalAttainmentYears2025203520452055206520752085HighschooldropoutBaseline19,89418,78318,86820,82922,48825,26627,072Scenario219,89419,24218,90219,31320,73222,43623,314(0.
0%)(2.
4%)(0.
2%)(-7.
3%)(-7.
8%)(-11.
2%)(-13.
9%)Scenario320,20019,16419,18220,26120,89823,29925,011(1.
5%)(2.
0%)(1.
7%)(-2.
7%)(-7.
1%)(-7.
8%)(-7.
6%)HighschoolgraduateBaseline38,31735,72836,17039,74342,93347,39150,455Scenario238,11935,35434,84937,48840,26943,68946,171(-0.
5%)(-1.
0%)(-3.
7%)(-5.
7%)(-6.
2%)(-7.
8%)(-8.
5%)Scenario338,03935,30934,75038,39942,12643,96146,069(-0.
7%)(-1.
2%)(-3.
9%)(-3.
4%)(-1.
9%)(-7.
2%)(-8.
7%)SomecollegeBaseline49,83049,32450,95952,77458,23163,63966,462Scenario249,13648,57649,07649,32955,21258,11060,985(-1.
4%)(-1.
5%)(-3.
7%)(-6.
5%)(-5.
2%)(-8.
7%)(-8.
2%)Scenario349,48248,61949,23849,70254,67857,40361,964(-0.
7%)(-1.
4%)(-3.
4%)(-5.
8%)(-6.
1%)(-9.
8%)(-6.
8%)CollegegraduateBaseline76,99083,18183,01982,83488,54097,620102,409Scenario277,48984,26286,82388,12495,769103,465112,713(0.
6%)(1.
3%)(4.
6%)(6.
4%)(8.
2%)(6.
0%)(10.
1%)Scenario377,24383,18285,01386,65693,484104,000113,288(0.
3%)(0.
0%)(2.
4%)(4.
6%)(5.
6%)(6.
5%)(10.
6%)Source:DYNASIM4ID942,ID942B,andID942CNotes:Netpercapitaannuityincomeincludesearnings,SocialSecurity,DBpension,annuitizedassetincome,SSI,imputedrent,federalassistanceprograms,lessfederalincometax,stateincometax,OASDItax,HItax,Medicaresurtax,andMedicarePartBandPartDpremiums.
Percentagechangerelativetothebaselineisinparentheses.
APPENDIX53APPENDIXTABLE6EffectsofGrowingWageInequalityandRaisingMinimumWageonIncomeAmongPeopleAges67to75in2065byIncomeSource,TaxCategory,andEducationPercentSource:DYNASIM4ID942,ID942B,andID942CNotes:Netpercapitaannuityincomeincludesearnings,SocialSecurity,DBpension,annuitizedassetincome,SSI,imputedrent,federalassistanceprograms,lessfederalincometax,stateincometax,OASDItax,HItax,Medicaresurtax,MedicarePartBandPartDpremiums.
Upperpanelshowsdifferencesbetweenscenario2andthebaseline,bottompaneldifferencesbetweenscenarios3and2,bothexpressedaspercentageofthebaseline.
HighschooldropoutHighschoolgraduateSomecollegeCollegegraduateEffectsofgrowingwageinequalityIncome-7.
9-8.
0-8.
515.
6SocialSecuritybenefits-2.
1-2.
1-1.
41.
3SSI0.
10.
00.
00.
0Annuitizedfinancialincome-0.
6-2.
0-1.
93.
1DBpensionincome0.
1-0.
1-0.
60.
0Earnedincome-3.
6-3.
1-3.
99.
4Imputedrentalincome-1.
8-0.
6-0.
81.
7Federalassistanceprograms0.
00.
00.
10.
0Tax-0.
1-1.
8-3.
37.
4Federalincometax-0.
1-1.
3-2.
35.
3Stateincometax0.
0-0.
1-0.
51.
1tax-0.
2-0.
2-0.
10.
3HItax0.
00.
00.
00.
1Medicaresurtax0.
1-0.
1-0.
20.
5MedicarePartBpremium0.
2-0.
1-0.
20.
1MedicarePartDpremium0.
10.
0-0.
10.
0Netincome-7.
8-6.
2-5.
28.
2EffectsofraisingtheminimumwageIncome-0.
13.
3-0.
1-5.
0SocialSecuritybenefits1.
10.
7-0.
3-0.
4SSI0.
00.
00.
00.
0Annuitizedfinancialincome-0.
60.
7-0.
10.
3DBpensionincome0.
0-0.
10.
10.
1Earnedincome-1.
41.
70.
7-4.
1Imputedrentalincome0.
70.
3-0.
3-1.
0Federalassistanceprograms0.
10.
0-0.
10.
0Tax-0.
9-0.
20.
8-4.
0Federalincometax-0.
9-0.
20.
7-2.
7Stateincometax-0.
10.
00.
2-0.
8OASDItax-0.
10.
20.
0-0.
1HItax0.
00.
00.
0-0.
1Medicaresurtax-0.
10.
00.
0-0.
3MedicarePartBpremium0.
2-0.
1-0.
10.
0MedicarePartDpremium0.
00.
00.
00.
0Netincome0.
83.
4-0.
9-1.
154NOTESNotes1.
Forapopulationofnworkersarrangedinorderofincreasingwageincome,theGinicoefficientGforwageswcanbedefinedas=1(+12∑(+1)=1∑=1).
2.
SeeNeumark(2017)foranoverviewofresultsintherecentliterature.
3.
DYNASIM4imputesa3percentannualrateofreturntohousingequitythatrepresentsthesavingsinrentfromowningahome,netofinterestandhomemaintenancecosts.
4.
Theannuityfactorisrecalculatedeachyeartoreflectchangesinwealthasindividualsage,basedonDYNASIM4projectionsofwealthaccumulationandspenddownandchangesinlifeexpectancyandmaritalstatus.
Formarriedcouples,DYNASIM4assumesa50percentsurvivorannuity.
5.
FollowingAutor,weclassifiedone-halfofworkerswithsomecollegebutlessthanafour-yeardegree,randomlyselected,withthecollege-educatedgroupandtheotherhalfwiththehigh-school-educatedgroup.
6.
Weincludedtheaverageannualwageinthewageequationtoimprovethemodel'spredictivepower,asdescribedbelow,animportantconsiderationforaprojectionmodel.
7.
WeperformedinitialestimateswithboththeCPSandthePanelStudyofIncomeDynamics.
WepreferredtheCPSresultsbecausetheyweremorepreciseandhadgreaterpredictivepower.
TheCPSlikelyperformedbetterbecauseitgeneratedamuchlargersample,withmorethan2.
5millionobservations,comparedwithfewerthan200,000observationsinthePanelStudyofIncomeDynamics,andthusbetterrepresentedtheentirepopulation.
8.
Thismethoddoesnotreplicatethetopoftheearningsdistributionwell.
Tosimulatehighearners,weinsteadrandomlydrewearningsfromadistributionoftopearningsobtainedfromtheSocialSecurityAdministration.
DYNASIM4adjuststheseearningssothataveragesimulatedearningsmatchaverageearningsprojectedbytheSocialSecurity'sBoardofTrustees(2016).
9.
TheFairLaborStandardsActdefinesmanycategoriesofemployeesthatareexemptfromthefederalminimumwage,includingexecutive,administrative,professionalandoutsidesalesemployeespaidonasalarybasis,farmworkers,andsomeworkersemployedbyfirmswithannualgrossvolumeofsaleslessthan$500,000.
10.
Thewageelasticitywithrespecttotheminimumwageistheratioofthepercentagechangeinthehourlywagetothepercentagechangeintheminimumwage.
REFERENCES55ReferencesAaronson,Daniel.
2001.
"PricePass-throughandtheMinimumWage.
"ReviewofEconomicsandStatistics83(1):158–69.
Autor,DavidH.
2014.
"Skills,Education,andtheRiseofEarningsInequalityamongthe'Other99Percent.
'"Science34(6186):843–51.
Autor,DavidH.
,DavidDorn,andGordonH.
Hanson.
2013.
"TheChinaSyndrome:LocalLaborMarketEffectsofImportCompetitionintheUnitedStates.
"AmericanEconomicReview103(6):2121–68.
Autor,DavidH.
,AlanManning,andChristopherL.
Smith.
2016.
"TheContributionoftheMinimumWagetoUSWageInequalityoverThreeDecades:AReassessment.
"AmericanEconomicJournal:AppliedEconomics8(1):58–99.
Bee,Adam,andJoshuaMitchell.
2017.
"DoOlderAmericansHaveMoreIncomethanWeThink"SESHDWorkingPaperNo.
2017-39.
Washington,DC:USCensusBureau.
https://www.
census.
gov/content/dam/Census/library/working-papers/2017/demo/SEHSD-WP2017-39.
pdf.
Blau,FrancineD.
,andLawrenceM.
Kahn.
2016.
"TheGenderWageGap:Extent,Trends,andExplanations.
"Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w21913.
BoardofTrustees(BoardofTrustees,FederalOld-AgeandSurvivorsInsuranceandtheFederalDisabilityInsuranceTrustFunds).
2016.
The2016AnnualReportoftheBoardofTrusteesoftheFederalOld-AgeandSurvivorsInsuranceandFederalDisabilityInsuranceTrustFunds.
Washington,DC:BoardofTrustees.
Card,David.
1996.
"TheEffectofUnionsontheStructureofWages:ALongitudinalAnalysis.
"Econometrica64(4):957–79.
Card,David,andAlanB.
Krueger.
1994.
"MinimumWagesandEmployment:ACaseStudyoftheFast-FoodIndustryinNewJerseyandPennsylvania.
"AmericanEconomicReview84(4):772–93.
CBO(CongressionalBudgetOffice).
2014.
TheEffectsofaMinimum-WageIncreaseonEmploymentandFamilyIncome.
Washington,DC:CongressionalBudgetOffice.
https://www.
cbo.
gov/publication/44995.
———.
2016.
The2016Long-TermBudgetOutlook.
Washington,DC:CongressionalBudgetOffice.
https://www.
cbo.
gov/publication/51580.
Crystal,Stephen,andDennisG.
Shea.
1990.
"CumulativeAdvantage,CumulativeDisadvantage,andInequalityamongElderlyPeople.
"Gerontologist30(4):437–43.
———.
1990b.
"TheEconomicWell-BeingoftheElderly.
"ReviewofIncomeandWealth36(3):227–47.
Crystal,Stephen,DennisG.
Shea,andAdrianaM.
Reyes.
2016.
"CumulativeAdvantage,CumulativeDisadvantage,andEvolvingPatternsofLate-LifeInequality.
"Gerontologist.
https://doi.
org/10.
1093/geront/gnw056.
Ellwood,DavidT.
,andNishaG.
Patel.
2018.
RestoringtheAmericanDream:WhatWouldItTaketoDramaticallyIncreaseMobilityfromPovertyWashington,DC:USPartnershiponMobilityfromPoverty.
https://www.
mobilitypartnership.
org/restoring-american-dream.
Favreault,MelissaM.
,KarenE.
Smith,andRichardW.
Johnson.
2015.
TheDynamicSimulationofIncomeModel(DYNASIM):AnOverview.
Washington,DC:UrbanInstitute.
http://www.
urban.
org/research/publication/dynamic-simulation-income-model-dynasim-overview.
Fernald,JohnG.
2016.
"ReassessingLonger-RunU.
S.
Growth:HowLow"WorkingPaperSeries2016–18.
SanFrancisco,CA:FederalReserveBankofSanFrancisco.
https://ideas.
repec.
org/p/fip/fedfwp/2016-18.
html.
Ferraro,KennethF.
,andTetyanaPylypivShippee.
2009.
"AgingandCumulativeInequality:HowDoesInequalityGetUndertheSkin"Gerontologist49(3):333–43.
56REFERENCESFortin,NicoleM.
,andThomasLemieux.
1997.
"InstitutionalChangesandRisingWageInequality:IsThereaLinkage"JournalofEconomicPerspectives11(2):75–96.
Ganapati,Sharat,andJeffreyWeaver.
2017.
"MinimumWageandRetailPricePass-Through:EvidenceandEstimatesfromConsumptionData.
"SSRNScholarlyPaperID2968143.
Rochester,NY:SocialScienceResearchNetwork.
https://papers.
ssrn.
com/abstract=2968143.
Gordon,RobertJ.
2014.
"TheDemiseofU.
S.
EconomicGrowth:Restatement,Rebuttal,andReflections.
"NBERWorkingPaperNo.
19895.
Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w19895.
Hakobyan,Shushanik,andJohnMcLaren.
2016.
"LookingforLocalLaborMarketEffectsofNAFTA.
"ReviewofEconomicsandStatistics98(4):728–41.
Heckman,JamesJ.
1976.
"TheCommonStructureofStatisticalModelsofTruncation,SampleSelectionandLimitedDependentVariablesandaSimpleEstimatorforSuchModels.
"AnnalsofEconomicandSocialMeasurement5(4):475–92.
———.
1979.
"SampleSelectionBiasasaSpecificationError.
"Econometrica47(1):153–61.
Helpman,Elhanan.
2016.
"GlobalizationandWageInequality.
"NBERWorkingPaperNo.
22944.
Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w22944.
Hirsch,BarryT.
,andMacpherson,DavidA.
"UnionMembershipandCoverageDatabasefromtheCurrentPopulationSurvey.
"AccessedJuly26,2018.
http://www.
unionstats.
com/.
Jardim,Ekaterina,MarkC.
Long,RobertPlotnick,EmmavanInwegen,JacobVigdor,andHilaryWething.
2017.
"MinimumWageIncreases,Wages,andLow-WageEmployment:EvidencefromSeattle.
"NBERWorkingPaperNo.
23532.
Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w23532.
Katz,LawrenceF.
,andKevinM.
Murphy.
1992.
"ChangesinRelativeWages,1963-1987:SupplyandDemandFactors.
"QuarterlyJournalofEconomics107(1):35–78.
King,Miriam,StevenRuggles,SarahFlood,andRobertJ.
Warren.
2015.
"IntegratedPublicUseMicrodataSeries,CurrentPopulationSurvey:Version4.
0.
[Machine-ReadableDatabase].
"https://cps.
ipums.
org/cps/.
Kopczuk,Wojciech,EmmanuelSaez,andJaeSong.
2010.
"EarningsInequalityandMobilityintheUnitedStates:EvidencefromSocialSecurityDatasince1937*.
"QuarterlyJournalofEconomics125(1):91–128.
Leamer,EdwardE.
1996.
"InSearchofStolper-SamuelsonEffectsonUSWages.
"NBERWorkingPaperNo.
5427.
Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w5427.
Neumark,David.
2017.
"TheEmploymentEffectsofMinimumWages:SomeQuestionsWeNeedtoAnswer.
"NBERWorkingPaperNo.
23584.
Cambridge,MA:NationalBureauofEconomicResearch.
http://www.
nber.
org/papers/w23584.
Neumark,David,MarkSchweitzer,andWilliamWascher.
2004.
"MinimumWageEffectsthroughouttheWageDistribution.
"JournalofHumanResources39(2):425–50.
OECD(OrganisationforEconomicCooperationandDevelopment).
2017.
PreventingAgeingUnequally.
Paris:OECDPublishing,https://doi.
org/10.
1787/9789264279087-en.
O'Rand,AngelaM.
1996.
"ThePreciousandthePrecocious:UnderstandingCumulativeDisadvantageandCumulativeAdvantageovertheLifeCourse.
"Gerontologist36:230–38.
———.
2003.
"CumulativeAdvantageTheoryinLifeCourseResearch.
"InAnnualReviewofGerontologyandGeriatrics:FocusonEconomicOutcomesinLaterLife,editedbyStephanCrystalandDennisG.
Shea,14–30.
NewYork,NY:Springer.
Piketty,Thomas,andEmmanuelSaez.
2006.
"TheEvolutionofTopIncomes:AHistoricalandInternationalPerspective.
"AmericanEconomicReview96(2):200–05.
REFERENCES57Reich,Michael,SylviaAllegretto,andAnnaGodoey.
2017.
"Seattle'sMinimumWageExperience2015-16.
"CWEDPolicyBrief.
Berkeley,CA:CenteronWageandEmploymentDynamics,UniversityofCalifornia,Berkeley.
SocialSecurityAdvisoryBoard.
2015.
2015TechnicalPanelonAssumptionsandMethods,AReporttotheBoard,September2015.
Washington,DC:SocialSecurityAdvisoryBoard.
http://www.
ssab.
gov/Details-Page/ArticleID/656/2015-Technical-Panel-on-Assumptions-and-Methods-A-Report-to-the-Board-September-2015.
Stolper,WolfgangF.
,andPaulA.
Samuelson.
1941.
"ProtectionandRealWages.
"ReviewofEconomicStudies9(1):58–73.
Tinbergen,Jan.
1974.
"SubstitutionofGraduatebyOtherLabour.
"Kyklos27(2):217–26.
USBureauofLaborStatistics.
2016.
"CharacteristicsofMinimumWageWorkers,2015.
"BLSReportsNumber1061.
Washington,DC:USBureauofLaborStatistics.
https://www.
bls.
gov/opub/reports/minimum-wage/2015/home.
htm.
UrbanInstitute.
2015.
DYNASIM:ProjectingOlderAmericans'FutureWell-Being.
Washington,DC:UrbanInstitute.
http://www.
urban.
org/research/publication/dynasim-projecting-older-americans-future-well-being.
- Time200gana-894相关文档
- "鸽主姓名","地区","羽色","足环号","第三关排名","第二关排名","第一关排名","综合","综合排名"
- investments200gana-894
- pancreatic200gana-894
- weighted200gana-894
- excluded200gana-894
- transition200gana-894
cera：秋季美国便宜VPS促销，低至24/月起，多款VPS配置，自带免费Windows
介绍：819云怎么样？819云创办于2019，由一家从2017年开始从业的idc行业商家创办，主要从事云服务器，和物理机器819云—-带来了9月最新的秋季便宜vps促销活动，一共4款便宜vps，从2~32G内存，支持Windows系统，…高速建站的美国vps位于洛杉矶cera机房，服务器接入1Gbps带宽，采用魔方管理系统，适合新手玩耍！官方网站：https://www.8...
什么是BGP国际线路及BGP线路有哪些优势
我们在选择虚拟主机和云服务器的时候，是不是经常有看到有的线路是BGP线路，比如前几天有看到服务商有国际BGP线路和国内BGP线路。这个BGP线路和其他服务线路有什么不同呢？所谓的BGP线路机房，就是在不同的运营商之间通过技术手段时间各个网络的兼容速度最佳，但是IP地址还是一个。正常情况下，我们看到的某个服务商提供的IP地址，在电信和联通移动速度是不同的，有的电信速度不错，有的是移动速度好。但是如果...
咖啡主机22元/月起,美国洛杉矶弹性轻量云主机仅13元/月起,高防云20G防御仅18元/月
咖啡主机怎么样?咖啡主机是一家国人主机销售商，成立于2016年8月，之前云服务器网已经多次分享过他家的云服务器产品了，商家主要销售香港、洛杉矶等地的VPS产品，Cera机房 三网直连去程 回程CUVIP优化 本产品并非原生地区本土IP，线路方面都有CN2直连国内，机器比较稳定。咖啡主机目前推出美国洛杉矶弹性轻量云主机仅13元/月起,高防云20G防御仅18元/月;香港弹性云服务器，香港HKBN CN...
-
insomniac英文歌中有一句歌词是这样的：“here tonight”，谁知道这首歌曲叫什么名？百度商城百度商城里抽奖全是假的lunwenjiance论文检测，知网的是32.4%，改了以后，维普的是29.23%。如果再到知网查，会不会超过呢？www.yahoo.com.hk香港的常用网站www.ijinshan.com在电脑看港台电视台那个网站最好而又不用钱速度又快dadi.tv1223tv影院首页地址是什么？1223tv影院在哪里可以找到？机器蜘蛛《不思议迷宫》四个机器蜘蛛怎么得 获得攻略方法介绍dpscycle寻求LR 高输出宏本冈一郎本冈一郎是什么东西？？谁知道？？铂金血痕“斑斑的血痕”是什么意思？