國際勞工組織（ILO）：是什么在推動美國的財富不平等？-生產(chǎn)力、稅收和技能的作用

International

Labour

Organization

ILOWorkingPaper105

February/2024

?WhatIsDrivingWealthInequalityintheUnitedStatesofAmerica?

TheRoleofProductivity,TaxationandSkills

Authors/EkkehardErnst,Fran?oisLangot,RossanaMerola,FabienTripier

二—

二(

ofsucharbitrationasthefinaladjudicationofsuchadispute.

CopyrightInternationalLabourOrganization2024

ThisisanopenaccessworkdistributedundertheCreativeCommonsAttribution4.0InternationalLicense(

/licenses/by/4.0/

).Userscanreuse,share,adaptandbuildupontheoriginalwork,asdetailedintheLicense.TheILOmustbedearlycreditedastheown-eroftheoriginalwork.TheuseoftheemblemoftheILOisnotpermittedinconnectionwithusers'work.

Attribution-Theworkmustbecitedasfollows:Ernst,E,Langot,F.,Merola,R.,Tripier,F.What

IsDrivingWealthInequalityintheUnitedStatesofAmerica?:TheRoleofProductivity,Taxationand

Skills.ILoWorkingPaper105.Geneva:InternationalLabourOffice,2024.

Translations-Incaseofatranslationofthiswork,thefollowingdisclaimermustbeaddedalong

withtheattribution:ThistranslationwasnotcreatedbytheInternationalLabourOrganization(ILO)andshouldnotbeconsideredanofficialILOtranslation.TheILOisnotresponsibleforthecontentoraccuracyofthistranslation.

Adaptations-Incaseofanadaptationofthiswork,thefollowingdisclaimermustbeaddedalongwiththeattribution:ThisisanadaptationofanoriginalworkbytheInternationalLabour

Organization(ILO).ResponsibilityfortheviewsandopinionsexpressedintheadaptationrestssolelywiththeauthororauthorsoftheadaptationandarenotendorsedbytheILO.

ThisCClicensedoesnotapplytonon-ILOcopyrightmaterialsincludedinthispublication.Ifthematerialisattributedtoathirdparty,theuserofsuchmaterialissolelyresponsibleforclearingtherightswiththerightholder.

Anydisputearisingunderthislicensethatcannotbesettledamicablyshallbereferredtoarbitra-

TradeLaw(UNCITRAL).Thepartiesshallbeboundbyanyarbitrationawardrenderedasaresult

tioninaccordancewiththeArbitrationRulesoftheUnitedNationsCommissiononInternational

AllqueriesonrightsandlicensingshouldbeaddressedtotheILOPublishingUnit(RightsandLicensing),1211Geneva22,Switzerland,orbyemailtorights@.

ISBN9789220404904(print),ISBN9789220404911(webPDF),ISBN9789220404928(epub),ISBN9789220404935(mobi),ISBN9789220404942(html).ISSN2708-3438(print),ISSN2708-3446(digital)

/10.54394/AHFP2990

ThedesignationsemployedinILOpublications,whichareinconformitywithUnitedNationspractice,andthepresentationofmaterialthereindonotimplytheexpressionofanyopinionwhatsoeveronthepartoftheILOconcerningthelegalstatusofanycountry,areaorterritoryorofitsauthorities,orconcerningthedelimitationofitsfrontiers.

Theresponsibilityforopinionsexpressedinsignedartidles,studiesandothercontributionsrestssolelywiththeirauthors,andpublicationdoesnotconstituteanendorsementbytheILOoftheopinionsexpressedinthem.

Referencetonamesoffirmsandcommercialproductsandprocessesdoesnotimplytheiren-dorsementbytheILO,andanyfailuretomentionaparticularfirm,commercialproductorpro-cessisnotasignofdisapproval.

InformationonILOpublicationsanddigitalproductscanbefoundat:/publns

ILOWorkingPaperssummarizetheresultsofILOresearchinprogress,andseektostimulatediscussionofarangeofissuesrelatedtotheworldofwork.CommentsonthisILOWorkingPaperarewelcomeandcanbesenttoresearch@.

Authorizationforpublication:RichardSamans,DirectorofRESEARCH

ILOWorkingPaperscanbefoundat:/global/publications/working-papers

Suggestedcitation:

Ernst,E.,Langot,F.,Merola,R.,Tripier,F.2024.WhatIsDrivingWealthInequalityintheUnited

StatesofAmerica?:TheRoleofProductivity,TaxationandSkills,ILOWorkingPaper105(Geneva,

ILO).

/10.54394/AHFP2990

01J|?ILOWorkingPaper105

Abstract

OutoffourmajorstructuralchangesaffectingtheUSeconomy-namelyarisingshareofskilledworkers,skill-biasedtechnologicalchange,decreasingprogressivenessoftaxationandproductiv-ityslowdown-weshowthatthedeclineinproductivitygrowthnotonlyisthemaindriverofthewideningwealthdisparitiesobservedintheUnitedStatesofAmericaoverthepastfewdecades,butisalsotheonlymechanismthatcanexplaininequalitiesbothwithinandbetweenskillgroups.

AbouttheAuthors

EkkehardErnst:ILO,ResearchDepartment.E-mail:ernste@

Fran?oisLangot:LeMansUniversity;InstitutUniversitairedeFrance;ParisSchoolofEconomics;CEPREMAP.E-mail:francois.langot@univ-lemans.fr

RossanaMerola:ILO,ResearchDepartment.E-mail:merola@

FabienTripier:UniversitéParisDauphine-PSL,Laboratoired'conomiedeDauphine(LEDa);CEPREMAP.E-mail:fabien.tripier@dauphine.psl.eu

References

34

02?ILOWorkingPaper105

Tableofcontents

Abstract

AbouttheAuthors

01

01

?1Introduction05

2StylizedFacts

2.1.Inequalitybetweenskillgroups

2.2.Inequalitywithinskillgroups

2.3.Istherea"wealthpolarization"?

08

08

09

11

3Model

13

3.1.Households

13

3.2.Production

15

3.3.Aggregation16

4Calibration

4.1.Calibrationoftheconstantparameters

4.2.CalibrationofthestructuralchangesintheUSeconomy

17

17

18

5Identifyingthesourcesofrisinginwealthinequalities20

6Extendingtheanalysis:Changesinincomerisk

23

7Conclusionsandpolicyimplications

25

8Appendix:DetailedstatisticsfromtheSurveyofConsumerFinances26

03ILOWorkingPaper105

ListofFigures

Figure1:WealthandwageinequalitybetweenskillsintheUnitedStates

10

Figure2:WealthinequalityacrossandwithinskillgroupsintheUnitedStates,1989-201611

Figure3:WealthinequalitybetweenskillsintheUnitedStates,1989-2016

12

Figure4:StructuralchangesintheUSeconomy

19

Figure5:ChangesinincomeriskintheUSeconomy,skilledversusunskilledworkers,1968-

2010

23

04?ILOWorkingPaper105

ListofTables

Table1:WealthinequalityacrossskillgroupsintheUnitedStates:Shareofunskilledhouse-

holdsinthetotalpopulationandinthefourquartilesofthewealthdistribution,1989-

2016(percentage)

08

Table2:WealthinequalityacrossskillgroupsintheUnitedStates:Meanvalueofwealth

heldbyhouseholds,skilledversusunskilled,1989-2016(thousandsof2016USdollars)08

Table3:WealthinequalitywithinskillgroupsintheUnitedStates,1989-2016

09

Table4:WealthinequalitybetweenskillgroupsintheUnitedStates,1989-2016

12

Table5:Calibratedparameters

17

Table6:Calibrationofshocks

19

Table7:Comparisonbetweenempiricaldataandmodelresults

20

Table8:Calibrationofshocks

24

Table9:Comparisonbetweenempiricaldataandmodellingresults(withincomerisk)

24

TableA1:Averageearnings,incomeandwealthofUShouseholds,1989

27

TableA2:Averageearnings,incomeandwealthofUShouseholds,1995

27

TableA3:Averageearnings,incomeandwealthofUShouseholds,2001

27

TableA4:Averageearnings,incomeandwealthofUShouseholds,2004

28

TableA5:Averageearnings,incomeandwealthofUShouseholds,2007

28

TableA6:Averageearnings,incomeandwealthofUShouseholds,2010

28

TableA7:Averageearnings,incomeandwealthofUShouseholds,2013

29

TableA8:Averageearnings,incomeandwealthofUShouseholds,2016

29

TableA9:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,1989

31

TableA10:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,1995

31

TableA11:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,2001

31

TableA12:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,2004

31

TableA13:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,200732

TableA14:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,201032

TableA15:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,2013

32

TableA16:MeasuresofinequalityamongUShouseholds,totalandbyskillgroup,2016

33

05?ILOWorkingPaper105

?1Introduction

Overthepasttwodecades,incomeinequalityhasincreasedinalargemajorityofmembercoun-triesoftheOrganizationforEconomicCo-operationandDevelopment.Theaverageincomegapbetweenthetop10percentandthebottom50percentofindividualswithincountrieshasal-mostdoubled.Ifwelookatwealthinequality,thewealthgapbetweenupper-incomefamiliesandmiddle-andlower-incomefamiliesiswiderthantheincomegapandisgrowingmorerap-idly(Chanceletal.2022).AccordingtoastudyconductedbythePewResearchCenterin2020,"[t]herichestfamiliesintheU.S.haveexperiencedgreatergainsinwealththanotherfamiliesinrecentdecades,atrendthatreinforcesthegrowingconcentrationoffinancialresourcesatthetop"(Horowitz,Igielnik,andArditi2020).Identifyingthedrivingforcesofinequalityisakeychal-engeindesigningeffectivepoliciestoachieveinclusiveandequitablegrowthanddevelopment.

Existingstudieshavepointedtoseveralstructuralchangesthatmaybeleadingtogreaterin-equality.Weconsiderfouroftheminthispaper:decliningproductivity,thegrowingshareofskilledworkersintheeconomy,skill-basedtechnologicalchange,andatrendwherebytaxationisbecominglessprogressive.Withregardtothefirstoftheseprocesses,thatis,theslowdowninproductivitygrowth,thereisclearlyadiscrepancybetweenthelevelofinvestmentinthein-formationtechnologysectorandnationalproductivitylevels.Backin1993,ErikBrynjolfssonre-ferredtothissituationasthe"productivityparadox",whichisencapsulatedinthefollowingquip:“Youcanseethecomputerageeverywherebutintheproductivitystatistics”(Solow1987).!IntheUnitedStatesofAmerica,thegrowthinlabourproductivity-realoutputperhourworked-sloweddownmarkedlyaround2004.Overtheprevioustenyears,labourproductivityinthebusinesssectorhadrisenatanannualaveragerateofmorethan3percent.Itslowedtoabout2percentayearduring2004-10beforedroppingtoapaltry0.5percentduring2010-16.Moreprecisely,usingdatafromFernald(2014),weestimatethatthegrowthrateoftotalfactorpro-ductivitydecreasedsignificantlyfrom2.1percentbefore1989to0.89percentin2010.Iftherateofreturnoncapital,r,ismuchlargerthantherateofgrowth,g,thenexistingcapitalgrowsfasterthannewcapitalcreatedoutofaccumulatedincomeandalreadyrichownersofcapitalwillbecomeevenricher.Incontrast,ahighergrowthrategimplieshigherincomegrowthandena-blesthoseatthebottomofthedistributiontoaccumulatenewwealth,therebyreducingwealthconcentration.AsarguedbyPiketty(2014)andPikettyandZucman(2015),arateofreturnthatexceedseconomicgrowth(ther>gmechanism)isakeyfactordrivingeconomicinequality.Asecondfactorisrelatedtotheevolutionofsupplyof,anddemandfor,skills.Arisingshareofskilledworkerscanleadtoareductionininequalitiesastherelativescarcityofsuchworkersde-creases.?Athirddriverisskill-biasedtechnologicalchange,whichpushestheskillpremiumup,ultimatelyincreasinginequalitiesbetweenskillgroups(GoldinandKatz2009).Thefactthattheskillpremiumisrisingsuggeststhattheskillsrequiredtohandlenewtechnologiesareinscarcesupply(AcemogluandAutor2011).Moreover,itisgenerallyagreedthatskill-biasedtechnologi-calchangeledtoahollowing-outofthewagedistributioninthe2000s,whenmanymiddle-wagecognitiveroutineoccupationswereautomated(Goos,Manning,andSalomons2014).Thefourthdriverofrisinginequalityhastodowiththedegreeofprogressivenessofthetaxsystem.Overthepastdecade,severalcountrieshaveimplementedtaxreformsmakingtheirtaxstructuresmuchflatter(seeDuncanandSabirianovaPeter2012).Theshifttoflattaxes,ortoataxstruc-turewithlowerlevelsofprogressiveness,isconsistentwithadeclineinthetopincometaxratesandhasbeenassociatedwithhighlevelsofinequality.Gerberetal.(2018)findquitestrongev-idenceoftaxprogressivenesshavinganimpactoninequality.

Thelackofdiffusionofnewtechnologiescanbeviewedasamajorfactorbehindtheproductivityparadox(Ernst2022).

Itisworthnotingthatthiseffectmaybeambiguous.AzizandCortes(2021)arguethat,insocietieswheretheshareofskilledwork-ersisbelow50percent,ariseineducationalattainmentincreasesinequalitybetweenskillgroups.Conversely,insocietieswherethatshareisabove50percent,anincreaseineducationalattainmentputsdownwardpressureontheskillpremiumandultimatelyreducesinequalitybetweenskillgroups.

06?ILOWorkingPaper105

Toanalysethemaindriversofwealthinequality,wedevelopaclosed-economygeneralequilib-riummodelbasedonAhnetal.(2018).Weinvestigatetheimpactoffourstructuralchangesonrisingwealthinequalities:(a)theincreasinggapbetweeninterestrateandproductivity;(b)therisingshareofeducatedworkers;(c)skill-biasedtechnologicalchange;and(d)thereductionofprogressivenessintaxation.Inordertomeasuretheimpactofthesestructuralchanges,wecon-sideraheterogeneousagentsmodelàlaAiyagari(1994)withskilledandunskilledworkers(asinAhnetal.2018),wheretaxdistortionsmodifylaboursupplythroughpreferencesalaGreenwood,Hercowitz,andHuffman(1988).Thetaxationismodelledbythenon-linearfunctionproposedbyHeathcote,Storesletten,andViolante(2017).Finally,inordertohavethemodelgeneratewealthinequalitiessimilartothoseobservedempirically,weintroduceheterogeneousdiscountfactorsamongworkersasinKrusellandSmith(1998).Tosolveandsimulatethemodel,weemploythemethodologyofAchdouetal.(2021).Theaforementionedstructuralchangesareintroducedinthemodelaspermanentshocksthatmodifythesteady-statewealthdistribution.Theshocksarecalibratedusingempiricaldataon(a)theriseintheshareofskilledworkersintheUSlabourmarket;(b)thetechnologicalchangewhichfavoursskilledworkers;(c)thedecreasingprogres-sivenessofthetaxsystem;and(d)theslowdowninaggregateproductivitygrowth.Bycompar-ingtheinitialsteadystate(withoutanyshocks)andsteadystatesinwhichallshocksoronlyoneofthemhasoccurred,wecanmeasuretherelativecontributionofeachshocktotheriseinin-equality,togetherwiththecontributionsofinteractionsbetweentheseshocks.Themodellingresultsshowthatallshocks,exceptforthegrowingshareofskilledworkers,induceanincreaseinwealthinequalitybetweenskillgroups.Whenitcomestowealthinequalitywithinskillgroups,onlytheproductivityslowdowncanexplainthesurgeininequalityofthiskind,whichiseitherunchangedby,ordecreasesinresponseto,shocks(a)to(c).

Ouranalysiscontributestoinequalityresearchinseveralrespects.First,averylargebodyoflit-eratureinvestigatingwealthinequalitythroughgeneralequilibriummodelshasemergedsincetheseminalcontributionbyAiyagari(1994),withsuchrecentimportantadditionsasKaymakandPoschke(2016)andHubmer,Krusell,andSmithJr(2020).InthispaperwedrawonsuchearlierstudiestoidentifypotentialdriversofwealthinequalityintheUnitedStates.Weextendtheanalyticalapproachtakeninthosestudiesbyconsideringnotonlyvariationsintheoverallwealthdistributionbutalsochangesininequalitybetweenandwithinskillgroups.AsnotedbyGottschalk(1997),"[i]nequalityincreasednotonlyamongthosewithdifferentobservabletraits,suchasgender,race,educationandexperience,butalsowithingroupsofworkerswiththesamegender,race,educationandexperience".Buildingonthisevidence,Heathcote,Storesletten,andViolante(2010),amongothers,analysethedynamicsofwage,earningsandincomeinequalitybothbetweenandwithingroups.Asfortheroleplayedbyproductivityslowdownindrivingwealthinequality,3manystudieshavelookedattheeffectofther>gmechanismdescribedbyPiketty(2014)(see,forexample,Jones2014;KrusellandSmith2015;DeNardi,Fella,andYang2016).Tothebestofourknowledge,wearethefirsttoexplorehowthismechanismaffectswealthine-qualitybothbetweenandwithinskillgroups.Ouremphasisonskills(and,therefore,alsoedu-cation)isjustified,sincetheempiricalevidenceshowsthatthedistributionofskillsisanessen-tialpieceofthewealth-inequalitypuzzle(Castaneda,Diaz-Gimenez,andRios-Rull2003;CooperandZhu2016).ThistiesinwiththeobservationbyBartscher,Kuhn,andSchularick(2019)thatthecollegewealthpremiumintheUnitedStateshasincreasedsubstantiallysincethe1980sandismuchlargerthanthecollegeincomepremium.Therearegrowingconcernsthat,ontheonehand,greaterconcentrationofwealthmayleadtobiggerinheritancesastherichesthouseholdspassonfortunestotheirchildrenandthat,ontheotherhand,inheritancefurther

increaseswealthpersistence(Adermon,Lindahl,andWaldenstr?m2018).Existingstudiessug-gestthatinheritedwealthaccountsforbetween15and31percentoftotalhouseholdwealth.4Withinheritedwealthplayinganincreasinglygreaterrolethanself-madewealth,intergenera-tionalsocialmobilitymaywelldecline.Inarecentsurveyofaround1,000individualswhoheld

AsforcefullydemonstratedbyPiketty(2014),wealthinequalityisparticularlyrelevantbecausewealthisevenmoreunequallydistrib-

utedthanincome.

Forareviewoftheliteratureonthistopic,seeKopczukandLupton(2007).

07?ILOWorkingPaper105

overUS$3millionininvestableassets,theBankofAmerica(2023)foundthatonly27percentwereselfmade,whiletheresthadeithersomeinheritedwealthoranaffluentupbringingorboth.Thenegativeeffectsofwealthconcentrationandinheritanceonintergenerationalsocialmobilityareamplifiedwhencombinedwiththeotheradvantagesthatwealthyparentsprovidetheirchildrenwith,suchasinvestingmoreinthemandsendingthemtobetter-qualityschools(Blanden,Gregg,andMacmillan2007;KeaneandWolpin2001;DaruichandKozlowski2020).Wealthconcentrationandinheritancecouldendupcreatinga"glassfloor"belowwhichchildreninwealthyfamiliescannotfall,therebylimitingdownwardmobility.5

Thispaperisstructuredasfollows.Section2presentsanumberofstylizedfactsregardingwealthinequality.Section3discussedthemodel,whilesection4explainsthecalibration.Section5dis-cussestheresults.Additionalanalysisconsideringchangesinincomeriskovertimeispresentedinsection6.Finally,section7offerssomeconclusionsandpolicyimplications.

ForfurtheranalysisofupwardanddownwardintergenerationalmobilityintheUnitedStates,seeLangotetal.(2023).

08?ILOWorkingPaper105

?2StylizedFacts

TheSurveyofConsumerFinances(SCF)hasbeenmeasuringthewealthofUShouseholdseverythreeyearssince1989.UsingSFCdata,earlierpapershaveanalysedthedistributionofwealthbetweenhouseholdstakingintoaccounthouseholdcharacteristicssuchasthelevelofeducationofthefamilyreferenceperson(seeCooperandZhu2016;KuhnandRios-Rull2016;Pfeffer,Gross,andSchoeni2019).Recently,Kuhn,Schularick,andSteins(2017)haveproposedtheHistoricalSurveyofConsumerFinances(HSCF)whichtracksbackthedistributionofwealthto1949.Bartscher,Kuhn,andSchularick(2019)reportthedistributionofwealthbyskillgroupsusingthissurveysince1956.Wecomplementthisliteraturebydescribingextensivelythedynamicsofwealthine-qualitybothbetweenskillgroups(subsection2.1)andwithinskillgroups(subsection2.2).?

2.1.Inequalitybetweenskillgroups

IntheSFC,householdsareclassifiedintofourgroupsaccordingtothelevelofeducation:"Nohighschooldiploma","Highschooldiploma","Somecollege"and"Collegedegree".Consistentlywithourtheoreticalmodel,weuseatwo-levelclassification(skilledandunskilled)inwhichthethreegroupswithlowereducationalattainment("Nohighschooldiploma","Highschooldiplo-ma"and"Somecollege")arecombinedtomakeupthe"unskilled"group,whilethe"skilled"groupcomprisesthosewith"Collegedegree"education.

Table1showsthattheshareofunskilledhouseholdsinthetotalpopulationhasfallenfrom76.76percentinthefirstwaveofthesurvey(1989)to65.97percentinthe2016wave,reflectingtherisinglevelineducationalattainmentintheUnitedStates.

?Table1:WealthinequalityacrossskillgroupsintheUnitedStates:Shareofunskilledhouseholdsintheto-talpopulationandinthefourquartilesofthewealthdistribution,1989-2016(percentage)

A

p0-25

p25-50

p50-75

p75-100

Shareofunskilledin1989

76.76

86.83

84.68

78.77

56.75

Shareofunskilledin2016

65.97

81.06

76.74

67.11

38.99

Variation(inpercentagepoints)

-10.79

-5.77

-7.93

-11.65

-17.76

Note:Source:Authors'calculationsbasedondatafromtheSurveyofConsumerFinances.

?Table2:WealthinequalityacrossskillgroupsintheUnitedStates:Meanvalueofwealthheldbyhouse-holds,skilledversusunskilled,1989-2016(thousandsof2016USdollars)

AllSkilledUnskilledWealthRatio

Meanvaluein1989

353.33

765.69

228.52

3.35

Meanvaluein2016

689.51

150417

269.29

5.59

Annualgrowthrate

2.51

2.53

0.61

1.91

Note:Source:Authors'calculationsbasedondatafromtheSurveyofConsumerFinances.

KuhnandRios-Rull(2016)reportthedistributionofearnings,incomeandwealthfortheUSeconomyusingthe2013SCFsurveybyeducationlevel(5groups).Table2inPfeffer,Gross,andSchoeni(2019)reportsthemedianandGinicoefficientbyeducationlevel(5groups)for3waves(1989,2007,2016)oftheSCFsurvey.Table14inCooperandZhu(2016)showstheassetallocationforfouredu-

cationlevels.

DetailedstatisticsarereportedinAppendix7.

09ILOWorkingPaper105

Table2documentstheriseinwealthinequalityacrossskills.Themeanvalueofwealthincreasedatarateof2.51percentperyearforallhouseholdsraisingfrom353to689thousandsof2016USdollars.However,thisincreaseinwealthhasbeensharedveryunevenly:theannualrateofwealthgrowthwas2.53percentforskilledhouseholds,comparedwithjust0.61percentforunskilledones.Thewidegapinwealthaccumulationhasmarkedlyincreasedthewealthratiobetweenskilledandunskilledhouseholdsfrom3.35to5.59overthesameperiod.Thismeansthatunskilledhouseholdshavebeenpushedoutofthetopofthewealthdistribution,ascanbeseenintable1.In1989,57percentofwealthyhouseholds(definedasthetopquartile)wereun-skilled,comparedwithjust39percentin2016.Thisdecreaseofroughly18percentagepointsismuchlargerthanthedecreaseintheshareofunskilledhouseholdsinthetotalpopulation(11percentagepoints)orthedecreaseintheshareofunskilledhouseholdsinthefirstquartileofthewealthdistribution(6percentagepoints).

2.2.Inequalitywithinskillgroups

WenowlookattheseconddimensionofinequalitythatcanbeanalysedusingtheinformationcontainedintheSCF:within-skillinequality.Table3comparesthreecommonindicatorsofine-quality(thewealthGinicoefficientforskilledandunskilled,theinterquartileratioandtheshareofwealthpossessedbythetopquartileinthedistribution)forallhouseholdsandthenbyskilllevel,in1989and2016.Thistableshowstwostrikingfacts:first,wealthinequalityisalmostiden-ticalatthegloballevelandattheskilllevels-exceptfortheinterquartileratiowhichsuggestsamoreunequalwealthdistributionamongunskilledhouseholdsthanamongskilledhouse-holds.Second,wealthinequalityhasincreasedbothoverallandineachoftheskillgroups.For

example,theincreasesintheshareofwealthpossessedbythetopquartilebetween1989and

2016