Journal of Economic Perspectives: Vol. 27 No. 3 (Summer 2013)

The Top 1 Percent in International and Historical Perspective

Article Citation

Alvaredo, Facundo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez. 2013. "The Top 1 Percent in International and Historical Perspective." Journal of Economic Perspectives, 27(3): 3­20.

DOI: 10.1257/jep.27.3.3

Abstract

The top 1 percent income share has more than doubled in the United States over the last 30 years, drawing much public attention in recent years. While other English­speaking countries have also experienced sharp increases in the top 1 percent income share, many high­income countries such as Japan, France, or Germany have seen much less increase in top income shares. Hence, the explanation cannot rely solely on forces common to advanced countries, such as the impact of new technologies and globalization on the supply and demand for skills. Moreover, the explanations have to accommodate the falls in top income shares earlier in the twentieth century experienced in virtually all high­income countries. We highlight four main factors. The first is the impact of tax policy, which has varied over time and differs across countries. Top tax rates have moved in the opposite direction from top income shares. The effects of top rate cuts can operate in conjunction with other mechanisms. The second factor is a richer view of the labor market, where we contrast the standard supply­side model with one where pay is determined by bargaining and the reactions to top rate cuts may lead simply to a redistribution of surplus. Indeed, top rate cuts may lead managerial energies to be diverted to increasing their remuneration at the expense of enterprise growth and employment. The third factor is capital income. Overall, private wealth (relative to income) has followed a U­shaped path over time, particularly in Europe, where inherited wealth is, in Europe if not in the United States, making a return. The fourth, little investigated, element is the correlation between earned income and capital income, which has substantially increased in recent decades in the United States.

Article Full­Text Access

Full­text Article (Complimentary)

Additional Materials

Author Disclosure Statement(s) (83.25 KB)

Authors

Alvaredo, Facundo (U Oxford and CONICET, Buenos Aires) Atkinson, Anthony B. (Nuffield College, U Oxford and London School of Economics and Political Science) Piketty, Thomas (Paris School of Economics) Saez, Emmanuel (U CA, Berkeley)

JEL Classifications

D31: Personal Income, Wealth, and Their Distributions F66: Globalization: Labor J24: Human Capital; Skills; Occupational Choice; Labor Productivity J31: Wage Level and Structure; Wage Differentials

Comments

View Comments on This Article (0) | Login to post a comment

| A E A | J o u r n a l s | A n n u a l M e e t i n g s | E c o n L i t | J O E | R F E | M e m b e r s | Journal of Economic Perspectives—Volume 27, Number 3—Summer 2013—Pages 3–20

The Top 1 Percent in International and Historical Perspective†

Facundo Alvaredo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez

oorr tthreehree ddecades,ecades, tthehe ddebateebate aaboutbout risingrising incomeincome inequalityinequality inin thethe UnitedUnited SStatestates hashas ccenteredentered onon thethe ddispersionispersion ofof wageswages andand thethe increasedincreased premiumpremium F fforor sskilled/educatedkilled/educated workers,workers, attributedattributed inin varyingvarying proportionsproportions toto skill-skill- bbiasediased ttechnologicalechnological changechange andand toto globalizationglobalization (for(for example,example, seesee KatzKatz andand AAutorutor 11999999 fforor a survey).survey). InIn recentrecent years,years, however,however, therethere hashas beenbeen a growinggrowing realizationrealization tthathat mostmost ofof thethe actionaction hashas beenbeen atat thethe veryvery top.top. ThisThis hashas attractedattracted a greatgreat dealdeal ofof ppublicublic attentionattention (as(as witnessedwitnessed byby thethe numbernumber ofof visitsvisits ttoo aandnd ppressress ccitationsitations ooff oourur WWorldorld TopTop IncomesIncomes DatabaseDatabase atat http://topincomes.parisschoolofeconomics.eu/)http://topincomes.parisschoolofeconomics.eu/) aandnd hashas representedrepresented a challengechallenge toto thethe economicseconomics profession.profession. StoriesStories basedbased onon thethe ssupplyupply andand demanddemand forfor skillsskills areare notnot enoughenough toto explainexplain thethe extremeextreme toptop tailtail ofof tthe earningshe earnings ddistribution;istribution; nornor iiss iitt eenoughnough ttoo llookook oonlynly aatt eearnedarned iincomes.ncomes. DDifferentifferent aapproachespproaches areare necessarynecessary toto explainexplain whatwhat hashas happenedhappened inin thethe UnitedUnited StatesStates overover tthehe pastpast centurycentury andand alsoalso toto explainexplain thethe differingdiffering experienceexperience inin otherother high-incomehigh-income ccountriesountries overover recentrecent decades.decades. WeWe beginbegin withwith thethe internationalinternational comparisoncomparison inin thethe fi rrstst ssectionection andand thenthen turnturn toto thethe causescauses andand implicationsimplications ofof thethe evolutionevolution ofof toptop iincome shares.ncome shares.

■ Facundo Alvaredo is Research Fellow at Nuffi eld College and Department of Economics, Oxford, United Kingdom, and CONICET (Consejo Nacional de Investigaciones Científi cas y Técnicas), Buenos Aires, Argentina, and Affi liate Member, Paris School of Economics, Paris, France. Anthony B. Atkinson is Fellow of Nuffi eld College, Oxford, and Centennial Professor at the London School of Economics, London, United Kingdom. Thomas Piketty is Professor of Economics at the Paris School of Economics, Paris, France. Emmanuel Saez is Professor of Economics, Univer- sity of California at Berkeley, United States. Their email addresses are [email protected], tony.atkinson@nuffi eld.ox.ac.uk, [email protected], and [email protected], respectively. † To access the disclosure statements, visit http://dx.doi.org/10.1257/jep.27.3.3 doi=10.1257/jep.27.3.3 4 Journal of Economic Perspectives

Figure 1 Top 1 Percent Income Share in the United States

25%

20%

15%

10%

5% Top 1% income share excluding capital gains Top 1% income share including capital gains

0% 1913 1920 1927 1934 1941 1948 1955 1962 1969 1976 1983 1990 1997 2004 2011

Source: Source is Piketty and Saez (2003) and the World Top Incomes Database. Notes: The fi gure reports the share of total income earned by top 1 percent families in the United States from 1913 to 2011. Income is defi ned as pre-tax market income; it excludes government transfers and nontaxable fringe benefi ts. The fi gure reports series including realized capital gains (solid squares) and series excluding realized capital gains (hollow squares).

WWee shouldshould sstarttart bbyy eemphasizingmphasizing tthehe ffactualactual iimportancemportance ooff tthehe ttopop 11 percent. percent. IItt isis temptingtempting ttoo ddismissismiss tthehe sstudytudy ooff tthishis ggrouproup aass a ppassingassing ppoliticalolitical ffadad ddueue ttoo tthehe slogansslogans ofof thethe OccupyOccupy movementmovement oror asas thethe academicacademic equivalentequivalent ofof realityreality TTV.V. ButBut thethe magnitudesmagnitudes aarere ttrulyruly ssubstantial.ubstantial. BBasedased oonn ppre-taxre-tax aandnd ppre-transferre-transfer mmarketarket iincomencome ((excludingexcluding nnontaxableontaxable ffringeringe bbenefienefi ttss ssuchuch aass healthhealth insuranceinsurance bbutut includingincluding realizedrealized ccapitalapital ggains)ains) pperer ffamilyamily rreportedeported oonn ttaxax rreturns,eturns, tthehe ssharehare ooff totaltotal annualannual iincomencome rreceivedeceived bbyy tthehe ttopop 11 percent percent hhasas mmoreore tthanhan ddoubledoubled ffromrom 99 percent percent inin 19761976 toto 20 percent20 percent inin 20112011 (Piketty(Piketty andand Saez,Saez, 2003,2003, andand thethe WorldWorld TTopop IncomesIncomes Database).Database). ThereThere havehave beenbeen risesrises forfor otherother toptop shares,shares, butbut thesethese hhaveave bbeeneen mmuchuch ssmaller:maller: dduringuring tthehe ssameame pperiod,eriod, tthehe ssharehare ooff tthehe ggrouproup ffromrom 995th5th toto 99th percentile99th percentile rroseose oonlynly bbyy 33 percentage percentage ppoints.oints. TThehe rriseise iinn tthehe ssharehare ooff tthehe toptop 1 percent1 percent hhasas hhadad a nnoticeableoticeable eeffectffect oonn ooverallverall iincomencome iinequalitynequality iinn tthehe UUnitednited StatesStates ((Atkinson,Atkinson, PPiketty,iketty, aandnd SSaezaez 22011,011, SSection 2.2).ection 2.2).

The United States Top 1 Percent in International Perspective

FFigureigure 1 ddepictsepicts tthehe UUSS ttopop 1 ppercentercent iincomencome ssharehare ssinceince 11913.913. SSimonimon KKuznetsuznets ((1955)1955) ffamouslyamously hhypothesizedypothesized tthathat eeconomicconomic ggrowthrowth wwouldould fi rstrst bebe accompaniedaccompanied byby a riserise inin inequalityinequality andand thenthen byby a declinedecline inin inequality.inequality. AtAt fi rstrst glance,glance, itit isis temptingtempting Facundo Alvaredo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez 5

ttoo cconcludeonclude ffromrom FFigureigure 1 tthathat tthehe KKuznetsuznets ccurveurve hhasas bbeeneen tturnedurned uupside-down.pside-down. BButut tthishis ssuggestionuggestion isis tootoo facile.facile. AfterAfter all,all, thethe interwarinterwar periodperiod diddid notnot exhibitexhibit a secularsecular ddownwardownward ttrendrend iinn sshareshares ooff ttopop iincomes.ncomes. AApartpart ffromrom tthehe bbubbleubble ooff tthehe llateate 11920s,920s, tthehe USUS toptop 1 percent1 percent shareshare waswas betweenbetween 1515 andand 20 percent20 percent throughoutthroughout thisthis time.time. AAtt thethe timetime ofof PearlPearl HarborHarbor inin 1941,1941, thethe shareshare ofof thethe toptop 1 percent1 percent waswas essentiallyessentially tthehe ssameame aass iinn 11918.918. TThehe ddownwardownward ttrendrend iinn ttopop sshareshares sstartedtarted aatt tthehe ttimeime ooff WWorldorld WWar IIar II aandnd ccontinuedontinued untiluntil thethe endend ofof thethe 1960s.1960s. ThereThere waswas thenthen a sharpsharp reversalreversal ssuchuch thatthat thethe toptop shareshare isis ttodayoday bbackack iinn tthehe samesame rrangeange asas iinn tthehe 11920s.920s. Interest-Interest- iingly,ngly, tthehe GGreatreat RecessionRecession ofof 20082008 –2009–2009 doesdoes notnot seemseem toto havehave reversedreversed thethe upwardupward ttrend.rend. ThereThere waswas a fallfall inin thethe toptop 1 percent1 percent shareshare inin 20082008 –2009–2009 butbut a reboundrebound inin 22010.010. TThishis wwouldould bbee cconsistentonsistent wwithith tthehe eexperiencexperience ooff tthehe ppreviousrevious eeconomicconomic ddown-own- tturn:urn: ttopop iincomencome sshareshares ffellell iinn 22001–2002001–2002 bbutut qquicklyuickly rrecoveredecovered aandnd rreturnedeturned ttoo tthehe ppreviousrevious trendtrend inin 20032003 –2007.–2007. AnotherAnother piecepiece ofof evidenceevidence thatthat isis consistentconsistent withwith thisthis iinterpretationnterpretation isis thethe smallersmaller cyclicalcyclical variationvariation inin thethe seriesseries eexcludingxcluding capitalcapital gainsgains ((shownshown bbyy tthehe hhollowollow ssquaresquares inin Figure 1).Figure 1). HHasas thethe USUS experienceexperience beenbeen reproducedreproduced inin otherother high-incomehigh-income countries?countries? TThehe evolutionevolution ofof thethe sharesshares ofof thethe toptop 1 percent1 percent isis shownshown forfor fourfour Anglo-SaxonAnglo-Saxon ccountriesountries inin Figure 2AFigure 2A andand forfor France,France, Germany,Germany, Sweden,Sweden, andand JapanJapan inin Figure 2BFigure 2B ((itit shouldshould bebe notednoted thatthat thethe estimatesestimates forfor FranceFrance andand thethe UnitedUnited KingdomKingdom dodo nnotot includeinclude capitalcapital gains,gains, thethe estimatesestimates forfor Canada,Canada, Germany,Germany, Japan,Japan, andand SwedenSweden iincludenclude realizedrealized capitalcapital gainsgains afterafter thethe yearyear thereintherein shown,shown, andand thethe estimatesestimates fforor AustraliaAustralia includeinclude themthem onlyonly partiallypartially andand atat varyingvarying degreesdegrees overover time).time). TheThe ootherther Anglo-SaxonAnglo-Saxon countries—Australia,countries—Australia, Canada,Canada, andand thethe UnitedUnited Kingdom—allKingdom—all sshowhow a strongstrong asymmetricasymmetric U-shape.U-shape. However,However, thethe risesrises werewere lessless markedmarked inin twotwo ooff thesethese countries.countries. OverOver thethe periodperiod 19801980 toto 2007,2007, whenwhen thethe toptop 1 percent1 percent shareshare rroseose byby somesome 135 percent135 percent inin thethe UnitedUnited StatesStates andand thethe UnitedUnited Kingdom,Kingdom, itit roserose bbyy somesome 105 percent105 percent inin AustraliaAustralia andand 76 percent76 percent inin CanadaCanada (and(and byby 39 percent39 percent iinn NewNew Zealand,Zealand, notnot shown).shown). TheThe experienceexperience isis markedlymarkedly differentdifferent inin continentalcontinental EEuropeurope andand Japan,Japan, wherewhere thethe longlong patternpattern ofof incomeincome inequalityinequality isis muchmuch closercloser toto aann L-shapedL-shaped thanthan a U-shapedU-shaped curve.curve. (Sweden(Sweden andand otherother ScandinavianScandinavian countriescountries ssuchuch asas NorwayNorway (not(not shown)shown) areare intermediateintermediate cases.)cases.)1 TTherehere hashas beenbeen somesome riserise inin rrecentecent yearsyears inin thethe toptop sharesshares inin thesethese countries,countries, butbut thethe toptop 1 percent1 percent sharesshares areare nnotot farfar todaytoday fromfrom theirtheir levelslevels inin thethe latelate 1940s,1940s, whereaswhereas inin thethe UnitedUnited StatesStates thethe ssharehare ofof tthehe ttopop 1 ppercentercent iiss hhigherigher bbyy mmoreore tthanhan aa half. half. TToo uus,s, thethe factfact thatthat high-incomehigh-income countriescountries withwith similarsimilar technologicaltechnological andand pproductivityroductivity developmentsdevelopments havehave gonegone throughthrough differentdifferent patternspatterns ofof iincomencome iinequalitynequality atat thethe veryvery toptop supportssupports thethe viewview thatthat institutionalinstitutional andand policypolicy differencesdifferences pplaylay a keykey rolerole inin thesethese transformations.transformations. PurelyPurely technologicaltechnological storiesstories basedbased solelysolely uuponpon ssupplyupply aandnd ddemandemand ooff sskillskills ccanan hhardlyardly eexplainxplain ssuchuch ddivergingiverging ppatterns.atterns. WWhathat iiss more,more, withinwithin countries,countries, wewe havehave toto explainexplain notnot onlyonly whywhy toptop sharesshares roserose (in(in thethe UU-shaped-shaped countries)countries) butbut alsoalso whywhy theythey fellfell forfor a sustainedsustained periodperiod ofof timetime earlierearlier inin

1 The Swedish top 1 percent share was very high during World War I. The same is observed in Denmark— see the discussion in Atkinson and Søgaard (2013). 6 Journal of Economic Perspectives

Figure 2 The Evolution of the Shares of the Top 1 Percent in Different Countries

A: Top 1 Percent Income Shares in English-speaking Countries (U-Shape) 25% United States—including capital gains Australia Canada-including capital gains from 1972 20% United Kingdom—families United Kingdom—adults

15%

10%

5%

0% 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

B: Top 1 Percent Income Shares in Continental Europe and Japan (L-Shape) 25% France Germany—including capital gains from 1950 Japan—including capital gains from 1947 20% Sweden—including capital gains

15%

10%

5%

0% 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Source: The World Top Incomes Database. Notes: The fi gure reports the share of total income earned by the top 1 percent in four English-speaking countries in panel A, and in four other OECD countries ( Japan and three continental European countries) in panel B. Income is defi ned as pre-tax market income. The estimates for Australia include realized capital gains partially and at varying degrees over time. The Top 1 Percent in International and Historical Perspective 7

Figure 3 Top Marginal Income Tax Rates, 1900 – 2011

100%

90%

80%

70%

60%

50%

40%

30% US UK 20% France Germany 10%

0% 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

Source: Piketty and Saez (2013, fi gure 1). Notes: The fi gure depicts the top marginal individual income tax rate in the United States, United Kingdom, France, and Germany since 1900. The tax rate includes only the top statutory individual income tax rate applying to ordinary income with no tax preference. State income taxes are not included in the case of the United States. For France, we include both the progressive individual income tax and the flat rate tax “Contribution Sociale Generalisée.” tthehe twentiethtwentieth century.century. TheThe mostmost obviousobvious policypolicy difference—betweendifference—between countriescountries andand ooverver ttime—regardsime—regards ttaxation,axation, aandnd iitt iiss hhereere tthathat wwee bbegin.egin.

Taxes and Top Shares DDuringuring thethe twentiethtwentieth century,century, toptop incomeincome taxtax ratesrates havehave followedfollowed anan inverseinverse UU-shaped-shaped time-pathtime-path inin manymany ccountries,ountries, asas illustratedillustrated inin Figure Figure 3.3. InIn tthehe UUnitednited SStates,tates, toptop incomeincome taxtax ratesrates werewere consistentlyconsistently aboveabove 60 percent60 percent fromfrom 19321932 toto 1981,1981, aandnd aatt tthehe sstarttart ooff tthehe 11920s,920s, ttheyhey wwereere aabovebove 770 percent0 percent ((ofof ccourse,ourse, vvaryingarying ppropor-ropor- ttionsions ofof taxpayerstaxpayers werewere subjectsubject toto thethe toptop rate).rate). HighHigh incomeincome taxtax ratesrates areare notnot justjust a featurefeature ofof thethe post-Worldpost-World WarWar IIII period,period, andand theirtheir cumulativecumulative effecteffect contributedcontributed ttoo thethe earlierearlier declinedecline inin toptop incomeincome shares.shares. WhileWhile manymany countriescountries havehave cutcut toptop ttaxax rratesates iinn rrecentecent decades,decades, thethe depthdepth ofof thesethese cutscuts hhasas vvariedaried considerably.considerably. FForor eexample,xample, thethe toptop taxtax raterate inin FranceFrance inin 20102010 waswas onlyonly 10 percentage10 percentage pointspoints lowerlower tthanhan iinn 11950,950, wwhereashereas tthehe ttopop ttaxax rrateate iinn tthehe UUSS wwasas llessess thanthan halfhalf itsits 19501950 value.value. FFigure igure 4 plotsplots thethe changeschanges inin toptop marginalmarginal incomeincome taxtax ratesrates (combining(combining bbothoth centralcentral andand locallocal governmentgovernment incomeincome taxes)taxes) sincesince thethe earlyearly 1960s1960s againstagainst tthehe changeschanges overover thatthat periodperiod inin toptop 1 percent1 percent incomeincome sharesshares forfor 18 high-income18 high-income ccountriesountries inin thethe WorldWorld TopTop IncomesIncomes Database.Database. ItIt showsshows thatthat therethere isis a strongstrong corre-corre- llationation betweenbetween thethe reductionsreductions inin toptop taxtax ratesrates andand thethe increasesincreases inin toptop 1 percent1 percent 8 Journal of Economic Perspectives

Figure 4 Changes in Top Income Shares and Top Marginal Income Tax Rates since 1960 (combining both central and local government income taxes)

10 US

8 Elasticity = .47 (.11)

6 UK

Ireland

4 Norway Portugal Canada

Italy 2 Australia

NZ Spain Denmark Japan Sweden 0

France Germany Finland Change in top 1% income share (percentage points)

Netherlands Switzerland –40 –30 –20 –10 0 10 Change in top marginal income tax rate (percentage points)

Source: Piketty, Saez, and Stantcheva (2011, revised October 2012, fi gure 3). Source for top income shares is the World Top Incomes Database. Source for top income tax rates is OECD and country- specifi c sources. Notes: The fi gure depicts the change in the top 1 percent income share against the change in the top income tax rate from 1960– 64 to 2005–2009 for 18 OECD countries. If the country does not have top income share data for those years, we select the fi rst available fi ve years after 1960 and the most recent 5 years. For the following fi ve countries, the data start after 1960: Denmark (1980), Ireland (1975), Italy (1974), Portugal (1976), Spain (1981). For Switzerland, the data end in 1995 (they end in 2005 or after for all the other countries). Top tax rates include both the central and local government top tax rates. The correlation between those changes is very strong. The elasticity estimates of the ordinary least squares regression of log(top 1% share) on log(1 – MTR) based on the depicted dots is 0.47 (0.11). ppre-taxre-tax incomeincome shares.shares. ForFor example,example, thethe UnitedUnited StatesStates experiencedexperienced a reductionreduction ooff 447 7 percentagepercentage pointspoints inin itsits toptop incomeincome taxtax raterate andand a 10 10 percentagepercentage pointpoint iincreasencrease inin itsits toptop 1 1 percentpercent pre-taxpre-tax incomeincome share.share. ByBy contrast,contrast, countriescountries suchsuch aass GGermany,ermany, Spain,Spain, oror Switzerland,Switzerland, whichwhich diddid notnot experienceexperience anyany signifisignifi cantcant toptop rrateate taxtax cut,cut, diddid notnot showshow increasesincreases inin toptop 1 percent1 percent incomeincome shares.shares. Hence,Hence, thethe eevolutionvolution ofof toptop ttaxax rratesates iiss sstronglytrongly nnegativelyegatively ccorrelatedorrelated wwithith cchangeshanges iinn ppre-taxre-tax iincome concentration.ncome concentration. TThishis nnegativeegative correlationcorrelation cancan bebe explainedexplained inin a varietyvariety ofof ways.ways. AsAs pointedpointed outout ooriginallyriginally byby SlemrodSlemrod (1996),(1996), itit isis possiblepossible thatthat thethe riserise inin toptop USUS incomeincome sharesshares ooccurredccurred because,because, whenwhen toptop taxtax ratesrates declined,declined, thosethose withwith highhigh incomesincomes hadhad lessless Facundo Alvaredo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez 9

rreasoneason ttoo sseekeek ooutut ttaxax aavoidancevoidance sstrategies.trategies. TThishis aargumentrgument hhasas mmoreore rrecentlyecently bbeeneen uusedsed toto denydeny thatthat anyany realreal increaseincrease inin incomeincome concentrationconcentration actuallyactually tooktook place—place— tthathat iitt iiss a ppureure sstatisticaltatistical aartifact.rtifact. UUndernder tthishis sscenario,cenario, thethe realreal USUS toptop incomeincome sharesshares wwereere asas highhigh inin thethe 1960s1960s asas theythey areare today,today, butbut a smallersmaller fractionfraction ofof toptop incomesincomes wwasas rreportedeported onon taxtax rreturns.eturns. WhileWhile tthishis ffactoractor mmayay hhaveave aaffectedffected tthehe ppatternattern ooff tthehe ddataata atat certaincertain times—fortimes—for example,example, thethe jumpjump inin toptop USUS incomeincome sharesshares followingfollowing tthehe 19861986 TaxTax ReformReform Act—closerAct—closer examinationexamination ofof thethe USUS casecase suggestssuggests thatthat thethe taxtax aavoidancevoidance responseresponse cannotcannot accountaccount forfor a signifisignifi cantcant fractionfraction ofof thethe long-runlong-run surgesurge iinn toptop incomes.incomes. TopTop incomeincome sharesshares basedbased onon a broaderbroader defidefi nitionnition ofof incomeincome thatthat iincludesncludes realizedrealized capitalcapital gains,gains, andand hencehence a majormajor portionportion ofof avoidanceavoidance channels,channels, hhaveave iincreasedncreased virtuallyvirtually asas muchmuch asas ttopop iincomencome sshareshares bbasedased oonn a nnarrowerarrower ddefiefi ni-ni- ttionion ofof incomeincome subjectsubject toto thethe progressiveprogressive taxtax scheduleschedule (see(see Figure 1Figure 1 andand Piketty,Piketty, SSaez,aez, aandnd SStantchevatantcheva 22011011 fforor a ddetailedetailed aanalysis).nalysis). TThehe explanationexplanation thatthat changeschanges inin taxtax ratesrates inin thethe toptop taxtax bracketsbrackets dodo leadlead toto ssubstantiveubstantive behavioralbehavioral changechange hashas indeedindeed receivedreceived somesome support.support. AfterAfter notingnoting thatthat ttopop UUSS iincomesncomes ssurgedurged followingfollowing thethe largelarge toptop marginalmarginal taxtax raterate cutscuts ofof thethe 1980s,1980s, LLindseyindsey ((1987)1987) aandnd FFeldsteineldstein ((1995)1995) pproposedroposed a sstandardtandard ssupply-sideupply-side sstorytory wwherebyhereby llowerower ttaxax rratesates sstimulatetimulate eeconomicconomic aactivityctivity aamongmong ttopop eearnersarners iinvolvingnvolving mmoreore wwork,ork, ggreaterreater entrepreneurship,entrepreneurship, andand thethe like.like. InIn thisthis scenario,scenario, lowerlower toptop taxtax ratesrates wouldwould lleadead ttoo mmoreore eeconomicconomic aactivityctivity bbyy tthehe rrichich aandnd hhenceence mmoreore eeconomicconomic ggrowth.rowth. BBehavioralehavioral changechange isis atat tthehe hhearteart ooff tthehe ooptimalptimal iincomencome ttaxax aanalysisnalysis ppioneeredioneered bbyy MirrleesMirrlees (1971)(1971) andand publiclypublicly evokedevoked inin thethe debatedebate aboutabout toptop taxtax ratesrates inin thethe UUK,K, wwherehere thethe ChancellorChancellor ofof thethe ExchequerExchequer hashas arguedargued thatthat reducingreducing thethe toptop taxtax rrateate bbelowelow 50 50 percentpercent (for(for broadlybroadly thethe toptop 1 1 percent)percent) willwill notnot reducereduce revenue.revenue. TThehe sstandardtandard optimaloptimal ttaxax fformulaormula ((DiamondDiamond aandnd SSaezaez 22011)011) iimplies,mplies, wwithith aann eelas-las- tticityicity ooff ttaxableaxable iincomencome ooff 00.5,.5, tthathat tthehe rrevenue-maximizingevenue-maximizing ttopop ttaxax rrateate wwouldould bbee 557 percent.7 percent.2 WhenWhen allowanceallowance isis mademade forfor otherother taxestaxes leviedlevied inin thethe UnitedUnited Kingdom,Kingdom, ssuchuch aass tthehe ppayrollayroll ttax,ax, tthishis iimpliesmplies a ttopop iincomencome ttaxax rrateate iinn tthehe UUnitednited KKingdomingdom ooff ssomeome 440 percent0 percent ((AtkinsonAtkinson 22012).012).

Richer Models of Pay Determination TThehe optimaloptimal taxtax literatureliterature has,has, however,however, remainedremained rootedrooted inin anan oversimpli-oversimpli- fi eded modelmodel ofof paypay determinationdetermination thatthat takestakes nono accountaccount ofof developmentsdevelopments inin laborlabor eeconomics,conomics, andand thethe samesame aappliespplies ttoo tthehe eexplanationsxplanations ofof changingchanging toptop incomeincome sshares.hares. ChangesChanges inin thethe paypay ofof a workerworker areare assumedassumed toto havehave nono impactimpact onon eithereither tthehe otherother sideside ofof thethe llaborabor mmarketarket oorr oonn ootherther wworkers.orkers. TThehe wworkerorker ggeneratesenerates mmoreore ooutpututput andand paypay adjustsadjusts byby thethe samesame amount.amount. EachEach personperson isis anan island.island. However,However, inin tthehe nnow-standardow-standard modelsmodels ofof job-matching,job-matching, a jobjob emergesemerges asas thethe resultresult ofof thethe costlycostly ccreationreation ofof a vacancyvacancy byby thethe employeremployer andand ofof jobjob searchsearch byby thethe employee.employee. A matchmatch

2 The revenue-maximizing top tax rate formula takes the form τ = 1/(1 + a · e) where a is the Pareto parameter of the top tail of the income distribution, and e is the elasticity of pre-tax income with respect to the net-of-tax rate 1 – τ. With e = 0.5 (as estimated from Figure 4) and a = 1.5 (the current Pareto parameter of the US income distribution), we get τ = 1/(1 + 0.5 · 1.5) = 57 percent. 10 Journal of Economic Perspectives

ccreatesreates a positivepositive surplus,surplus, andand therethere isis NashNash bargainingbargaining overover thethe divisiondivision ofof thethe ssurplus,urplus, lleadingeading toto a pproportionroportion β goinggoing toto thethe workerworker andand (1(1 – β ) toto thethe employer.employer. TTypically,ypically, β iiss aassumedssumed fi xed,xed, butbut itit isis possiblepossible thatthat whatwhat wewe havehave observed,observed, atat leastleast atat tthehe top,top, iiss aann iincreasencrease iinn β, wwhichhich cancan leadlead toto changeschanges inin thethe distributiondistribution ofof income.income.3 WWhyhy sshouldhould β hhaveave increased?increased? TheThe extentextent toto whichwhich toptop earnersearners exercisedexercised bbargainingargaining powerpower maymay havehave interactedinteracted withwith thethe changeschanges inin thethe taxtax system.system. WhenWhen ttopop marginalmarginal taxtax ratesrates werewere veryvery high,high, thethe netnet rewardreward toto a highlyhighly paidpaid executiveexecutive forfor bbargainingargaining forfor moremore compensationcompensation waswas modest.modest. WhenWhen toptop marginalmarginal taxtax ratesrates fell,fell, hhighigh earnersearners startedstarted bargainingbargaining moremore aggressivelyaggressively toto increaseincrease theirtheir compensation.compensation. IInn tthishis sscenario,cenario, cutscuts inin toptop taxtax ratesrates cancan increaseincrease toptop incomeincome shares—consistentshares—consistent wwithith thethe observedobserved trendtrend iinn Figure 1—butFigure 1—but thethe increasesincreases inin toptop 1 percent1 percent incomesincomes nnowow ccomeome aatt tthehe eexpensexpense ooff tthehe rremainingemaining 999 percent.9 percent. OOnene cancan alsoalso weaveweave thisthis notionnotion ofof greatergreater incentivesincentives forfor bargainingbargaining intointo a bbroaderroader scenario,scenario, inin whichwhich thethe improvedimproved informationinformation andand communicationscommunications tech-tech- nnologyology aandnd gglobalizationlobalization werewere iincreasingncreasing thethe demanddemand fforor hhigh-skilledigh-skilled llabor,abor, aandnd tthehe dderegulationeregulation ooff fi nancenance andand ofof otherother industriesindustries waswas bothboth raisingraising thethe demanddemand forfor skillskill aatt tthehe ttopop aandnd cchanginghanging tthehe rulesrules underunder whichwhich compensationcompensation hadhad beenbeen calculatedcalculated iinn tthehe ppast.ast. IInn tthishis pperspective,erspective, hhighigh mmarginalarginal ttaxax rratesates hhadad sservederved aass a bbrakerake oonn tthehe llevelevel ooff ssurplusurplus eextractionxtraction inin thethe past,past, butbut thenthen thisthis brakebrake waswas releasedreleased atat thethe samesame ttimeime tthathat eeconomicconomic aandnd iinstitutionalnstitutional conditionsconditions allowedallowed forfor higherhigher compensationcompensation aatt tthehe ttopop ooff tthehe iincomencome ddistributionistribution ((Piketty,Piketty, SSaez,aez, aandnd SStantchevatantcheva 22011).011). IInn thisthis scenario,scenario, thethe higherhigher shareshare ofof incomeincome goinggoing toto thethe toptop 1 percent1 percent doesdoes nnotot reflrefl ectect higherhigher economiceconomic growth—whichgrowth—which isis a keykey differencedifference withwith thethe supply-supply- ssideide scenario.scenario. ItIt isis eveneven possiblepossible thatthat reductionsreductions inin toptop marginalmarginal taxtax ratesrates maymay hhaveave adverseadverse effectseffects onon growth,growth, asas maymay bebe seenseen ifif wewe gogo backback toto thethe theoriestheories ofof mmanagerialanagerial fi rmsrms andand thethe separationseparation ofof ownershipownership andand controlcontrol developeddeveloped byby OOliver E.liver E. Williamson,Williamson, WilliamWilliam Baumol,Baumol, andand RobinRobin MarrisMarris inin thethe 1960s1960s andand 1970s1970s ((forfor discussion,discussion, seesee SolowSolow 1971).1971). InIn thesethese models,models, managersmanagers areare concernedconcerned withwith ttheirheir remunerationremuneration (both(both monetarymonetary andand nonmonetary)nonmonetary) butbut alsoalso withwith otherother dimen-dimen- ssionsions suchsuch asas thethe scalescale oror raterate ofof growthgrowth ofof theirtheir fi rrms,ms, andand allocateallocate theirtheir efforteffort aaccordingly.ccordingly. WhereWhere toptop taxtax ratesrates werewere high,high, therethere waswas a lowlow returnreturn toto efforteffort spentspent oonn negotiatingnegotiating higherhigher pay.pay. TopTop corporatecorporate executivesexecutives maymay havehave concentratedconcentrated onon ssecuringecuring alternativealternative sourcessources ofof utility,utility, suchsuch asas unproductiveunproductive corporatecorporate expenses,expenses, bbutut theythey maymay alsoalso havehave ploughedploughed backback profiprofi tsts intointo securingsecuring fasterfaster expansionexpansion thanthan iinn thethe traditionaltraditional stockstock marketmarket valuation-maximizingvaluation-maximizing fi rrm.m. CutsCuts inin toptop taxtax rates,rates, hhowever,owever, meantmeant thatthat toptop executivesexecutives switchedswitched effortsefforts backback toto securingsecuring a largerlarger shareshare ooff thethe profiprofi tts,s, inin whichwhich casecase increasesincreases inin remuneration,remuneration, oror bonuses,bonuses, maymay havehave comecome aatt thethe expenseexpense ofof employmentemployment and growth.and growth. TThehe ccorrelationorrelation sshownhown iinn FFigure 4igure 4 bbetweenetween ttopop mmarginalarginal ttaxax rratesates aandnd cchangeshanges iinn ttopop iincomencome sshareshares mmayay ooff ccourseourse rreflefl eectct iinn ppartart ccoincidenceoincidence rratherather tthanhan ccausality.ausality. TThehe ppoliticalolitical ffactorsactors tthathat lleded ttoo ttopop ttaxax rrateate ccutsuts ——suchsuch aass tthosehose bbyy RReaganeagan aandnd TThatcherhatcher

3 Kleven, Landais, Saez, and Schultz (2013) fi nd evidence of such bargaining effects in the pay determi- nation of high earners, using the Danish preferential tax scheme for highly paid immigrants. The Top 1 Percent in International and Historical Perspective 11

iinn tthehe 11980s980s iinn tthehe UUnitednited SStatestates aandnd tthehe UUnitednited KKingdom—wereingdom—were aaccompaniedccompanied bbyy ootherther llegislativeegislative cchanges,hanges, ssuchuch aass dderegulation,eregulation, wwhichhich mmayay hhaveave ccausedaused ttopop iincomesncomes ttoo rrise,ise, nnotot lleasteast oonn aaccountccount ooff tthehe iimpetusmpetus ttheyhey ggaveave ttoo tthehe ggrowthrowth ooff tthehe fi nancialnancial sserviceservices ((PhilipponPhilippon aandnd RReshefeshef 22012)012) aandnd llegalegal sserviceservices ssectors.ectors. MMoreore ggenerally,enerally, tthehe eeffectsffects ooff ttaxationaxation mmayay iinteractnteract wwithith ootherther cchanges,hanges, ssuchuch aass tthosehose iinn rremunerationemuneration ppractices.ractices. WWherehere ttherehere iiss a ssurplusurplus ttoo bbee sshared,hared, tthehe ddivisionivision mmayay rreflefl ectect rrelativeelative bbargainingargaining sstrength,trength, aass aabove,bove, bbutut iitt mmayay aalsolso bbee iinflnfl uuencedenced bbyy ssocialocial nnorms.orms. NNotionsotions ooff ffairness,airness, oorr a ““paypay ccode,”ode,” mmayay ccomeome iintonto pplaylay ttoo rremoveemove tthehe iindeterminacyndeterminacy wwherehere ““individualindividual incentivesincentives areare notnot byby themselvesthemselves . . . suffisuffi cientcient ttoo ddetermineetermine a uuniquenique eequilibrium”quilibrium” ((MacLeodMacLeod aandnd MMalcomsonalcomson 11998,998, pp. 400).. 400). AA “pay “pay ccode”ode” llimitsimits tthehe eextentxtent ttoo wwhichhich eearningsarnings aarere iindividuallyndividually ddetermined,etermined, a ssituationituation tthathat bbothoth wworkersorkers aandnd eemployersmployers aacceptccept oonn rreputationaleputational ggrounds.rounds. AAss aarguedrgued iinn AAtkinsontkinson ((2008),2008), ttherehere mmayay bbee a ttipping-pointipping-point wwherehere ttherehere iiss a sswitchwitch ffromrom a hhighigh llevelevel ooff aadherencedherence ttoo ssuchuch a ccodeode ttoo a ssituationituation wwherehere ppayay bbecomesecomes llargelyargely iindividuallyndividually ddetermined.etermined. TThishis hhasas bbeeneen ddocumentedocumented iinn tthehe ccasease ooff tthehe UUnitednited SStatestates bbyy LLemieux,emieux, MMacLeod,acLeod, aandnd PParentarent ((2009),2009), wwhoho fi ndnd anan increaseincrease inin thethe proportionproportion ofof performance-payperformance-pay jobsjobs overover thethe pperioderiod 11976976 ttoo 1998.1998. AsAs theythey note,note, thethe increasedincreased extentextent ofof performance-payperformance-pay maymay bebe a channelchannel byby whichwhich otherother factorsfactors areare expressedexpressed inin greatergreater wagewage dispersion,dispersion, andand theythey sstresstress tthehe eeffectffect aatt tthehe ttopop eendnd ooff tthehe wwage distribution.age distribution.

Top Tax Rates and Growth IIff wwee llookook aatt tthehe aaggregateggregate ooutcomes,utcomes, wwee fi ndnd nono apparentapparent correlationcorrelation betweenbetween ccutsuts inin toptop taxtax ratesrates andand growthgrowth ratesrates inin realreal perper capitacapita GDPGDP (Piketty,(Piketty, Saez,Saez, andand SStantchevatantcheva 2011).2011). CountriesCountries thatthat mademade llargearge ccutsuts iinn ttopop ttaxax rratesates ssuchuch aass tthehe UUnitednited KKingdomingdom oorr tthehe UUnitednited SStatestates hhaveave nnotot ggrownrown ssignifiignifi cantlycantly fasterfaster thanthan countriescountries thatthat ddidid nnot,ot, ssuchuch aass GGermanyermany oorr SSwitzerland.witzerland. TThishis llackack ooff ccorrelationorrelation iiss mmoreore cconsistentonsistent wwithith a storystory tthathat tthehe rresponseesponse ooff ppre-taxre-tax toptop iincomesncomes ttoo ttopop ttaxax rratesates ddocumentedocumented iinn FFigure 4igure 4 iiss ddueue toto increasedincreased bargainingbargaining powerpower oror moremore individualizedindividualized paypay atat thethe top,top, rratherather thanthan increasedincreased productiveproductive effort.effort. Naturally,Naturally, cross-countrycross-country comparisonscomparisons areare bboundound toto bebe fragile;fragile; exactexact resultsresults varyvary withwith thethe specifispecifi cation,cation, years,years, andand countries.countries. HHowever,owever, tthehe rregressionegression analysisanalysis byby Piketty,Piketty, Saez,Saez, andand StantchevaStantcheva (2011),(2011), usingusing thethe ccompleteomplete ttime-seriesime-series ddataata ssinceince 11960,960, sshowshows tthathat tthehe aabsencebsence ooff ccorrelationorrelation bbetweenetween eeconomicconomic growthgrowth andand toptop taxtax ratesrates isis quitequite robust.robust. ByBy andand large,large, thethe bottombottom lineline isis tthathat rrichich ccountriesountries hhaveave aallll ggrownrown aatt rroughlyoughly tthehe ssameame rrateate ooverver tthehe ppastast 440 years—in0 years—in sspitepite ooff hhugeuge vvariationsariations iinn taxtax ppolicies.olicies. MMoreore sspecifipecifi cally,cally, internationalinternational evidenceevidence showsshows thatthat currentcurrent paypay levelslevels forfor chiefchief eexecutivexecutive offioffi cerscers acrossacross countriescountries areare stronglystrongly negativelynegatively correlatedcorrelated withwith toptop taxtax rratesates eveneven controllingcontrolling forfor fi rm’srm’s characteristicscharacteristics andand performance,performance, andand thatthat thisthis ccorrelationorrelation isis sstrongertronger iinn fi rmsrms withwith poorpoor governancegovernance (Piketty,(Piketty, Saez,Saez, andand StantchevaStantcheva 22011).011).4 ThisThis fi ndingnding alsoalso suggestssuggests thatthat thethe linklink betweenbetween toptop taxtax ratesrates andand paypay ofof chiefchief

4 Governance is measured with an index that combines various governance measures: insider ownership, institutional ownership, the ratio of independent board directors, whether the CEO is also chairman of the board, and the average number of board positions held by board members. 12 Journal of Economic Perspectives

eexecutivexecutive ooffiffi cerscers doesdoes notnot runrun throughthrough fi rmrm performanceperformance butbut isis likelylikely toto bebe duedue toto bbargaining effects.argaining effects. SSuchuch fi ndingsndings havehave strongstrong implicationsimplications forfor toptop taxtax raterate policies.policies. TheThe optimaloptimal ttopop taxtax raterate risesrises dramaticallydramatically ifif a substantialsubstantial fractionfraction ofof thethe effecteffect ofof toptop taxtax ratesrates oonn pre-taxpre-tax toptop incomesincomes documenteddocumented inin Figure 4Figure 4 aboveabove isis duedue toto wage-bargainingwage-bargaining eeffectsffects insteadinstead ofof supply-sidesupply-side effects.effects. UsingUsing mid-rangemid-range parameterparameter valuesvalues wherewhere tthehe rresponseesponse ofof toptop earnersearners toto toptop taxtax raterate cutscuts isis three-fithree-fi fthsfths duedue toto increasedincreased bbargainingargaining behaviorbehavior andand ttwo-fiwo-fi fthsfths duedue toto increasedincreased productiveproductive work,work, Piketty,Piketty, Saez,Saez, aandnd StantchevaStantcheva (2011)(2011) fi ndnd thatthat thethe toptop taxtax raterate couldcould potentiallypotentially bebe setset asas highhigh asas 883 percent—as3 percent—as oopposedpposed ttoo 557 percent7 percent iinn tthehe ppureure ssupply-sideupply-side mmodel.odel.5

Capital Income and Inheritance

TThehe analysisanalysis justjust citedcited focused—likefocused—like muchmuch ofof thethe literature—literature— onon whatwhat isis ccommonlyommonly ccalledalled ““earnedearned iincomes,”ncomes,” rreferringeferring ttoo iincomencome rreceivedeceived iinn rreturneturn fforor wwork.ork. BButut capitalcapital incomeincome isis alsoalso anan importantimportant partpart ofof thethe story.story. OfOf course,course, thethe distinctiondistinction bbetweenetween thethe two typestwo types ofof incomeincome cancan becomebecome blurryblurry inin somesome cases—notably,cases—notably, entre-entre- ppreneurialreneurial incomeincome cancan havehave elementselements ooff bbothoth ccompensationompensation forfor workwork andand a rreturneturn ttoo ccapitalapital iinvestment.nvestment. HHere,ere, wwee ddefiefi nene “capital“capital income”income” asas rents,rents, dividends,dividends, interest,interest, aandnd realizedrealized capitalcapital gains.gains. TheThe declinedecline ofof toptop capitalcapital incomesincomes isis thethe mainmain driverdriver ofof tthehe ffallsalls iinn ttopop iincomencome sshareshares tthathat ooccurredccurred iinn mmanyany ccountriesountries eearlyarly iinn tthehe ttwentiethwentieth ccentury.entury. ForFor example,example, fromfrom 19161916 toto 1939,1939, capitalcapital incomeincome representedrepresented 50 percent50 percent ofof UUSS toptop 1 percent1 percent incomes,incomes, whereaswhereas byby thethe endend ofof thethe centurycentury fromfrom 19871987 toto 2010,2010, tthehe ssharehare hhadad ffallenallen toto one-thirdone-third (Piketty(Piketty andand SaezSaez 2003,2003, tables A7tables A7 andand A8).A8). InIn thethe UUnitednited Kingdom,Kingdom, thethe correspondingcorresponding shareshare fellfell fromfrom 60 percent60 percent inin 19371937 toto underunder 220 0 percentpercent byby thethe endend ofof thethe centurycentury (Atkinson(Atkinson 2007,2007, fi gure 4.11).gure 4.11). AtAt thethe samesame ttime,ime, iitt sshouldhould bbee bborneorne iinn mmindind tthathat tthesehese ccalculationsalculations ddependepend oonn tthehe ddefiefi nitionnition ofof ttaxableaxable incomes.incomes. InIn timestimes ppast,ast, a nnumberumber ooff iincomencome ttaxax systemssystems llikeike thosethose inin FranceFrance aandnd tthehe UUnitednited KKingdomingdom iincludedncluded iimputedmputed rrentsents ooff hhomeownersomeowners iinn tthehe iincomencome ttaxax bbase,ase, bbutut ttodayoday iimputedmputed rrentsents aarere ttypicallyypically eexcluded.xcluded. WhereWhere thethe taxtax basebase hashas beenbeen eextended,xtended, thisthis hashas inin somesome casescases takentaken thethe formform ooff sseparateeparate taxationtaxation (as(as withwith real-real- iizedzed capitalcapital gainsgains inin thethe UnitedUnited Kingdom),Kingdom), soso thatthat thisthis elementelement ofof capitalcapital incomeincome iiss notnot coveredcovered inin thethe incomeincome taxtax data.data. AsAs a resultresult ofof thesethese developments,developments, thethe shareshare ooff capitalcapital incomeincome thatthat isis reportablereportable onon incomeincome taxtax returnsreturns hashas oftenoften signifisignifi cantlycantly ddecreasedecreased oover time.ver time. EEarlierarlier wewe referredreferred toto thethe cumulativecumulative effecteffect ofof progressiveprogressive taxation.taxation. A longlong pperioderiod ofof hhighigh ttopop rratesates ooff iincomencome ttaxation,axation, ccoupledoupled wwithith hhighigh ttopop rratesates ooff ttaxationaxation oonn tthehe ttransmissionransmission ooff wwealthealth bbyy iinheritancenheritance aandnd ggift,ift, rreducededuced tthehe ccapacityapacity ooff llargearge

5 With wage-bargaining effects, the optimal top tax rate formula becomes τ = (1 + s · a · e)/(1 + a · e) where s is the fraction of the total behavioral elasticity due to bargaining effects. With a = 1.5, e = 0.5 (as above), and s = 3/5, we obtain τ = 83 percent. In the standard model with no wage-bargaining effects, we had s = 0 and τ = 57 percent. Facundo Alvaredo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez 13

Figure 5 Annual Inheritance Flow as a Fraction of Disposable Income, France 1820 –2008

40% Economic flow (computed from national wealth estimates, 36% mortality tables and observed age-wealth profiles) Fiscal flow (computed from observed bequest and gift tax data, including tax exempt assets) 32%

28%

24%

20%

16%

12%

8%

4%

0% 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000

Source: Piketty (2011). Notes: The annual inheritance fl ow is defi ned as the total market value of all assets (tangible and fi nancial assets, net of fi nancial liabilities) transmitted at death or through inter vivos gifts. Disposable income was as high as 90–95 percent of national income during the 19th century and early 20th century (when taxes and transfers were almost nonexistent), while it is now about 70 percent of national income.

wwealth-holdersealth-holders toto sustainsustain theirtheir preeminence.preeminence. TheThe keykey ffactoractor iinn ddeterminingetermining tthehe ccapacityapacity toto ttransmitransmit wwealthealth iiss tthehe ddifferenceifference bbetweenetween tthehe ““internalinternal rrateate ooff aaccumula-ccumula- ttion”ion” ((thethe ssavingsavings raterate timestimes tthehe rrateate ooff rreturneturn netnet ofof taxes)taxes) andand thethe raterate ofof growthgrowth ooff thethe economy.economy. TThishis mmeanseans thatthat thethe taxationtaxation ofof incomeincome andand wealthwealth transferstransfers cancan ccauseause tthehe ssharehare ooff ttopop wwealth-holdersealth-holders ttoo ffall,all, aass iinn tthehe UUnitednited KKingdomingdom ooverver tthehe fi rstrst tthree-quartershree-quarters ofof thethe twentiethtwentieth centurycentury ((AtkinsonAtkinson andand HHarrisonarrison 11978),978), ccontributingontributing ttoo thethe downwarddownward trajectorytrajectory ofof toptop incomeincome shares.shares. AlongsideAlongside thisthis waswas thethe growthgrowth ooff ““popularpopular wealth”wealth” ownedowned byby thethe bottombottom 99 percent.99 percent. BackBack inin 19081908 inin thethe UnitedUnited KKingdom,ingdom, thethe 17th17th EarlEarl ofof DerbyDerby hadhad a rentrent rollroll ofof somesome £100,000,£100,000, whichwhich waswas moremore tthanhan 1,000 times1,000 times tthehe aaverageverage incomeincome atat tthehe ttime.ime. MManyany ooff tthesehese hhousesouses aarere nnowow oownedwned bbyy ttheir occupiers.heir occupiers. IInn rrecentecent ddecades,ecades, hhowever,owever, tthehe rrelationelation bbetweenetween tthehe iinternalnternal rrateate ooff aaccumula-ccumula- ttionion ooff wwealthealth hholdingsoldings aandnd tthehe rrateate ooff ggrowthrowth ooff ccapitalapital hhasas nnowow bbeeneen rreversedeversed aass a rresultesult ofof thethe cutscuts inin capitalcapital taxationtaxation andand thethe declinedecline inin thethe macroeconomicmacroeconomic growthgrowth rrateate (Piketty(Piketty 2011).2011). AsAs a result,result, a numbernumber ofof countriescountries areare witnessingwitnessing a returnreturn ofof iinheritancenheritance asas a majormajor factor.factor. Figure 5Figure 5 showsshows thethe estimatesestimates ofof PikettyPiketty (2011)(2011) forfor FFrancerance forfor thethe periodperiod 18201820 toto 20082008 ofof thethe annualannual inheritanceinheritance fl owow (the(the amountamount ppassedassed onon throughthrough bequestsbequests andand giftsgifts inter vivos)),, eexpressedxpressed asas percentagepercentage ofof 14 Journal of Economic Perspectives

ddisposableisposable income.income.6 TTwo methodswo methods areare employed:employed: a constructiveconstructive calculationcalculation fromfrom nnationalational wealthwealth fi gures,gures, mortalitymortality rates,rates, andand observedobserved age-wealthage-wealth profiprofi les,les, andand anan eestimatestimate basedbased onon thethe estateestate andand giftgift taxtax records.records. TheThe twotwo methodsmethods differdiffer inin levelslevels ((thethe fi scalscal fl owsows areare lower),lower), butbut thethe time-pathstime-paths areare veryvery similar.similar. TThehe iinheritancenheritance fl ooww iinn FFrancerance wwasas relativelyrelatively stablestable aroundaround 20–25 percent20–25 percent ooff disposabledisposable incomeincome throughoutthroughout thethe 18201820 –1910–1910 periodperiod (with(with a slightslight upwardupward ttrend),rend), beforebefore bbeingeing ddividedivided bbyy a ffactoractor ooff aaboutbout 5 ttoo 6 bbetweenetween 11910910 aandnd tthehe 11950s.950s. SSinceince tthen,hen, iitt hhasas bbeeneen rrisingising rregularly,egularly, wwithith aann aaccelerationcceleration ooff tthehe ttrendrend dduringuring thethe pastpast 30 years.30 years. TheseThese ttrulyruly eenormousnormous hhistoricalistorical vvariationsariations bbringring FFrancerance bbackack toto a situationsituation similarsimilar toto thatthat ofof 100 years100 years ago.ago. AnAn annualannual inheritanceinheritance fl owow aaroundround 20 percent20 percent ofof disposabledisposable incomeincome isis veryvery large.large. ItIt isis typicallytypically muchmuch largerlarger tthanhan thethe annualannual fl owow ofof newnew ssavingsavings aandnd aalmostlmost aass bbigig aass tthehe aannualnnual fl owow o off ccapitalapital income.income. TThishis iimpliesmplies tthathat iinheritancenheritance iiss aagaingain bbecomingecoming a vveryery iimportantmportant ffactoractor ofof lifetimelifetime economiceconomic inequality.inequality. AsAs shownshown iinn PikettyPiketty aandnd SSaezaez ((2012),2012), iinn a wworldorld wherewhere inheritanceinheritance iiss qquantitativelyuantitatively ssignifiignifi cant,cant, thosethose receivingreceiving nono bequestsbequests wwillill lleaveeave ssmaller-than-averagemaller-than-average bbequestsequests tthemselveshemselves aandnd hhenceence sshouldhould ssupportupport sshiftinghifting laborlabor taxationtaxation towardtoward bequestbequest taxation.taxation. InIn thisthis situation,situation, inheritanceinheritance taxa-taxa- ttionion (and(and moremore generallygenerally capitalcapital taxation,taxation, ggiveniven ccapitalapital mmarketarket iimperfections)mperfections) bbecomesecomes a powerfulpowerful andand desirabledesirable tooltool forfor redistributionredistribution towardtoward thosethose receivingreceiving nno inheritance.o inheritance. TThehe rreturneturn ooff iinheritednherited wwealthealth mmayay wwellell ddifferiffer iinn mmagnitudeagnitude aacrosscross ccoun-oun- ttries.ries. TheThe historicalhistorical seriesseries availableavailable soso farfar regardingregarding thethe inheritanceinheritance fl owsows aarere ttoooo scarcescarce toto reachreach fi rmrm conclusions.conclusions. EExistingxisting eestimatesstimates ssuggestuggest tthathat tthehe FFrenchrench UU-shaped-shaped ppatternattern aalsolso aappliespplies ttoo GGermany,ermany, aandnd ttoo a llesseresser eextentxtent ttoo tthehe UUnitednited KKingdomingdom aandnd tthehe UUnitednited SStatestates ((AtkinsonAtkinson 22013;013; SSchinkechinke 22012;012; sseeee PPikettyiketty aandnd ZZucman,ucman, forthcoming,forthcoming, forfor a survey).survey). SuchSuch variationsvariations couldcould bebe duedue toto differencesdifferences iinn pensionpension systemssystems andand thethe shareshare ofof privateprivate wealthwealth thatthat isis annuitizedannuitized (and(and there-there- fforeore nontransmissible).nontransmissible). FFromrom a theoreticaltheoretical perspective,perspective, iitt isis unclearunclear howeverhowever whywhy ttherehere shouldshould bebe muchmuch crowdingcrowding outout betweenbetween lifecyclelifecycle wealthwealth andand transmissibletransmissible wwealthealth inin anan oopenpen eeconomyconomy (that(that iis,s, tthehe ffactact tthathat iindividualsndividuals ssaveave mmoreore fforor ttheirheir ppensionension sshouldhould nnotot mmakeake tthemhem ssaveave llessess fforor ttheirheir cchildren;hildren; tthehe eextraxtra ppensionension wwealthealth comingcoming ffromrom tthehe llifecycleifecycle mmotiveotive sshouldhould bbee iinvestednvested aabroad).broad). IItt ccouldould bbee tthathat ttherehere aarere ddifferencesifferences iinn ttastesastes fforor wwealthealth ttransmission.ransmission. MMaybeaybe wwealthyealthy iindividualsndividuals inin thethe UnitedUnited KingdomKingdom andand inin thethe UnitedUnited StatesStates havehave lessless tastetaste forfor bbequestequest tthanhan ttheirheir FFrenchrench aandnd GGermanerman ccounterparts.ounterparts. HHoweverowever iitt sshouldhould bbee kkeptept iinn mmindind tthathat ttherehere aarere iimportantmportant ddataata pproblemsroblems ((inin pparticular,articular, wwealthealth ssurveysurveys ttendend ttoo vastlyvastly underestimateunderestimate inheritanceinheritance receipts),receipts), whichwhich couldcould partlypartly explainexplain whywhy thethe rriseise ofof inheritanceinheritance fl owsows inin thethe recentrecent periodperiod appearsappears toto bebe moremore limitedlimited inin somesome

6 It is critical to include both bequests (wealth transmitted at death) and gifts (wealth transmitted inter vivos) in our defi nition of inheritance, fi rst because gifts have always represented a large fraction of total wealth transmission, and second because this fraction has changed a lot over time. The Top 1 Percent in International and Historical Perspective 15

Figure 6 Private Wealth/National Income Ratios, 1870 – 2010

800%

United States 700% Europe

600%

500%

400%

300%

200%

100% 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

Source: Piketty and Zucman (2013). Notes: Europe is the (unweighted) average of France, Germany, and the United Kingdom. Private wealth is defi ned as the sum of nonfi nancial assets, fi nancial assets, minus fi nancial liabilities in the household and nonprofi t sectors.

ccountriesountries tthanhan iinn oothers.thers.7 AAnothernother ssourceource ooff ddifferenceifference bbetweenetween ccountriesountries ccouldould ccomeome fromfrom vvariationsariations iinn tthehe ttotalotal mmagnitudeagnitude ooff wwealthealth aaccumulation.ccumulation. TTherehere mmayay iinn tthishis respectrespect bbee aann iimportantmportant ddifferenceifference bbetweenetween tthehe UUnitednited SStatestates aandnd EEurope,urope, aass isis indeedindeed ssuggesteduggested wwhenhen wwee llookook aatt ttotalotal pprivaterivate wwealthealth ((expressedexpressed aass a rratioatio ttoo nnationalational iincome),ncome), sshownhown iinn FFigure 6igure 6 ((seesee PPikettyiketty aandnd ZZucman,ucman, 22013,013, fforor a ddiscus-iscus- ssionion onon thethe ddifferencesifferences bbetweenetween pprivaterivate aandnd nnational wealth).ational wealth). AAss maymay bebe seenseen fromfrom Figure Figure 6,6, thethe twentiethtwentieth centurycentury hashas seenseen a U-shapedU-shaped ttime-pathime-path inin thethe ratioratio ofof privateprivate wealthwealth toto nationalnational incomeincome thatthat isis moremore markedmarked inin EEuropeurope thanthan inin thethe UnitedUnited States.States. PrivatePrivate wealthwealth inin EuropeEurope waswas aroundaround six timessix times

7 In particular, the smaller rise of the UK inheritance fl ow (as compared to France and Germany) is entirely due to the much smaller rise of recorded inter vivos gifts, which according to fi scal data barely rose in the United Kingdom during recent decades, while they have become almost as large as bequests in France and Germany. This might simply be due to the fact that gifts are not properly recorded by the UK tax adminis- tration (Atkinson 2013). In the United States, due to the limitations of federal fi scal data on bequests and gifts, scholars often use retrospective wealth survey data. The problem is that in countries with exhaustive administrative data on bequests and gifts (such as France, and to some extent Germany), survey-based self-reported fl ows appear to be less than 50 percent of fi scal fl ows. This probably contributes to explaining the low level of inheritance receipts found in a number of US studies. An example of such a study is Wolff and Gittleman (2011); one additional bias in this study is that inherited assets are valued using asset prices at the time these assets were transmitted, and no capital gain or income is included. 16 Journal of Economic Perspectives

nnationalational incomeincome inin 1910,1910, andand thenthen fellfell afterafter thethe WorldWorld WarsWars toto lessless thanthan twotwo andand a halfhalf timestimes inin 1950.1950. InIn thethe pastpast 60 years,60 years, itit hashas risenrisen sharplysharply toto reachreach moremore thanthan fi ve timesve times nationalnational income.income. ThisThis patternpattern suggestssuggests thatthat capitalcapital isis “back”“back” andand thatthat thethe llowow wwealth–incomeealth–income ratiosratios observedobserved inin EuropeEurope fromfrom thethe 1950s1950s ttoo tthehe 11970s970s werewere aann anomaly.anomaly. TThishis cancan bebe wellwell accountedaccounted forfor byby thethe long-runlong-run wealthwealth accumulationaccumulation fformulaormula β = s/g, wwherehere β iiss tthehe HHarrod–Domar–Solowarrod–Domar–Solow wealth/incomewealth/income ratio,ratio, s iiss tthehe ssavingaving rate,rate, andand g isis thethe growthgrowth raterate includingincluding bothboth realreal perper capitacapita andand populationpopulation ggrowth.rowth. ForFor a givengiven savingsaving raterate (say(say s = 1010 percent),percent), youyou accumulateaccumulate a lotlot moremore wwealthealth relativerelative toto incomeincome inin thethe longlong runrun whenwhen thethe growthgrowth raterate isis 1.51.5 toto 2 percent2 percent tthanhan ifif thethe growthgrowth raterate isis 2.52.5 toto 3 percent.3 percent. GivenGiven thethe largelarge andand continuingcontinuing differ-differ- eencence iinn ppopulationopulation growthgrowth ratesrates betweenbetween OldOld EuropeEurope andand thethe NewNew World,World, thisthis cancan eexplainxplain notnot onlyonly thethe long-runlong-run changeschanges butbut alsoalso thethe differencedifference inin levelslevels betweenbetween EEuropeurope aandnd tthehe UUnitednited SStatestates ((PikettyPiketty 22011;011; PPikettyiketty andand ZucmanZucman 22013).013).8 OOnn thethe otherother hand,hand, itit shouldshould bebe notednoted thatthat wealth concentration ((asas oopposedpposed toto wwealthealth aaccumulation)ccumulation) isis ssignifiignifi cantlycantly greatergreater inin thethe UnitedUnited States,States, wherewhere thethe toptop 11 percent percent ownsowns aboutabout 35 percent35 percent ofof aggregateaggregate wealthwealth (for(for comparison,comparison, thethe shareshare isis aaboutbout 2020 –25–25 ppercentercent iinn EEurope).urope). SSoo ffar,ar, eexistingxisting sstudiestudies hhaveave ffoundound tthathat tthehe iincreasencrease iinn UUSS wwealthealth concentrationconcentration sincesince thethe 1970s1970s andand 1980s1980s hashas beenbeen relativelyrelatively moderatemoderate iinn ccontrastontrast ttoo tthehe hhugeuge iincreasencrease iinn UUSS iincomencome cconcentrationoncentration ddocumentedocumented aabovebove ((KennickellKennickell 2009;2009; KopczukKopczuk andand SaezSaez 2004).2004). However,However, wewe shouldshould bebe modestmodest aboutabout oourur aabilitybility ttoo mmeasureeasure tthehe ttrendsrends iinn ttopop bbillionaireillionaire wwealth.ealth. WWithith llowow aandnd ddiminishingiminishing ggrowthrowth ratesrates andand highhigh globalglobal returnsreturns toto capital,capital, thethe potentialpotential forfor divergencedivergence ofof thethe wwealthealth ddistributionistribution iiss nnaturallyaturally quitequite large.large.

Joint Distribution of Earned and Capital Income WWee hhaveave discusseddiscussed earnedearned incomeincome andand capitalcapital income.income. TheThe lastlast piecepiece ofof thethe ppuzzleuzzle concernsconcerns thethe joint ddistributionistribution ofof earnedearned andand capitalcapital incomes—anincomes—an aspectaspect thatthat iiss rarelyrarely givengiven explicitexplicit consideration.consideration. YetYet itit isis importantimportant toto knowknow whetherwhether thethe samesame ppeopleeople areare aatt tthehe ttopop ooff bbothoth tthehe ddistributionistribution ooff ccapitalapital iincomencome aandnd tthehe ddistributionistribution ooff eearnedarned income.income. SupposeSuppose thatthat wewe imagineimagine askingasking thethe populationpopulation fi rstrst toto lineline upup aalonglong oneone sideside ofof a roomroom inin increasingincreasing orderorder ofof theirtheir earnedearned incomeincome andand thenthen toto ggoo toto thethe otherother sideside ofof thethe roomroom andand lineline upup inin increasingincreasing orderorder ofof theirtheir capitalcapital iincome.ncome. HHowow mmuchuch wwillill ttheyhey ccrossross oover?ver? IInn tthehe RRicardianicardian cclasslass mmodel,odel, tthehe ccrossingrossing iiss ccomplete:omplete: thethe ccapitalistsapitalists ccomeome aatt tthehe ttopop iinn oonene ccasease aandnd aatt tthehe bbottomottom iinn tthehe oother.ther. HHasas a nnegativeegative correlationcorrelation inin thethe nineteenthnineteenth centurycentury beenbeen replacedreplaced todaytoday byby a zerozero ccorrelation?orrelation? OrOr isis therethere a pperfecterfect ccorrelation,orrelation, soso thatthat peoplepeople crosscross straightstraight over?over? TheThe ppatternattern ofof ccrossingrossing isis ggiveniven bbyy tthehe ccopula,opula, wwhichhich representsrepresents thethe jointjoint distributiondistribution iinn termsterms ofof a functionfunction ofof thethe ranksranks inin thethe twotwo distributionsdistributions ofof earningsearnings andand capitalcapital iincome.ncome. BBecauseecause tthehe ccopulaopula ccomparesompares rranks,anks, iitt iiss nnotot aaffectedffected bbyy wwhetherhether tthehe ddistri-istri- bbutionsutions tthemselveshemselves aarere wwideningidening oorr nnarrowing.arrowing.

8 In a way, this is equivalent to the explanation based upon lower bequest taste: with higher population growth and the same bequest taste (per children), the United States should save more. However a signifi cant part of US population growth historically comes from migration, so this interpretation is not fully accurate. Facundo Alvaredo, Anthony B. Atkinson, Thomas Piketty, and Emmanuel Saez 17

Table 1 Relation between Top Labor Incomes and Top Capital Incomes in the United States

Year

1980 2000

A: Percent of top 1% capital incomes in various top labor income groups Labor income groups: Top 1% 17% 27% Top 5% 27% 45% Top 10% 32% 52% Top 20% 38% 61%

B: Percent of top 1% labor incomes in various top capital income groups Capital income groups: Top 1% 17% 27% Top 5% 36% 50% Top 10% 47% 63% Top 20% 68% 80%

Source: Aaberge, Atkinson, Königs, and Lakner (forthcoming). Notes: Panel A reports the percent of top 1 percent capital income earners in various top labor income groups in 1980 (column 1) and 2000 (column 2). In 2000, 27 percent of top 1 percent capital income earners were also in the top 1 percent of labor incomes, 45 percent were in the top 5 percent of labor incomes, etc. Panel B reports the percent of top 1 percent labor income earners in various top capital income groups in 2000 (column 1) and 1980 (column 2). The computations are based on the public use US tax return micro-datafi les (see Aaberge et al., forthcoming, for complete details).

WWhathat ccanan bbee learnedlearned byby consideringconsidering thethe copula?copula? TableTable 1 showsshows resultsresults forfor thethe UUnitednited StatesStates inin 20002000 andand inin 19801980 basedbased onon taxtax returnreturn datadata analysisanalysis fromfrom Aaberge,Aaberge, AAtkinson,tkinson, Königs,Königs, andand LaknerLakner (forthcoming).(forthcoming). ThreeThree conclusionsconclusions maymay bebe drawn.drawn. FFirst,irst, thethe jointjoint distributiondistribution isis asymmetric.asymmetric. InIn 2000,2000, ofof thosethose inin thethe toptop 1 percent1 percent ofof ccapitalapital income,income, 61 percent61 percent werewere inin thethe toptop 20 percent20 percent ofof earnedearned income.income. However,However, tturningurning thingsthings round,round, ofof thosethose inin thethe toptop 1 1 percentpercent ofof earnedearned income,income, a largerlarger pproportionroportion ofof 80 80 percentpercent werewere inin thethe toptop 20 20 percentpercent ofof capitalcapital income.income. InIn fact,fact, 663 percent3 percent ofof thethe toptop 1 percent1 percent ofof earnersearners werewere inin thethe toptop 10 percentpercent ofof capitalcapital iincome.ncome. SuchSuch asymmetryasymmetry couldcould easilyeasily bebe missedmissed byby thethe useuse ofof a measuremeasure suchsuch asas tthehe correlationcorrelation coefficoeffi cientcient oror a parametricparametric formform forfor thethe copulacopula function.function. Second,Second, tthehe ddegreeegree ooff aassociationssociation appearsappears strong.strong. EvenEven forfor capitalcapital income,income, overover halfhalf ofof thethe ttopop 11 percent percent fi ndnd themselvesthemselves inin thethe toptop tenthtenth ofof earners.earners. A quarterquarter areare inin thethe toptop 11 percent percent forfor both.both. Third,Third, thethe numbersnumbers forfor 19801980 areare allall smallersmaller thanthan theirtheir coun-coun- tterpartserparts forfor 2000.2000. TheThe degreedegree ofof associationassociation increasedincreased betweenbetween 19801980 andand 2000:2000: inin 11980980 onlyonly 17 percent17 percent werewere inin thethe toptop 1 percent1 percent forfor both.both. TheThe proportionproportion ofof thethe toptop 11 percent percent ofof earnersearners whowho werewere inin thethe toptop 5 percent5 percent ofof capitalcapital incomeincome roserose fromfrom one-one- tthirdhird ttoo oone-half,ne-half, aandnd tthehe rreverseeverse pproportionroportion roserose fromfrom 2277 ttoo 445 percent.5 percent. 18 Journal of Economic Perspectives

TToo understandunderstand thethe changingchanging relationshiprelationship betweenbetween earnedearned andand capitalcapital incomes,incomes, wwee needneed toto considerconsider thethe mechanismsmechanisms thatthat linklink thethe twotwo sources.sources. InIn oneone direction,direction, ttherehere isis thethe accumulationaccumulation ofof wealthwealth outout ofof earnedearned income.income. HereHere thethe opportuni-opportuni- ttiesies hhaveave changedchanged inin Anglo-SaxonAnglo-Saxon countries.countries. A thirdthird ofof a centurycentury ago,ago, KayKay andand KKinging (1980,(1980, p. 59)p. 59) describeddescribed thethe hypotheticalhypothetical positionposition ofof a seniorsenior executiveexecutive withwith a largelarge corporationcorporation inin thethe UnitedUnited KingdomKingdom whowho hadhad savedsaved a quarterquarter ofof hishis after-after- ttaxax eearnings:arnings: “[F]eeling“[F]eeling . . . thatthat hehe hashas beenbeen unusuallyunusually fortunatefortunate inin hishis careercareer andand uunusuallynusually thriftythrifty . . . hehe maymay bbee ssomewhatomewhat ssurprisedurprised ttoo ddiscoveriscover tthathat ttherehere aarere iinn BBritainritain atat leastleast 100,000 people100,000 people richerricher thanthan hehe is.”is.” Today,Today, a chiefchief executiveexecutive offioffi cercer mmayay bbee bbothoth bbetteretter ppaidaid aandnd mmoreore aableble ttoo aaccumulate.ccumulate. IInn tthehe ootherther ddirection,irection, ttherehere iiss tthehe eeffectffect ooff llargearge ffamilyamily wwealthealth onon earnings.earnings. InIn thethe past,past, thethe linklink maymay havehave beenbeen nnegative,egative, whereaswhereas todaytoday itit mmayay bbee ssociallyocially uunacceptablenacceptable toto livelive purelypurely offoff unearnedunearned iincome.ncome. WWealth/familyealth/family connectionsconnections maymay provideprovide accessaccess toto high-payinghigh-paying employmentemployment ((toto assessassess this,this, itit isis necessarynecessary toto investigateinvestigate thethe cross-generationcross-generation correlationcorrelation ofof allall iincome,ncome, nnotot jjust earnings).ust earnings).

Conclusions

TThehe rriseise iinn ttopop iincomencome sshareshares iinn tthehe UUnitednited SStatestates hhasas bbeeneen ddramatic.ramatic. IInn sseekingeeking eexplanations,xplanations, hhowever,owever, iitt wwouldould bbee mmisleadingisleading ttoo ffocusocus jjustust oonn tthehe ddoublingoubling ooff tthehe ssharehare ooff iincomencome ggoingoing ttoo tthehe ttopop 11 p percentercent ooff tthehe UUSS ddistributionistribution ooverver tthehe ppastast 440 years.0 years. WWee aalsolso hhaveave ttoo aaccountccount fforor tthehe ffactact tthathat a nnumberumber ooff hhigh-incomeigh-income ccoun-oun- ttriesries hhaveave sseeneen mmoreore mmodestodest oorr llittleittle iincreasencrease iinn ttopop sshares.hares. HHence,ence, tthehe eexplanationxplanation ccannotannot rrelyely ssolelyolely oonn fforcesorces ccommonommon ttoo aadvanceddvanced ccountries,ountries, llikeike tthehe iimpactmpact ooff nnewew ttechnologiesechnologies aandnd gglobalizationlobalization oonn tthehe ssupplyupply aandnd ddemandemand fforor sskills.kills. MMoreover,oreover, tthehe eexplanationsxplanations hhaveave ttoo aaccommodateccommodate tthehe ffallsalls iinn ttopop iincomencome sshareshares eearlierarlier iinn tthehe ttwen-wen- ttiethieth ccenturyentury tthathat ccharacterizeharacterize tthehe ccountriesountries ddiscussediscussed hhere.ere. IInn thisthis paper,paper, wewe havehave highlightedhighlighted fourfour mainmain factorsfactors thatthat havehave contributedcontributed toto tthehe ggrowingrowing incomeincome sharesshares atat thethe veryvery toptop ofof thethe incomeincome distribution,distribution, notingnoting thatthat ttheyhey maymay ooperateperate toto differingdiffering extentsextents inin thethe UnitedUnited StatesStates andand otherother countries,countries, pparticularlyarticularly inin ccontinentalontinental Europe.Europe. TThehe fi rstrst isis taxtax policy:policy: toptop taxtax ratesrates havehave movedmoved inin tthehe ooppositepposite directiondirection fromfrom ttopop ppre-taxre-tax iincomencome sshares.hares. TThehe ssecondecond ffactoractor iiss a rrichericher vviewiew ofof thethe laborlabor market,market, wherewhere wewe havehave contrastedcontrasted thethe standardstandard supply-sidesupply-side modelmodel wwithith thethe alternativealternative possibilitypossibility thatthat therethere maymay havehave beenbeen changeschanges toto bargainingbargaining ppowerower andand greatergreater individualizationindividualization ofof pay.pay. TaxTax cutscuts maymay havehave ledled managerialmanagerial ener-ener- ggiesies toto bebe diverteddiverted toto increasingincreasing theirtheir remunerationremuneration atat thethe expenseexpense ofof enterpriseenterprise ggrowthrowth aandnd eemployment.mployment. TThehe tthirdhird ffactoractor iiss ccapitalapital iincome.ncome. IInn EEurope—buturope—but llessess ssoo iinn tthehe UUnitednited SStates—privatetates—private wwealthealth ((relativerelative ttoo nnationalational iincome)ncome) hhasas ffollowedollowed a sspectacularpectacular U-shapedU-shaped pathpath overover time,time, aandnd iinheritednherited wwealthealth mmayay bbee mmakingaking a rreturn,eturn, iimplyingmplying thatthat inheritanceinheritance andand capitalcapital incomeincome taxationtaxation willwill becomebecome againagain centralcentral ppolicyolicy toolstools forfor curbingcurbing inequality.inequality. TheThe fi nal,nal, little-investigated,little-investigated, elementelement isis thethe ccorrelationorrelation betweenbetween earnedearned incomeincome andand capitalcapital income,income, whichwhich havehave becomebecome moremore ccloselylosely aassociatedssociated iinn tthehe UUnitednited SStates.tates. The Top 1 Percent in International and Historical Perspective 19

■ We are grateful to the journal’s Editor, David Autor, the Managing Editor, Timothy Taylor, and Coeditors, Chang-Tai Hseih and Ulrike Malmendier, for most helpful comments. Financial support from the MacArthur Foundation, the Center for Equitable Growth at UC Berkeley, the Institute for New Economic Thinking at the Oxford Martin School, and the ESRC-DFID Joint Fund is thankfully acknowledged.

References

Aaberge, Rolf, Anthony B. Atkinson, Sebastian 2013. “The Long-Run History of Income Inequality Königs, and Christoph Lakner. Forthcoming. in Denmark: Top Incomes from 1870 to 2010.” “Wages, Capital and Top Incomes.” Unpublished EPRU working paper 2013-01, Economic Policy paper, not completed. Research Unit, University of Copenhagen. Alvaredo, Facundo, Anthony B. Atkinson, Diamond, Peter, and Emmanuel Saez. 2011 Thomas Piketty, and Emmanuel Saez. 2011. The “The Case for a Progressive Tax: From Basic World Top Incomes Database, online at http:// Research to Policy Recommendations.” Journal of topincomes.g-mond.parisschoolofeconomics.eu/, Economic Perspectives 25(4): 165 – 90. May 15, 2013. Feldstein, Martin. 1995. “The Effect of Marginal Atkinson, Anthony B. 2007. “The Distribution of Tax Rates on Taxable Income: A Panel Study of the Top Incomes in the United Kingdom 1908–2000.” 1986 Tax Reform Act.” Journal of Political Economy In Top Incomes over the Twentieth Century—A Contrast 103(3): 551–72. between Continental European and English-Speaking Katz, Lawrence F., and David H. Autor. 1999. Countries, edited by A. B. Atkinson and T. Piketty, “Changes in the Wage Structure and Earnings 82 –140. Oxford University Press. Inequality.” Chap. 26 in Handbook of Labor Atkinson, Anthony B. 2008. The Changing Economics, edited by O. C. Ashenfelter and D. Card, Distribution of Earnings in OECD Countries. Oxford Vol. 3A. Amsterdam: North-Holland. University Press. Kay, John A., and Mervyn A. King. 1980. The Atkinson, Anthony B. 2012. “The Mirrlees British Tax System. Oxford University Press. Review and the State of Public Economics.” Journal Kennickell, Arthur B. 2009. “Ponds and of Economic Literature 50(3): 770 – 80. Streams: Wealth and Income in the U.S., 1989 to Atkinson, Anthony B. 2013. “Wealth and 2007.” Finance and Economics Discussion Series Inheritance in Britain from 1896 to the Present.” 2009 -13, Federal Reserve Board, Washington, DC. Unpublished paper. Kleven, Henrik Jacobsen, Camille Landais, Atkinson, Anthony B., and Allan J. Harrison. Emmanuel Saez, and Esben Schultz. 2013. “Migra- 1978. Distribution of Personal Wealth in Britain. tion and Wage Effects of Taxing Top Earners: Cambridge University Press. Evidence from the Foreigners’ Tax Scheme in Atkinson, Anthony B., and Thomas Piketty. Denmark.” NBER Working Paper 18885. 2007. Top Incomes over the Twentieth Century—A Kopczuk, Wojciech, and Emmanuel Saez. Contrast between Continental European and English- 2004. “Top Wealth Shares in the United States, Speaking Countries. Oxford University Press. 1916 –2000: Evidence from Estate Tax Returns.” Atkinson, Anthony B., and Thomas Piketty. National Tax Journal 57(2, part 2): 445 – 88. 2010. Top Incomes: A Global Perspective. Oxford Kuznets, Simon. 1955. “Economic Growth and University Press. Income Inequality.” American Economic Review Atkinson, Anthony B., Thomas Piketty, and 45(1): 1– 28. Emmanuel Saez. 2011. “Top Incomes in the Long Lemieux, Thomas, W. Bentley MacLeod, and Run of History.” Journal of Economic Literature Daniel Parent. 2009. “Performance Pay and Wage 49(1): 3 –71. Inequality.” Quarterly Journal of Economics 124(1): Atkinson, Anthony B., and Jakob E. Søgaard. 1– 49. 20 Journal of Economic Perspectives

Lindsey, Lawrence B. 1987. “Individual Taxpayer Piketty, Thomas, Emmanuel Saez, and Stefanie Response to Tax Cuts: 1982–84: With Implications Stantcheva. 2011. “Optimal Taxation of Top Labor for the Revenue Maximizing Tax Rate.” Journal of Incomes: A Tale of Three Elasticities.” NBER Public Economics 33(2): 173 – 206. Working Paper 17616. (Forthcoming in American MacLeod, W. Bentley, and Malcomson, James Economic Journal: Economic Policy.) M. 1998. “Motivation and Markets.” American Piketty, Thomas, and Gabriel Zucman. 2013. Economic Review 88(3): 388 – 411. “Capital is Back: Wealth-Income Ratios in Rich Mirrlees, James A. 1971. “An Exploration in the Countries, 1700–2010.” http://piketty.pse.ens.fr Theory of Optimum Income Taxation.” Review of /fi les/PikettyZucman2013WP.pdf. Economic Studies 38(2): 175 – 208. Piketty, Thomas, and Gabriel Zucman. Philippon, Thomas, and Ariell Reshef. 2012. Forthcoming. “Wealth and Inheritance in the “Wages and Human Capital in the U.S. Finance Long Run.” In Handbook of Income Distribution, Industry: 1909 –2006.” Quarterly Journal of Economics Vol. 2, edited by A. Atkinson and F. Bourguignon. 127(4): 1551–1609. Elsevier-North Holland. Piketty, Thomas. 2011. “On the Long-run Schinke, Christoph. 2012. “Inheritance in Evolution of Inheritance: France 1820–2050.” Germany 1911 to 2009: A Mortality Multiplier Quarterly Journal of Economics 126(3): 1071–1131. Approach.” DIW-Berlin SOEP Paper 462. Piketty, Thomas, and Emmanuel Saez. Slemrod, Joel. 1996. “High Income Families 2003. “Income Inequality in the United States, and the Tax Changes of the 1980s: The Anatomy 1913–1998.” Quarterly Journal of Economics 118(1): of Behavioral Response.” In Empirical Foundations 1–39. (Series updated to 2011 in January 2013; of Household Taxation, edited by Martin Feldstein available at Saez’s website: http://elsa.berkeley.edu and James M. Poterba, 169 – 92. University of /~saez/.) Chicago. Piketty, Thomas, and Emmanuel Saez. 2012. Solow, Robert. 1971. “Some Implications of “A Theory of Optimal Inheritance Taxation.” Alternative Criteria for the Firm.” In The Corporate CEPR Discussion Paper 9241. (Forthcoming in Economy, edited by R. Marris and A. Wood, 318 – 42. Econometrica.) Macmillan. Piketty, Thomas, and Emmanuel Saez. 2013. Wolff, Edward N., and Maury Gittleman. 2011. “Optimal Labor Income Taxation.” Chap. 9 in “Inheritances and the Distribution of Wealth or Handbook of Public Economics, Vol. 5, edited by A. Whatever Happened to the Great Inheritance Auerbach, R. Chetty, M. Feldstein, and E. Saez. Boom?” BLS Working Paper 445, US Bureau of Elsevier-North Holland. Labor Statistics.