Seventy-five years of measuring income inequality in Latin America

Al a edo, Facundo; Bou guignon, F ançois; Fe ei a, F ancisco; Lus ig, No a Wo king Pape Se en y- i e yea s o measu ing income inequali y in La in Ame ica IDB Wo king Pape Se ies, No. IDB-WP-01521 P o ided in Coope a ion wi h: In e -Ame ican De elopmen Bank (IDB), Washing on, DC Sugges ed Ci a ion: Al a edo, Facundo; Bou guignon, F ançois; Fe ei a, F ancisco; Lus ig, No a (2023) : Se en y- i e yea s o measu ing income inequali y in La in Ame ica, IDB Wo king Pape Se ies, No. IDB-WP-01521, In e -Ame ican De elopmen Bank (IDB), Washing on, DC, h ps://doi.o g/10.18235/0005211 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/299499 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h p://c ea i ecommons.o g/licenses/by-nc-nd/3.0/igo/legalcode Facundo Al a edo F ançois Bou guignon F ancisco Fe ei a No a Lus ig IDB WORKING PAPER SERIES No IDB-WP-01521 Se en y- i e Yea s o Measu ing Income Inequali y in La in Ame ica In e -Ame ican De elopmen Bank Oc obe , 2023 Se en y- i e Yea s o Measu ing Income Inequali y in La in Ame ica Facundo Al a edo F ançois Bou guignon F ancisco Fe ei a No a Lus ig In e -Ame ican De elopmen Bank Oc obe , 2023 Ca aloging-in-Publica ion da a p o ided by he In e -Ame ican De elopmen Bank Felipe He e a Lib a y Se en y- i e yea s o measu ing income inequali y in La in Ame ica / Facundo Al a edo, F ançois Bou guignon, F ancisco H.G. Fe ei a, No a Lus ig. p. cm. — (IDB Wo king Pape Se ies ; 1521) Includes bibliog aphical e e ences. 1. Income dis ibu ion-La in Ame ica. 2. Income dis ibu ion-Ca ibbean A ea. 3. Household su eys-La in Ame ica. 4. Household su eys-Ca ibbean A ea. I. Al a edo, Facundo. II. Bou guignon, F ançois. III. Fe ei a, F ancisco H. G. IV. Lus ig, No a. V. In e -Ame ican De elopmen Bank. Vice P esidency o Sec o s and Knowledge. VI. Se ies. IDB-WP-1521 JEL codes: D31, D63, O54 Keywo ds: Inequali y; Equali y; measu emen ; Household Su ey; La in Ame ica and The Ca ibbean; Equali y Indica o ; Income Equali y; Taxa ion; Na ional Accoun h p://www.iadb.o g Copy igh © 2023 In e -Ame ican De elopmen Bank. This wo k is licensed unde a C ea i e Commons IGO 3.0 A ibu ion- NonComme cial-NoDe i a i es (CC-IGO BY-NC-ND 3.0 IGO) license (h p://c ea i ecommons.o g/licenses/by-nc- nd/3.0/igo/legalcode) and may be ep oduced wi h a ibu ion o he IDB and o any non-comme cial pu pose, as p o ided below. No de i a i e wo k is allowed. Any dispu e ela ed o he use o he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he UNCITRAL ules. The use o he IDB’s name o any pu pose o he han o a ibu ion, and he use o IDB’s logo shall be subjec o a sepa a e w i en license ag eemen be ween he IDB and he use and is no au ho ized as pa o his CC-IGO license. Following a pee e iew p ocess, and wi h p e ious w i en consen by he In e -Ame ican De elopmen Bank (IDB), a e ised e sion o his wo k may also be ep oduced in any academic jou nal, including hose indexed by he Ame ican Economic Associa ion’s EconLi , p o ided ha he IDB is c edi ed and ha he au ho (s) ecei e no income om he publica ion. The e o e, he es ic ion o ecei e income om such publica ion shall only ex end o he publica ion’s au ho (s). Wi h ega d o such es ic ion, in case o any inconsis ency be ween he C ea i e Commons IGO 3.0 A ibu ion-NonComme cial-NoDe i a i es license and hese s a emen s, he la e shall p e ail. No e ha link p o ided abo e includes addi ional e ms and condi ions o he license. The opinions exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews o he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun ies hey ep esen . Se en y- i e yea s o measu ing income inequali y in La in Ame ica* Facundo Al a edo, F ançois Bou guignon, F ancisco Fe ei a, No a Lus ig Abs ac : D awing on a comp ehensi e compila ion o quan ile sha es and inequali y measu es o 34 coun ies, including o e 5,600 es ima ed Gini coe icien s, we e iew he measu emen o income inequali y in La in Ame ica and he Ca ibbean o e he las se en decades. Al hough he e idence om he i s qua e cen u y – oughly un il he 1970s – is oo agmen a y and di icul o compa e, clea e pa e ns eme ge o he las i y yea s. The cen al ea u e o hese pa e ns is a b oad in e ed U cu e, wi h inequali y ising in mos coun ies p io o he 1990s, and alling du ing he ea ly 21s Cen u y, a leas un il he mid-2010s, when ends appea o di e ge ac oss coun ies. This b oad pa e n is modi ied by coun y speci ici ies, wi h conside able a ia ion in iming and magni ude. Whe eas his b oad pic u e eme ges o income inequali y dynamics, he e is much mo e unce ain y abou he exac le els o inequali y in he egion. The unce ain y a ises om he dispa i y in es ima es o he same coun y/yea combina ions, depending on whe he hey come om household su eys exclusi ely; om some combina ion o su eys and adminis a i e ax da a; and on whe he hey a emp o scale income agg ega es o achie e consis ency wi h Na ional Accoun s es ima es. Since no single me hod is ully con incing a p esen , we a e le wi h (o en wide) anges, o bands, o inequali y as ou bes summa ies o inequali y le els. Reassu ingly, howe e , he dynamic pa e ns a e gene ally obus ac oss he bands. Keywo ds: income inequali y, measu emen , La in Ame ica and he Ca ibbean JEL classi ica ions : D31, D63, O54 * We a e g a e ul o Leona do Gaspa ini, Mau icio De Rosa, Ignacio Flo es Beale, Ma c Mo gan, Ch is oph Lakne and Nishan Yonzan o p o iding se ies and cla i ica ions behind some o he igu es; and o Leopoldo To na olli o kindly answe ing a numbe o inqui ies conce ning he ecen e olu ion o inequali y in B azil. Gas ón Nie as O idani and Pie o Geuna p o ided excellen esea ch assis ance. We a e also indeb ed o commen s and discussion o Se gio Fi po, he panel membe s o LACIR, and o pa icipan s in he LACIR wo kshops in Washing on, DC (Augus , 2022) and Ca agena (Ma ch, 2023). 2 1. In oduc ion The e a e many ‘windows’ h ough which a coun y’s income dis ibu ion can be obse ed and i s inequali y measu ed.1 Use ul in o ma ion abou people’s incomes can be ob ained om a leas i e sou ces: censuses; na ional accoun s a is ics; household su eys; ax eco ds; and o he adminis a i e da ase s on social secu i y and wages.2 Some imes, esea che s p e e o bypass income al oge he and examine he dis ibu ion o consump ion expendi u es, which come p ima ily om household su eys, al hough no necessa ily he same household su eys ha cap u e incomes. In La in Ame ica, long conside ed one o he wo ld’s mos unequal egions, all o hese di e en ypes o sou ces ha e been used o measu e inequali y. The egion has a ich and long adi ion in he de elopmen and p oduc ion o inequali y and po e y s a is ics. S udies a emp ing o es ima e inequali y and o desc ibe he dis ibu ion o income can be aced back o mo e han se en decades. This pape e iews he s a is ical e idence on income inequali y in La in Ame ica and he Ca ibbean (LAC) o e he las se en y- i e yea s o so.3 We do no collec any da a o p oduce new p ima y se ies ou sel es. Ins ead, we examine a b oad a ay o s a is ics and esea ch indings p oduced in he egion o e ime, as epo ed by s a is ical agencies, in academic s udies, o in o he eposi o ies o p ima y and seconda y sou ces. In so doing – as in any a emp o co e a la ge li e a u e on income inequali y in a la ge egion – we mus con end wi h a conside able he e ogenei y o es ima es and a ange o compa abili y issues, which no only a ec he ex en o which a pa icula measu e o a gi en coun y in a gi en yea can be compa ed ac oss ime o space bu also, mo e wo yingly, gene a e dispa a e es ima es o he same coun y and yea . In ying o make sense o his landscape, ou p ima y objec i e is o summa ize wha can be said wi h con idence abou he le els and dynamics o income inequali y in La in Ame ica, while ecognizing he unce ain y ha a ises om con lic ing in o ma ion om di e en sou ces. We each h ee main conclusions. Fi s , he pic u e a ising om household su eys alone is, b oadly speaking, in e nally consis en , in he sense ha he causes o mos o he di e ences in es ima es can be unde s ood and de i e om di e ences in he cons uc ion o he wel a e indica o o he 1 In an o en-quo ed passage, Z i G iliches w o e: “The a ailable economic s a is ics a e ou main window on economic beha io . In spi e o he sc a ches and pe sis en ogging, we canno s op pee ing h ough i and ying o unde s and wha is happening.” (1986, p.1509) 2 E en his is no an exhaus i e lis . Income da a is also p esen in egis ies, p i a e company da abases, ich lis s, and so on. 3 Ou emi ocuses na owly on incomes, wi h occasional o ays in o consump ion expendi u es and ea nings. Inequali ies in weal h, land, educa ion, heal h, oppo uni ies, and poli ical powe a e he subjec o companion pape s in he LACIR p ojec . 3 uni o analysis.4 The e is some esidual a ia ion among da a sou ces ha p esen inequali y numbe s o pu po edly he same dis ibu ion, bu he anges a e ela i ely na ow and, again, can usually be unde s ood. Second, he ange o es ima es is conside ably augmen ed when in o ma ion is d awn om ax and/o social secu i y eco ds, and e en mo e so when a emp s a e made o econcile income o als om he ‘mic oeconomic’ sou ces wi h Na ional Accoun s agg ega es. While some easons o he dispa i ies can be su mised – ela ing, o example, o di e ences in he co e age o iche o poo e households; in he epo ing o capi al incomes; o in he ea men o ea nings e ained by i ms – i is much ha de o zoom in on a single p e e ed es ima e. All o he a ious sou ces ha e s eng hs and weaknesses and, u he mo e, a emp s o combine hem inhe en ly equi e assump ions and me hodological decisions ha a e – o a g ea e o lesse ex en – a bi a y. This gi es ise o a si ua ion o genuine unce ain y, whe e we become ai ly con iden ha inequali y is unde es ima ed when es ima ed om household su eys alone, bu a e less con iden ha he es ima es combining a ious sou ces unde hose mul iple assump ions a e accu a e. This unce ain y, in u n, leads us o deal wi h anges, o bands, o inequali y es ima es, which we a gue ep esen he ue (and unce ain) s a e o knowledge abou inequali y le els in La in Ame ica. Thi d, he unce ain y abou le els does no , in he main, a ec he b oad dynamic pa e n o how inequali y has e ol ed o e ime. He e, he e is ema kable consis ency, wi h mos coun ies in he egion ha ing ollowed an in e ed U cu e be ween he 1970s and he 2010s. The le els, he magni ude o he changes, and he exac iming o he peaks and oughs a y ac oss bo h coun ies and sou ces, as one would expec . Bu he o e all pic u e o ising inequali y in he i s hal o he pe iod, and declining inequali y om he mid-1990s o ea ly 2000s o he mid-2010s, is consis en and obus o di e en da a cons ella ions – wi h ew excep ions. The pape p oceeds as ollows. Sec ion 2 b ie ly discusses he ou main sou ces o he e ogenei y in he measu emen o income inequali y: di e en da a; di e en concep s; di e en ea men s o he same da a; and di e en summa y s a is ics. I also b ie ly desc ibes he (me a) da ase we ha e assembled. Sec ion 3 zooms in on es ima es de i ed om household su eys and p esen s a pic u e o he le els and dynamics o inequali y in eigh een coun ies in he egion as desc ibed by mos o he se ies a ailable since he 1940s. The pic u e o i een addi ional coun ies, including he Ca ibbean, o which e idence is spa se , is gi en in he appendix. We discuss he exis ing dispe sion o es ima es a ising om di e en analys s wo king on he same unde lying da a bu 4 Inequali y, as used he e, is a measu e o dispe sion in a dis ibu ion: a dis ibu ion o some hing, amongs a g oup o uni s. We use he e m ‘wel a e indica o ’ o e e o he a iable ha is dis ibu ed – income, consump ion, e c. – and ‘uni o analysis’ o e e o he uni s among whom he a iable is dis ibu ed – indi iduals, adul s, households, e c. 4 wi h di e en concep s o assump ions and show ha , while le els can a y and while he di ec ion o yea -on-yea changes can di e , he b oad dynamic pa e ns a e consis en . Sec ion 4 in oduces in o ma ion om adminis a i e sou ces ( ax and social secu i y) and na ional accoun s, b ie ly discusses he ecen de elopmen s in me hods o analyze hem, and examines he implica ions o ou unde s anding o inequali y le els and dynamics in he egion. We no e ha inco po a ing hese addi ional sou ces gi es ise o an addi ional dispe sion in es ima ed inequali y le els, and ha he new es ima es also su e om di e ences in me hods, in he ea men o da a, in ax codes, and in da a quali y. Sec ion 5 discusses he likely economic mechanisms d i ing he main obus inding, namely he in e ed U in inequali y dynamics which is b oadly common o all coun ies in he egion o which he a ailable s a is ical e idence allows a sui able ime co e age. I also in oduces some o he impo an di e ences among coun ies in he egion, he b oad common pa e n no wi hs anding. These inal ema ks a e necessa ily b ie and non-exhaus i e, and ce ainly do no a emp o p eemp he de ailed discussion o inequali y d i e s in he 27 chap e s ha ollow his one in he LACIR p ojec . Sec ion 6 concludes. 2. Why is he e disag eemen on how much inequali y he e is? When su eying he li e a u e and da a compila ions on income inequali y in LAC, one is aced wi h he ac ha he e a e o en a ew di e en es ima es o a pa icula coun y in any gi en yea .5 Any a emp o unde s and wha is eally going on wi h inequali y in he egion mus con on his ac head on and seek o unde s and i . A e all bu one o hese es ima es w ong? O pe haps, a e hey all w ong? Is he e one, unequi ocally igh way o measu e income inequali y? And i no , wha do we eally know abou he phenomenon?6 The e a e mul iple sou ces o his a ia ion in inequali y s a is ics which, o ou pu poses, can be help ully classi ied in o ou b oad g oups o a ie ies. The i s g oup (Va ie y 1) is he exis ence o di e en da a sou ces ha gi e ise o di e en dis ibu ions o any pa icula coun y and ime. In de eloping coun ies, household su eys and ax eco ds seldom ha e he same popula ion co e age. Household su eys ypically ail o cap u e op incomes p ope ly, whe eas ax egis e s will almos by de ini ion miss ou on wo ke s ha do no ha e o decla e incomes o he ax sys em, o who wo k ou side he o mal sec o en i ely. I em and uni non- esponses, as well as 5 Al hough i is no pa o ou emi , we suspec he same o be ue o mos o he pa s o he wo ld, pa icula ly de eloping ones. 6 We a e ce ainly no he i s o aise hese ques ions. Fo an ea lie discussion o he sensi i i y o inequali y measu es in La in Ame ica o how da a is ea ed, see Székely and Hilge (2007). 5 mis epo ing, can c ea e se ious biases o some household su eys, jus as ax e asion and a oidance s a egies can gene a e se ious measu emen e o s o ax da a. Some s a is ical o ices y o add ess non esponse be o e eleasing he su eys (no necessa ily in a co ec and anspa en manne ) and o he s do no . In o mali y limi s he ep esen a i eness and accu acy o social secu i y and o he adminis a i e wage da abases in he egion. When incomes om (any o ) hese sou ces a e agg ega ed o a na ional o al, ha o al is almos always a odds wi h income o als compu ed h ough he Na ional Accoun s Sys em (SNA) and econciling hem mus deal wi h possible e o s in all da a sou ces – including he SNA i sel – as well as concep ual di e ences in how some incomes a e ea ed. The second se o sou ces o he a ia ion in inequali y s a is ics (Va ie y 2) is ha , e en o one pa icula da ase , s a is ical o ices and esea che s may choose o wo k wi h di e en wel a e indica o s al oge he . Common choices include household incomes, indi idual ea nings, o consump ion expendi u es. Incomes may include impu ed en o owne ’s occupied housing and consump ion o own p oduc ion, o jus mone a y income. Household incomes o consump ion may be used as a o al, o in pe capi a e ms, o using some al e na i e equi alence scale ha seeks o compensa e o di e ences in needs wi hin he household, o o he exis ence o economies o scale in consump ion.7 They may also be epo ed in g oss e ms (i.e., be o e axes and social secu i y con ibu ions), o ne o axes and ans e s. The e may also be di e ences in he uni s o analysis: household income pe capi a, o example, may be dis ibu ed by households o indi iduals – and his will a ec o e all inequali y. Some hing as simple as mixing income-based wi h consump ion-based inequali y measu es can in oduce signi ican misconcep ions, some o which may pe mea e o he public discou se and ha e long-las ing e ec s. Conside , o example, he long-held iew ha La in Ame ica and he Ca ibbean is unambiguously he wo ld’s mos unequal egion, wi h Gini coe icien s a e aging app oxima ely 0.50, a le el highe han in any o he con inen .8 This iew, like Figu e 1 below which depic s a e age inequali y le els o a ious egions o he wo ld be ween 1990 and 2020, d aws on inequali y s a is ics om mul iple coun ies, some o which a e based on income pe capi a, and some on consump ion expendi u e pe capi a. When he c oss-coun y compa ison is made on a mo e consis en basis, using consump ion as common a iable, Sub-Saha an A ica comes ou as he geog aphic a ea wi h he highes a e age 7 See Coul e , Cowell and Jenkins (1992) o a classic su ey o equi alence scales. 8 D awing on a ela i ely long and s ong adi ion o na ionally ep esen a i e household su eys, esea che s el gene ally con iden o hei assessmen o inequali y, including global ankings. Deininge and Squi e (1996), o ins ance, s a ed ha hei da a se con i med he “ amilia ac ha inequali y in La in Ame ica is conside ably highe han in he es o he wo ld” (p. 566). 3. Income inequali y measu es om household su ey da a We y o begin ou e iew o inequali y s a is ics in LAC by ocusing on hose ha d aw only on household su eys. Howe e , ha is easie said han done, pa icula ly i one wan s o go back o he 1950s and 1960s, because many o he ea lies s udies on income dis ibu ion in he egion combined in o ma ion om he i s household su eys wi h o he sou ces, including censuses and he ( hen incipien ) na ional accoun s compu a ions. The de elopmen o economic s a is ics is a leng hy his o ical p ocess ha in ol es he na u e o p e ailing pa adigms, he cons uc ion o a body o con en ions, and he limi s o a ailable da a. The p oduc ion o such s a is ics engages go e nmen s, cen al banks, o icial s a is ics o ices, esea ch ins i u ions, as well as independen schola s a di e en s ages o he p ocess. And in he ea ly yea s o he Sys em o Na ional Accoun s, na ional accoun an s we e also expe s in dis ibu ional issues, as he in e -linkages be ween he es ima ion o na ional income and i s dis ibu ion we e clea ly ecognized. La in Ame ica pa icipa ed ully in his adi ion. Mos s udies o income dis ibu ion in he 1940s, 1950s, and 1960s sys ema ically a emp ed o p esen hei esul s in he con ex o he es ima ion o na ional income, o o inpu -ou pu ma ices, o as he esul o he in o ma ion p o ided by popula ion censuses.11 In A gen ina, o example, in pa allel wi h he es ablishmen o he sys em o na ional accoun s, CONADE-CEPAL (1965) se ou o es ima e he pe sonal and he unc ional dis ibu ion o income in g ea de ail o he yea s 1953, 1959 and 1961, making use o su eys, popula ion and indus ial censuses, income ax egis ies, and social secu i y eco ds, and a emp ed a econcilia ion wi h na ional income.12 In B azil, he se en h popula ion census o 1960 was he i s o enqui e abou incomes. Samples we e exploi ed by Ho man (1971), Fishlow (1972), and Langoni (1973) o es ima e he dis ibu ion o income a he na ional le el o he labou o ce ( he População Economicamen e A i a, o PEA). In Mexico, he 1950 census included ques ions on income; addi ionally, su eys we e in oduced a he ea ly in he coun y; he i s one in 1956. Bo h sou ces, oge he wi h he na ional accoun s and inpu -ou pu ma ices, we e exploi ed by Na a e e (1960) and la e c i ically e iewed by Al imi (1982) o p oduce an es ima e o income inequali y o 1950.13 11 In se e al Ca ibbean coun ies (in pa icula , hose o me colonies o e i o ies o he B i ish Empi e), i is possible o es ima e he sha es acc uing o op income g oups o e se e al decades o he i s hal o he 20 h cen u y based on ax da a. While hese coun ies ha e oday spa se and mo e p oblema ic in o ma ion on he dis ibu ion o incomes han he es o he egion, hey had he bes exis ing aw da a a ha ime. The e is hus scope o expanding he ime co e age o he pic u e p esen ed in Figu e 3. 12 I onically, la e and o his da e, A gen ina’s su eys co e u ban a eas only. I is he only coun y in La in Ame ica whose inequali y es ima es sys ema ically lea e ou abou one hi d o he popula ion. 13 A comp ehensi e desc ip ion o inequali y se ies and s udies co e ing he yea s 1940-1985 can be ound in se en epo s published by CEPAL/ECLAC on A gen ina, B azil, Chile, Colombia, Mexico, Pe u and Venezuela (1986a,b, 1987a,b, 1988a,b, 1989). 13 All he ea ly s udies con ain a weal h o in o ma ion on inequali y in he egion in he 1940s, 50s and 60s. Un o una ely, hey su e om a numbe o he sou ces o non-compa abili y discussed in Sec ion 2 – o Va ie ies 1, 2 and 3 – which hampe hei use ulness o c oss-coun y compa ison o inequali y le els, o o long- un dynamic s udies. The uni o analysis, he de ini ions o income, he se o impu a ions, he spa ial co e age, and he expe s’ judgemen s a y o e ime and ac oss coun ies. Fo be e o wo se, in he decades ha ollowed, beginning in he la e 1960s and 1970s, he link be ween SNA agg ega es and dis ibu ional da a weakened and he dominan li e a u e in he wo ields (na ional accoun s and dis ibu ion s udies) wen hei sepa a e ways, globally and in La in Ame ica. A his poin , he applied analysis o he dis ibu ion o income began o ely almos exclusi ely on household su eys, which we e i s ielded a di e en imes in di e en coun ies in he egion. The i s a emp s we e no always led by go e nmen s h ough he s a is ics ins i u es o o he o icial bodies, and in some cases, hey we e isola ed and explo a o y inqui ies. The p ocess o make su eys sys ema ic and ha e hem a egula in e als ook mo e ime. Many i no mos o hese su eys co e ed u ban a eas only, o e en jus he capi al ci y – a leas a i s . Fo ins ance, in Chile, he i s ep esen a i e su eys s a ed in 1956 wi h he Encues as de Ocupación y Desocupación, a cons an -me hodology da a sou ce o he G an San iago conduc ed by he Uni e sidad de Chile un il 1979, which is he sou ce o he inequali y se ies p oduced by Heskia, 1980). These we e ollowed in 1963/64 by he Encues a de P esupues os Familia es, conduc ed by he Uni e sidad Ca ólica. In Colombia, na ionally ep esen a i e household su eys a e a he ecen , s a ing only in 2001. Ye , labo o ce/household su eys began in 1970 by he Depa amen o Adminis a i o Nacional de Es adís ica (DANE), wi h some i egula i y and a ying co e ages. I is only a e 1975 ha qua e ly su eys became a ailable o he u ban sec o (7 ci ies) and a consis en annual se ies could be cons uc ed, al hough co e ing only pa o he u ban sec o and mos ly labo incomes. A gen ina launched he o icial P og ama de Encues a Pe manen e de Hoga es, EPH, in 1972. B azil began he de elopmen o a sys ema ic su ey sys em in 1967, leading o he Pesquisa Nacional po Amos a de Domicílios, PNAD. Bo h he EPH and PNAD su i e, wi h changes, o his day. Viewed om a cu en pe spec i e, when a g ea deal o wo k is once again seeking o combine hose di e en sou ces o in o ma ion, his may appea eg e able. None heless, ha pa h had a leas one posi i e consequence: he inc easingly sys ema ic collec ion o su eys since he 1970s, as well as he possibili y o exploi ing mic oda a se s, expanded he ime co e age and imp o ed 14 he c oss-coun y compa abili y. This has allowed schola s, o icial s a is ical o ices, and esea ch ins i u ions o p oduce mo e compa able inequali y se ies, bo h o e ime and ac oss coun ies. Thus, we ha e easonably consis en – albei impe ec – inequali y se ies s a ing in he 1970s, 1980s o he 1990s, depending on he coun y, and a pa chwo k o sho e ime se ies o indi idual obse a ions, many imes p oduced by independen esea che s, o he yea s going back o he 1940s. A ew in e na ional ins i u ions and esea ch cen e s, bo h in La in Ame ica and beyond, ha e made e o s o compile hese se ies in public-use da abases, and we build on hei wo k in ou own compila ion. Fou o hese da abases p oduce hei own es ima es om a ailable mic oda a collec ed h ough household su eys. These a e: (a) he Socioeconomic Da abase o La in Ame ica and he Ca ibbean (SEDLAC), p oduced by he Cen e o Dis ibu i e, Labo and Social S udies (CEDLAS) a he Uni e si y o La Pla a; (b) Po calNe , now eplaced by Po e y and Inequali y Pla o m (PIP) o he Wo ld Bank; (c) he se ies p oduced by he UN Economic Commission o La in Ame ica and he Ca ibbean (CEPAL o ECLAC); and (d) he Luxembou g Income S udy (LIS), which has been p og essi ely inco po a ing La in Ame ican coun ies (so a B azil, Chile, Colombia, Gua emala, Mexico, Panama, Pe u, Pa aguay and U uguay) o i s da abase.14 The e a e also seconda y-sou ce da ase s, which compile summa y s a is ics epo ed elsewhe e, a he han gene a ing hei own om he mic oda a. These include, among o he s, he All he Ginis Da ase cons uc ed by B anko Milano ic a CUNY, and he Wo ld Income Inequali y Da abase (WIID) main ained by UNU-WIDER.15 Mic oda a-based da ase s a e o en p e e ed because hei se ies end o be in e nally consis en . Bu ou pu pose he e is p ecisely o ake s ock o he ull ange o a ailable es ima es, so as o assess he a iabili y ac oss he ield. To his end, we g ea ly bene i om he comp ehensi e collec ion o he exis ing his o ical se ies ha has been assembled by UNU-WIDER in WIID. We use i as he main building block o he analysis in his sec ion, o wo main easons. Fi s , i s ime co e age goes back u he in ime han any o he da abase o LAC. Second, i includes he co e o he inequali y indica o s p oduced by he leading p ima y, mic oda a-based da ase s men ioned abo e. As such, he WIID is a use ul in o ma ion agg ega o ha combines he me i s o he longe ime spans allowed by seconda y sou ce da ase s, wi h he e o s owa ds ime and 14 The e is a close ela ionship be ween SEDLAC and PIP, in ha he Wo ld Bank is a pa ne o CEDLAS in he p ojec and d aws on SEDLAC o build i s own es ima es o La in Ame ica. Tha said, he e is addi ional ‘ ea men ’ a he Bank, which leads o he e being some di e ences be ween he wo se ies, as we will see below. 15 The OECD also p oduces inequali y s a is ics, ollowing i s own guidelines, o h ee LAC membe s (Cos a Rica, Mexico and Chile), and one non-membe (B azil). They a e gi en in OECD.S a . These obse a ions a e also included in Figu e 3 abo e, and in Figu e 4 below. 15 c oss-coun y compa abili y allowed by he se ies buil on he o iginal mic oda a sou ces o he mo e ecen pe iod.16 As no ed in Sec ion 2, we complemen WIID wi h obse a ions om addi ional independen s udies. Res ic ing he da abase om Figu e 3 o Gini coe icien s a ising only om household su eys a e 1970 (bu also including he a o emen ioned ea ly s udies ha use da a om censuses p io o ha yea ) gi es ise o he sub-da ase p esen ed in Figu e 4, which co e s he same eigh een coun ies shown in Figu e 3. As be o e, an analogous igu e o he o he six een coun ies is shown in he appendix (Figu e A2). Gi en he spa si y o in o ma ion on his las se o coun ies, which un o una ely includes all o he Ca ibbean coun ies in ou da a, we ocus he emainde o ou discussion on Figu e 4. Because he obse a ions in Figu e 4 a e all Gini coe icien s and, a leas in he la e pe iod, exclusi ely om household su eys (and censuses, on he g ounds ha he p ocess o collec ion o incomes is ela ed o ha o su eys), he igu e elimina es he ou h a ie y o compa abili y p oblems (di e en indices), and some o he i s a ie y (di e en da a sou ces).17 Mos o he me hodological a ia ion in Figu e 4 he e o e comes om Va ie ies 2 and 3: di e ences in wel a e indica o o uni o analysis, and di e en ea men o he da a. We use a colo scheme o highligh some o hese di e ences and o s uc u e he discussion. 16 Ye he e a e no pe ec da a o da ase . Fo a ho ough e iew o WIID, see Jenkins (2015). The same pape explains why we do no include he S anda dized Wo ld Income Inequali y Da abase – ye ano he al e na i e – in ou analysis. 17 Analogous se ies o he income sha es o he op 10% and 1% a e a ailable om he au ho s on eques . Figu e 4. P edominan ly HHS-based Gini coe icien s in La in Ame ica and he Ca ibbean Sou ce: WIID complemen ed by he au ho s. No es: Band 2 is shown. The plo s o he emaining coun ies in La in Ame ica and he Ca ibbean, which ha e ewe obse a ions and mo e agmen ed se ies, can be ound in Appendix Figu e A2. 17 Figu e 4. P edominan ly HHS-based Gini coe icien s in La in Ame ica and he Ca ibbean (con inued) Sou ce: WIID complemen ed by he au ho s. No es: Band 2 is shown. The plo s o he emaining coun ies in La in Ame ica and he Ca ibbean, which ha e ewe obse a ions and mo e agmen ed se ies, can be ound in Appendix Figu e A2. The se ies highligh ed in ed (SEDLAC), g een (ECLAC new), yellow (Po calNe -PIP), and pu ple (LIS) come om he ou mic oda a-based compila ions men ioned abo e.18 All hese se ies use as wel a e indica o household pe capi a income, bu he e is conside able unce ain y as o whe he ha income is p e- o pos - axes. Some imes, he unce ain y o igina es om agueness in he ques ionnai es o in he enume a o ins uc ions, which means ha esponden s a e no asked clea ly and speci ically o epo ei he ne o g oss incomes. This is a su p isingly common and se ious issue in he egion, which equi es a en ion. The geog aphical co e age is always he coun y, excep o Colombia be o e 2001, A gen ina and U uguay, whe e i is u ban a eas (U uguay s a ed na ionally ep esen a i e se ies in 2009).19 The se ies ma ked in blue a e obse a ions om independen s udies, bu wi h an income de ini ion equal o close o household pe capi a disposable income. This is ypically he case o s udies co e ing he 1950s, 1960s and 1970s. All poin s in g ay e e o Gini coe icien s calcula ed using wel a e indica o s o he han household pe capi a income. They e e o se ies based on equi alized income, o al household income, income o he head o household, g oss income, ma ke income, consump ion, o ea nings. They also include na ional, u al, and u ban igu es; and some imes e e o a pa ial co e age o he na ional popula ion. They may also use di e en uni s o analysis: households, indi iduals, adul s only, e c. In his pape we do no discuss he me i s and he limi a ions o each o he se ies, no do we ocus on app aising he quali y o he da a.20 Ins ead, we ake he b oad span o he (me a) da a as ou subjec o s udy, and documen h ee main ac s ha jump ou o Figu e 4. Fi s , he dispe sion in le el es ima es ac oss he ull se o poin s, on any gi en yea , is conside ably na owe han in Figu e 3: na owing he scope o include p ima ily household su ey-based measu es emo es a la ge pa o he a ia ion. Second, hough, some a ia ion emains, e en hough mos o hese summa y s a is ics a e based on he same aw su eys. This emaining dispe sion should be elling o how impo an de ini ions a e, and o how much he analys ’s me hodological choices ma e when looking a he le els o yea - o-yea luc ua ions in inequali y.21 18 In WIID he e a e wo se ies o ECLAC: an olde one which includes a ( ela i ely coa se) adjus men o missing capi al incomes based on he gap wi h Na ional Accoun s and which has been discon inued, and a newe one ha is solely based on su eys. Bo h we e included in Figu e 3 bu only he newe one is included in Figu e 4. 19 Gini coe icien s a he coun y le el o A gen ina, Colombia, and U uguay, when a ailable, a e plo ed in g ey in Figu e 4. The Gini coe icien o u ban a eas allows o he longes his o ical se ies in hese coun ies. Mo e p ecisely, in A gen ina household su eys e e ed only o he G ea e Buenos Ai es be ween 1974 and 1991; he numbe o u ban a eas co e ed has been inc easing since hen. 20 In e es ed eade s a e e e ed o he special issue o he Jou nal o Economic Inequali y dedica ed o he app aisal o inequali y da abases (Fe ei a, Lus ig and Teles, 2015). 21 A he same ime, he e ha e also been a emp s o p oduce long- e m, compa able se ies based on hose many exis ing s udies shown in Figu e 3 and Figu e 4, including wi hin WIID (G adín, 2021). The le els and ends shown 19 Thi d, despi e his la ge a iabili y in le els (e en ac oss compa able de ini ions), Figu e 4 also sugges s ha household su ey es ima es o inequali y ell a consis en s o y in e ms o ends, o medium- e m dynamics. When conside ing he big pic u e o e his long pe iod ( ha is, abs ac ing om annual changes), he wide a ay o es ima es can be ead and in e p e ed in e ms o inequali y bands: in e als wi hin which a ailable es ima es lie e e y yea . Indeed, gi en ou colo -coded classi ica ion o he obse a ions in Figu e 4, one can de ine wo di e en bands. Band 1, he wides , is simply he annual ange o es ima es ac oss he en i e da ase (including blue and g ay poin s). Band 2 gi es ou p e e ed, na owe in e p e a ion o household su ey-based inequali y indices, namely he ange o es ima es ac oss he ou mos compa able mic oda a d i en se ies (SEDLAC, new CEPAL, Po calNe -PIP and LIS). While Band 1 can be seen o he ull imespan co e ed in each coun y, Band 2 begins la e , depending on when he mic oda a se ies commences. I can go as a back as he ea ly 1970s o A gen ina, o he ea ly 1980s o B azil and Mexico, o as la e a he ea ly 2000s o Gua emala. Fo ease o isualiza ion, only Band 2 is plo ed in Figu e 4, as do ed black lines. Band 1 is he easies o see on i s own: i is simply he ange o he da a, yea on yea . Table 1 p esen s ( he decadal a e ages o ) he ele an in o ma ion o bo h de ini ions o “inequali y bands’: he maximum Gini, he minimum Gini, and he di e ence be ween hem. Figu e 5 shows he a e age wid h o he wo bands ac oss he en i e pe iod, o each o he eigh een coun ies in Figu e 4. I shows ha when summa y measu es o inequali y use di e en concep s o he wel a e indica o o uni o analysis ( he g ay poin s), he he e ogenei y is s ill conside able, anging om a ound h ee Gini poin s (in A gen ina) o mo e han en in Pe u, Mexico, and Gua emala. The band wid h is conside ably educed when looking only a he ha monized se ies in Band 2. These colo ed se ies depic Gini coe icien s o he same wel a e concep s and om he same da ase s, so ha only di e ences in ea men o he da a (Va ie y 3) a e le . The a e age ange (o e ime) o his se o es ima es, shown in Figu e 5, is ypically below wo Gini poin s – a much mo e accep able deg ee o unce ain y o ha e abou inequali y le els. None heless, i is h ee poin s o abo e in h ee coun ies, namely Chile, Gua emala, and Venezuela. in hose a emp s - which lie beyond ou emi - should be ead wi h cau ion, in ligh o he discussions by A kinson and B andolini (2001), Fe ei a, Lus ig and Teles (2015), and Jenkins (2015). 20 Table 1. Gini coe icien ac oss compa able su ey-based se ies and inequali y bands 1980s-2020s 21 Figu e 5. Inequali y Band wid hs using HHS es ima es A e age 1980-2020 Ac oss he as majo i y o he coun ies shown in Figu es 4 and 5, he dynamics o inequali y can be cha ac e ized by h ee ‘episodes’, he cen e piece o which is an in e ed U. This is easily seen om he bands in Figu e 2, bu Table 2 below also lis s he maxima and minima (o peaks and oughs) o each o he ou ha monized se ies con ained wi hin Band 2. The i s episode begins in he 1970s o ea ly 1980s and is ma ked by a no able inc ease in income inequali y which peaked in he la e 1990s o ea ly 2000s, depending on he coun y. This ise in inequali y was sus ained, despi e sho - un luc ua ions. This episode includes he e ec s o he deb c isis and he subsequen “los decade” in he 1980s, he high-in la ion and hype in la ion yea s in A gen ina, B azil, Pe u, and U uguay o e 1989-1991, as well as he pe iod cha ac e ized by sweeping ma ke -o ien ed e o ms (i.e.: ade libe aliza ion, de egula ion, and p i a iza ion) in he 1990s. 28 Sou ce: Se ies om De Rosa, Flo es, and Mo gan (2022) 29 Figu e 8. Top 1% income sha e 2000-2020: household su eys and he e ec s o adjus men s wi h adminis a i e da a and na ional accoun s Sou ce: Se ies om De Rosa, Flo es, and Mo gan (2022) 30 Figu e 9. Top 10% income sha e 2000-2020: household su eys and he e ec s o adjus men s wi h adminis a i e da a and na ional accoun s Sou ce: Se ies om De Rosa, Flo es, and Mo gan (2022) 31 Figu e 10. Middle 40% income sha e 2000-2020: household su eys and he e ec s o adjus men s wi h adminis a i e da a and na ional accoun s Sou ce: Se ies om De Rosa, Flo es, and Mo gan (2022) 32 Figu e 11. Bo om 50% income sha e 2000-2020: household su eys and he e ec s o adjus men s wi h adminis a i e da a and na ional accoun s Sou ce: Se ies om De Rosa, Flo es, and Mo gan (2022) 33 The hi d (and uppe mos ) se ies in each plo in Figu es 7-11 inco po a es he impu a ion o he gap wi h Na ional Income om he SNA, including an es ima e o undis ibu ed p o i s, which a e e ained by i ms. In Figu e 7, he scaling up o NA p oduces a subs an ial upwa d shi in he se ies. Adding hese “missing” incomes o households in he su eys has i s and o emos a le el e ec : hey add be ween 5 and 10 pe cen age poin s o he Gini coe icien . Bu again, he dynamics o inequali y as depic ed by he su ey do no change subs an ially, wi h wo excep ions. In Mexico, he decline in inequali y gi en by su eys du ing he second hal o he 2010s disappea s, wi h inequali y s aying mo e o less cons an . In Chile, pe haps su p isingly, he adjus men u ns he su ey and ax-base decline in inequali y o e 2000-2010 in o an inc ease, and he inc ease o e 2010-2020 in a decline. Why is his inal adjus men made? E en aken in o conside a ion he adjus men s o su eys wi h ax in o ma ion, he e is s ill he la ge and some imes inc easing gap be ween agg ega es om inequali y s udies based on mic oeconomic da a and he SNA. The disc epancies can be seen in he le els o income, as well as in hei g ow h a es (see, o example Ra allion (2003); Dea on (2005); Bou guignon (2015); Nolan e al. (2019)) and can a ain pa icula ly high le els in de eloping coun ies. While i may no be su p ising ha na ional income is la ge han he income concep s adi ionally used o s udy inequali y, i has some imes been g owing as e oo. I has been a gued ha hese disc epancies make i ha d o assess how mac oeconomic g ow h is dis ibu ed ac oss income g oups, and o wha ex en exis ing dis ibu ional s a is ics a e a p ope ep esen a ion o he income lows in an economy. Implici ly oo ed in he adi ion o inequali y esea ch in he 1940s, 1950s and 1960s (men ioned in sec ion 2), as well as in he me hods de eloped by ECLAC and Al imi (1986, 1987), ecen wo k has emba ked on a p ocess o scaling up he a ious a ailable da a sou ces (su eys, adminis a i e eco ds, ich lis s) – h ough u he impu a ions – o SNA o als. These include, among o he s, Wo ld Inequali y Lab (2020), Fixle e al. (2017) and he DNA-Dis ibu ional Na ional Accoun s p ojec coo dina ed by he OECD (Zwijnenbu g, 2019). While he exis ing gaps ha e some imes s eng hened he eelings o unce ain y abou inequali y measu emen , hese new app oaches ha e aken o g an ed he numbe s p o ided by he na ional accoun s, a p ac ice ha does no always con ibu e o diminish hose eelings, a leas in he case o de eloping coun ies.33 The si ua ion in La in Ame ica conce ning he compa ison o agg ega es ac oss di e en sou ces is discussed in Al a edo e al. (2022) and in De Rosa, Flo es, and Mo gan (2022). They show ha 33 I should be emembe ed ha while he de ini ion o na ional income ce ainly includes an es ima e o undis ibu ed p o i s (which could be conside ed as a p oxy o co po a e capi al gains), i excludes o he i ems which may be e y ele an o he g oups in he middle o he dis ibu ion, such as capi al gains om eal es a e. This poin s o he ac ha dis ibu ing he Na ional Income does no close he ques ions on he mos sui able de ini ion o income o inequali y s udies. 34 he gap be ween he o al income epo ed in household su eys and he na ional income es ima e om he na ional accoun s is e y la ge, wi h o al su ey income co e ing om 40% o 60% o na ional income. Mexico appea s as an ex eme case, whe e he gap is close o 70%. Acco ding o Al a edo e al. (2022), he explana ion o he gap is sha ed by concep ual issues as much as by measu emen issues. O e all, hese pape s p esen a dismal message: scaling up incomes om he su eys, e en a e adjus ing wi h adminis a i e da a, o close he gap wi h he SNA agg ega es means impu ing he dis ibu ion o hal o he na ional income, mos ly composed o i ems abou whose dis ibu ion we know he leas : capi al incomes and e ained ea nings. This is no a mino de ail. Impu a ions ely on agmen a y in o ma ion as much as on he esea che judgemen . Gi en he disc epancies a s ake in La in Ame ica, his unde aking becomes a sensi i e issue, wi h po en ially la ge e isions o cu en ly accep ed inequali y ends. The gaps be ween inequali y es ima es om ax and o he adminis a i e da a and household su eys – pa icula ly when he o me is scaled up o ma ch Na ional Accoun s agg ega es – dwa es he inequali y bands among es ima es based exclusi ely on household su eys. Table 3 below p esen s he wid h o h ee inequali y bands, bu now only o he en coun ies in Figu es 7-11. Fi s , we e-s a e he inequali y Bands 1 and 2 om Sec ion 2, and hen we epo addi ional bands be ween he lowe bound o Band 2 and he wo es ima es om Figu e 7: he su ey adjus ed wi h ax da a/BFM es ima e (Band 3), and he es ima e scaled up o ma ch he SNA o al (Band 4). Table 3. Fou inequali y bands o LAC, 2000-2020 35 Band 1, in he i s column o Table 3, was al eady discussed in Sec ion 3. I includes es ima es om concep s o he han household pe capi a incomes, so he addi ional a ia ion wi h espec o Band 2 is accoun ed o by di e ences in he concep s o wel a e indica o and uni s o analysis (Va ie y 2). We now ocus on a compa ison ac oss he o he h ee inequali y bands, which s a om he lowe bound o Band 2 – ha is, he lowes inequali y es ima e ( o household pe capi a income om he household su eys) o a coun y/yea . Wi h ha lowe bound, Band 3 inco po a es he BFM ax adjus men s epo ed by De Rosa, Flo es, and Mo gan (2022), and Band 4 also accoun s o he scaling up o SNA ca ied ou by he same au ho s. The a e age wid hs (o e ime) o Bands 2, 3 and 4 o each coun y a e shown in Figu e 12. Figu e 12. A e age wid h o inequali y bands o e 2000-2020: among HHS-based measu es; including ax adjus men s; and including scaling up o SNA agg ega es. Ac oss he whole pe iod and all en coun ies, he a e age wid h o Band 2 is 1.7 pp, while i is 10.4 pp o Band 3 and 16.9 pp o Band 4. In o he wo ds, inco po a ing adjus men s om he ax 36 da a in LAC adds an a e age o 8.7 pe cen age poin o he a ia ion in inequali y measu es, and a emp ing o scale up o he SNA adds ano he 6.5 pp. Gi en he s eng h o he assump ions ha mus be made o make hese adjus men s, and he a ia ion in ou comes ac oss me hods, which we ha e summa ized abo e, we ind i di icul o ea hese numbe s as many appea o ha e – as unambiguous indica o s o how much highe inequali y eally is in he egion. We belie e ha hese esul s do con i m ha household su ey-based measu es unde es ima e inequali y in he egion. Bu hey do no ye ep esen c edible, obus poin es ima es o he ue, highe le el. Gi en he a ailable in o ma ion, we belie e looking a he h ee bands desc ibed abo e gi es us a sense o he ange whe e ue income inequali y in household pe capi a income is likely o lie. This ange is ypically qui e wide – a ound 10 pp o he ax adjus men only, and close o 17 pp when scaling up o SNA – bu he le els a e e y high ac oss he en i e plausible bands. In o he wo ds, e en unde he almos ce ainly unde es ima ed household su eys, LAC has Gini coe icien s ha a e almos always g ea e han 0.40. Depending on he adjus men s, hey can some imes exceed 0.70. This is e y high inequali y indeed, al hough p ecisely how high emains, o he momen , di icul o say. Fo una ely, as documen ed abo e, he dynamics and, in pa icula , he ough in e ed U pa e n be ween Episodes 1 and 2, a e much mo e obus . They appea o hold – albei wi h di e en iming and in di e en magni udes – o all bu one o wo coun ies, ac oss all h ee bands. In he nex sec ion, we b ie ly conside some o he economic – and poli ical – ac o s likely o ha e played a ole in unde pinning hese dynamics. Gi en all he unce ain y, he discussion should be ead as la gely specula i e. 5. In e p e ing he inequali y dynamics in LAC Despi e he di icul y o measu ing inequali y o income and he a ious no necessa ily consis en da a sou ces o do so, inequali y in LAC coun ies seems o ha e ollowed a easonably clea pa e n o e he las ew decades. Fi s , on he basis o inequali y es ima es g ounded on household su eys and judging om he coun ies o which such es ima es ha e been a ailable now o e 40 yea s o mo e, inequali y was highe in he mid- o la e-1990s han in he p eceding decades o e which da a a e a ailable. The ac ual pa h is o en unce ain. The ise was con inuous in A gen ina and Chile – excep o he Allende yea s whe e inequali y declined o a ew yea s – whe eas he e a e no enough obse a ions o say whe he his has also been he case in B azil o Mexico. Second, o mos coun ies in he egion, inequali y has s eadily declined o e he 2000-2015 pe iod, and he cumula i e decline has been subs an ial, eaching in some ins ances mo e han 10 Gini poin s. Thi d, his common end seems o ha e b oken down in mos coun ies in he mid-2010s, wi h inequali y mo ing up again in some coun ies (B azil, be o e Co id), la ening (A gen ina, Chile, 37 Colombia, Panama, U uguay) o s ill declining, possibly a a slowe a e (Boli ia, Pe u and Pa aguay) A he isk o o e -simpli ying he pic u e o he p e-1990 decades, a schema ic ep esen a ion o he e olu ion o LAC inequali y is ha o in e ed-U, wi h a op in he la e 1990s, and a la ening o he descending a m a ound 2015. Such a desc ip ion pe ec ly i s coun ies like A gen ina o Colombia, i may be mo e app oxima i e o Mexico and B azil. I is, o cou se, only a guess based on in ui ion ha he e mus be some simila i y in coun ies – such as Gua emala and Nica agua – whe e no inequali y es ima e is a ailable be o e 2000 and only he descending pa o he in e ed U is obse ed. Based on ha simple ep esen a ion o he e olu ion o inequali y, he ques ion we ask in his sec ion is ha o he in e p e a ion o such a long and p onounced inequali y cycle. Two ypes o in e p e a ions a e possible. They depend on he na u e o he dynamic p ocess behind he in e ed U shape and he ecen la ening o he ajec o y. In he i s scena io, he e olu ion o inequali y is a kind o andom p ocess wi h a sequence o shocks, wi h some simila i y ac oss coun ies, and some o hem wi h a long memo y. Then he ascending pa would simply co espond o a succession o shocks wi h a ne posi i e impac on inequali y, he mos ecen ones ending o ein o ce he e ec s o he p e ious ones. Then he p ocess e e ses in he ea ly 2000s when some opposi e shock, o a sequence o shocks, occu s, which p og essi ely e ases he e ec s o he p eceding sequence and igge s a i uous e olu ion. The bad sequence o common shocks ha hi La in Ame ica om he la e 1970s o he ea ly 1990s is a he ob ious: he oil p ice boom in he 1970s, he subsequen in la iona y p essu es and slowdown o economic ac i i y, he deb c isis caused by he ise o in e es a es in 1982, he se e e ecession ha ollowed, he IMF-led mac oeconomic adjus men s. Mos o hese shocks a ec ed he bo om o he income scale mo e se e ely han he op as, o ins ance, wi h he necessa y educ ion in eal labo ea nings a ising om a de alua ion esponding o an ad e se ex e nal posi ion. I is ue ha capi al was also se e ely hu , which could ha e mode a ed he impac on inequali y. Howe e , his canno be checked due o es ima es in ha pe iod elying mos ly on su eys and wi hou a simple solu ion o gene a e al e na i e es ima es ha would include op incomes and hei possible losses. The symme ic o he p eceding sequence o bad shocks has been he huge imp o emen in e ms o ade a e 2000, he p og essi e decline in in e es a es, global g ow h a unusually high a es, and a subs an ial d op in in la ion a es all a ound he wo ld. I is also he case ha cen e -le go e nmen s came o powe in se e al coun ies - some hing ha may no be o ally exogenous wi h espec o moun ing inequali y in he p eceding decades and sugges s ha , o e all, ising 44 Figu e A1. Gini coe icien s o six een smalle coun ies in La in Ame ica and he Ca ibbean (complemen ing Figu e 3) (con inued) No e: The igu e shows all he Gini coe icien s om WIID, om De Rosa, Flo es, and Mo gan (2022), and om addi ional his o ical se ies and s udies collec ed by he au ho s. 45 Figu e A2. P edominan ly HHS-based Gini coe icien s o six een smalle coun ies in La in Ame ica and he Ca ibbean (complemen ing Figu e 4) No e: The igu e shows all he Gini coe icien s om WIID, and om addi ional his o ical se ies and s udies collec ed by he au ho s. 46 Figu e A2. P edominan ly HHS-based Gini coe icien s o six een smalle coun ies in La in Ame ica and he Ca ibbean (complemen ing Figu e 4) (con inued) No e: The igu e shows all he Gini coe icien s om WIID, and om addi ional his o ical se ies and s udies collec ed by he au ho s. REFERENCES Al imi , O. (1982). La dis ibución del ing eso en México, 1950-1977. In Dis ibución del ing eso en México: Ensayos. Cuade no 2, Tomo 1. México, DF: Banco de México. Al imi , O. (1986) Es imaciones de la dis ibución del ing eso en la A gen ina, 1953-1980. Desa ollo Económico. 25(100): 521–566. Al imi , O. (1987). Income dis ibu ion s a is ics in La in Ame ica and hei eliabili y. Re iew o Income and Weal h. 33(2): 111–155. Al a edo, F. (2010) The ich in A gen ina o e he wen ie h cen u y, 1932-2004. In A.B. A kinson and T. Pike y (eds.), Top Incomes. A Global Pe spec i e. Ox o d: Ox o d Uni e si y P ess. Ch. 6, pp. 253–298. Al a edo, F., C uces, G. and Gaspa ini, L. (2018). A Sho Episodic His o y o Income Dis ibu ion in A gen ina. La in Ame ican Economic Re iew. 27(7) Al a edo, F. and Gaspa ini, L. (2015). Recen ends in inequali y and po e y in de eloping coun ies. In A.B. A kinson and F. Bou guignon (eds.) Handbook o Income Dis ibu ion. No h Holland. Ch. 9, pp. 697–805. Al a edo, F. and Londoño Vélez, J. (2013). High Incomes and Pe sonal Taxa ion in a De eloping Economy: Colombia 1993-2010. CEQ Wo king Pape 12. Al a edo, F. and Londoño Vélez, J. (2014). Al os ing esos e impues o de en a en Colombia 1993- 2010. Re is a de Economía Ins i ucional. 16(31): 157–194. Al a edo, F., De Rosa, M., Flo es Beale, I. and Mo gan, M. (2022). Desigualdad del ing eso en la República Dominicana 2012-2020. Una e isión a pa i de la combinación de uen es de in o mación. Publicación de las Naciones Unidas/CEPAL. Al a edo, F., De Rosa, M., Flo es, I., and Mo gan, M. (2022). The inequali y (o he g ow h) we measu e. Da a gaps and he dis ibu ion o incomes. CEPR DP 17135. A kinson, A. B. and B andolini, A. (2001) P omise and pi alls in he use o “seconda y” da a-se s: Income inequali y in OECD Coun ies as a case s udy. Jou nal o Economic Li e a u e, 39(3): 771–799. A kinson, A. B. and Pike y, T. (2007) Top Incomes o e he Twen ie h Cen u y. A Con as be ween Con inen al Eu opean and English-Speaking Coun ies. Ox o d: Ox o d Uni e si y P ess. A kinson, A. B. and Pike y, T. (2010) Top Incomes. A Global Pe spec i e. Ox o d: Ox o d Uni e si y P ess. 48 Aze edo, J. P., Inchaus e, G., San elice, V. (2013). Decomposing he ecen inequali y decline in La in Ame ica. Wo ld Bank Policy Resea ch Wo king Pape 6715. Blanche , T., Flo es, I. and Mo gan, M. (2022) The weigh o he ich: imp o ing su eys using ax da a. Jou nal o Economic Inequali y. 20: 119-150. Bo az, F. and Gonzalez-Pampillon, N. (2017). Assessing he dis ibu i e e ec s o minimum wage. Re iew o De elopmen Economics. 21(4): 1081-1112. Bou guignon, F. (2015). App aising income inequali y da abases in La in Ame ica, Jou nal o Economic Inequali y. 13(4): 557–578. B acco, J., Gaspa ini, L., Galeano, L. and Pis o io, M. (2021). Slowdown o mone a y inequali y educ ion in de eloping coun ies. Jou nal o Income Dis ibu ion. 29(1–2). Bu dín, G., De Rosa, M., Vigo i o, A. and Vilá, J. (2022). Falling inequali y and he g owing capi al income sha e: Reconciling di e gen ends in su ey and ax da a. Wo ld De elopmen . 152. Bus os, A. and Ley a, G. (2017). Towa ds a mo e ealis ic es ima e o he income dis ibu ion in Mexico. In Uni ed Na ions Economic Commission o Eu ope, Con e ence o Eu opean S a is icians, Expe mee ing on measu ing po e y and equali y, 26-27 Sep embe 2017. Bud a, Mon eneg o. Cano, L. (2015). Income mobili y in Ecuado : new e idence om indi idual income ax e u ns. UNU-WIDER Wo king Pape 2015/040. Cano, L. (2017). Pe sonal income ax and income inequali y in Ecuado be ween 2007 and 2011. Cepal Re iew. 123: 56-75. Cechini, S. and Mada iaga, A. (2011). Condi ional Cash T ans e P og ams: he ecen expe ience in La in Ame ica and he Ca ibbean, ECLAC, San iago, Chile Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1986a). An eceden es es adís icos de la dis ibución del ing eso. B asil 1960-1983. Se ie Dis ibución del Ing eso 2. Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1986b). An eceden es es adís icos de la dis ibución del ing eso. Colombia 1960-1983. Se ie Dis ibución del Ing eso 1. Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1987a). An eceden es es adís icos de la dis ibución del ing eso. Chile 1940-1982. Se ie Dis ibución del Ing eso 3. Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1987b). An eceden es es adís icos de la dis ibución del ing eso. A gen ina 1953-1982. Se ie Dis ibución del Ing eso 5. 49 Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1988a). An eceden es es adís icos de la dis ibución del ing eso. Mexico 1950-1977. Se ie Dis ibución del Ing eso 7. Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1988b). An eceden es es adís icos de la dis ibución del ing eso. Venezuela 1957-1985. Se ie Dis ibución del Ing eso 7. Comisión Económica pa a Amé ica La ina y el Ca ibe CEPAL/ECLAC (1989). An eceden es es adís icos de la dis ibución del ing eso. Pe u 1961-1982. Se ie Dis ibución del Ing eso 8 CONADE-CEPAL (1965). Dis ibución del ing eso y cuen as nacionales en la A gen ina. Tomos I-IV. Buenos Ai es. Co nia, G. A. (2010). Income dis ibu ion unde La in Ame ica’s New Le egime. Jou nal o Human De elopmen and Capabili ies. 11(1): 85-114. Co nia, G. A. (edi o .) (2014). Falling Inequali y in La in Ame ica: Policy Changes and Lessons. Ox o d Uni e si y P ess. Coul e , F. A. E., Cowell, F. A. and Jenkins, S. P. (1992). Equi alence scale ela i i ies and he ex en o inequali y and po e y. Economic Jou nal. 102, 1067-1082. Cowell, F. (1980). On he s uc u e o addi i e inequali y measu es. Re iew o Economic S udies. 47(3): 521-531. De Jan y, A. and E. Sadoule (2005). G ow h, po e y and inequali y in La in Ame ica: A causal analysis, 1970–94. Re iew o Income and Weal h. 46(3): 267-287. De Rosa, M., Flo es Beale, I. and Mo gan, M. (2022). Mo e unequal o no as ich? Re isi ing he La in Ame ican excep ion. S one Cen e Wo king Pape 53. De Rosa, M., Lus ig, N., and Ma inez Pabon, V. ( o hcoming). Fiscal edis ibu ion when accoun ing o he ich in La in Ame ica. CEQ Wo king Pape Se ies. Dea on, A. (2005). Measu ing po e y in a g owing wo ld (o measu ing g ow h in a poo wo ld). Re iew o Economics and S a is ics. 81(1): 1–19. Deininge , K. and Squi e, L. (1996). A new da a se measu ing income inequali y. Wo ld Bank Economic Re iew. 10(3): 565–591. Díaz Bazán, T. (2016). Measu ing inequali y om op o bo om. Wo ld Bank Policy Resea ch Wo king Pape 7237. Engbom, N. and Mose , C. (2022). Ea nings inequali y and he minimum wage. E idence om B azil. Ame ican Economic Re iew. 112(12):3803-47. 50 Fai ield, T. and Jo a , M. (2016). Top income sha es, business p o i s, and e ec i e ax a es in con empo a y Chile. Re iew o Income and Weal h., 62(S1): 120–144. Feie he d, G., La oule , P., Long, W. and Lus ig. N. (2023). The pink ide and income inequali y in La in Ame ica. La in Ame ican Poli ics and Socie y. 65(2): 110-144. Fe ei a de Souza, P. and Medei os, M. (2017). The concen a ion o income a he op in B azil 2006-2014. IPEA-Ins i u e o Applied Economic Resea ch Wo king Pape 163. Fe ei a, F., Fi po, S., and Messina, J. (2022). Labo ma ke expe ience and alling ea nings inequali y in B azil: 1995-2012. Wo ld Bank Economic Re iew. 36(1): 37-67. Fe ei a, F., Lei e, P and Li ch ield (2008), The Rise and Fall o B azilian Inequali y: 1981-2004. Mac oeconomic Dynamics 12 (S2): 199-230. Fe ei a, F., Lus ig, N. and Teles, D. (2015) App aising c oss-na ional income inequali y da abases: An in oduc ion. Jou nal o Economic Inequali y. 13: 497–526. Fe ei a, F. Lakne , C. and Silwal, A. (unpublished) Is Income Inequali y Rising E e ywhe e? Unpublished manusc ip , The Wo ld Bank. Fishlow, A. (1972). B azilian Size Dis ibu ion o Income. Ame ican Economic Re iew, 62(1/2): 391–402. Fixle , D., Johnson, D., C aig, A., and Fu long, K. (2017). A consis en da a se ies o e alua e g ow h and inequali y in he na ional accoun s. Re iew o Income and Weal h. 63: S437–S459. Flachai e, E., Lus ig, N. and Vigo i o, A. (2022) ‘Unde epo ing o op incomes and inequali y: A compa ison o co ec ion me hods using simula ions and linked su ey ax da a.’, Re iew o Income and Weal h [P ep in ]. Flo es, I., Sanhueza, C., A ia, J. and Maye , R. (2020) Top incomes in Chile: A his o ical pe spec i e on income inequali y, 1964–2017. Re iew o Income and Weal h. 66(4): 850–874. Fo s e , T., Ken ikelenis, A., Reinsbe g, B., S ubbs, T., King.L. (2019). How s uc u al adjus men p og ams a ec inequali y: A disagg ega ed analysis o IMF condi ionali y, 1980– 2014. Social Science Resea ch. 80: 83-113. Gaspa ini, L. and C uces, G. (2022). The changing pic u e o inequali y in La in Ame ica. E idence o h ee decades. UNDP LAC Wo king Pape 1. Gaspa ini, L., B acco, J., Galeano. L. and Pis o io, M.. (2018). Desigualdad en países en desa ollo: ¿Ajus ando las expec a i as? CEDLAS Documen o de T abajo 224. Gaspa ini, L., C uces, G. and To na olli, L. (2011). Recen ends in income inequali y in La in Ame ica. Economía LACEA Jou nal. 11(2): 147–201. 51 G adín, C. (2021). WIID companion (May 2021): in eg a ed and s anda dized se ies. UN WIDER. G iliches, Z i (1986). Economic da a issues, in Z i G iliches and Michael D. In iliga o , eds., Handbook o Econome ics. Ams e dam: Else ie , ol. 3. Heskia, I. (1980). Dis ibución del ing eso en el G an San iago 1957-1979. Documen o Se ie de In es igación 53, Uni e sidad de Chile, Depa amen o de Economía. Ho man, R. (1971). Con ibuiçao a analisis de dis ibuição de enda e de posse de e a no B asil. Tese de li e docência a Escola de Ag icul u a da Uni e sidade de S. Paulo. Pi acicaba. Jenkins, S. (2015). Wo ld income inequali y da abases: an assessmen o WIID and SWIID. Jou nal o Economic Inequali y. 13(4): 629–671. Langoni, C. (1973) Dis ibuição da Renda e Desen ol imen o Econômico do B asil. Rio de Janei o: Exp essão e Cul u a. Lopez-Cal a, L. F. and Lus ig, N. (2010). Declining Inequali y in La in Ame ica: A Decade o P og ess? B ookings Ins i u ion P ess. López, R., Figue oa, E. and Gu ié ez, P. (2013). La pa e del león: Nue as es imaciones de la pa icipación de los súpe icos en el ing eso de Chile. Se ie Documen os de T abajo de la Uni e sidad de Chile. SDT379. Lus ig, N. (2019). The “missing ich” in household su eys: causes and co ec ion app oaches. Commi men o Equi y (CEQ) Wo king Pape Se ies 75. Lus ig, N. (ed.) (1995). Coping wi h aus e i y: po e y and inequali y in La in Ame ica. The B ookings Ins i u ion, Washing on DC. Lus ig, N., Ma ínez Pabón, V., Neidhö e , G. and Tommasi, M. (2023). Sho and Long-Run Dis ibu ional Impac s o COVID-19 in La in Ame ica. Economía LACEA Jou nal. 22(1): 96–116. Mau izio, R. (2014). El impac o dis ibu i o del sala io mínimo en la A gen ina, el B asil, Chile y el U uguay. Polí icas Sociales 194, CEPAL Mau izio, R. and Vázquez, G. (2016). Dis ibu i e e ec s o minimum wages in La in Ame ica, In e na ional Labou Re iew. 155(1): 97-131. Meye , B., Mok, W. and Sulli an, J. (2015). Household su eys in c isis. Jou nal o Economic Pe spec i es, 29(4): 199–226. Mo gan, M. (2018). Falling inequali y s pe sis en concen a ion: econciling ecen e idence o B azil om su eys, adminis a i e da a and na ional accoun s. PhD Thesis. Pa is School o Economics/École des Hau es É udes en Sciences Sociales. 52 Mo gan, M. and Souza, P. (2019). Dis ibu ing g ow h du ing la e-de elopmen : New e idence on long- un inequali y in B azil. In CLADHE IV. Uni e sidad de San iago, Chile. Na a e e, I. (1960). La dis ibución del ing eso y el desa ollo económico de México. Ins i u o de In es igaciones Económicas, Escuela Nacional de Economía, Uni e sidad Au ónoma de México. México DF. Nolan, B., Rose , M. and Thewissen, S. (2019) GDP pe capi a e sus median household income: Wha gi es ise o di e gence o e ime? Re iew o Income and Weal h 65(3): 465– 494. Ra allion, M. (2003). Measu ing Agg ega e Wel a e in De eloping Coun ies: How Well Do Na ional Accoun s and Su eys Ag ee? Re iew o Economics and S a is ics. 85(3): 645–652. Rossignolo, D., Oli a, N. and Villac eses, N. (2016). Cálculo de la concen ación de los al os ing esos sob e la base de los da os imposi i os: un análisis pa a el Ecuado . Se ie Mac oeconomía del Desa ollo, Naciones Unidas Comisión Económica pa a Amé ica La ina y el Ca ibe (CEPAL). Saez, E. and Zucman, G. (2016). Weal h inequali y in he Uni ed S a es since 1913. E idence om capi alized income ax da a. Qua e ly Jou nal o Economics. 131(2): 519-578. Székely, M. and Hilge , M. (2007). Wha 's Behind he Inequali y We Measu e? An In es iga ion Using La in Ame ican Da a. Ox o d De elopmen S udies. 35(2): 197-217. Wo ld Inequali y Lab (2020). Dis ibu ional Na ional Accoun s Guidelines: Me hods and Concep s Used in he Wo ld Inequali y Da abase. WID.wo ld Wo king Pape se ies Zuñiga-Co de o, A. (2018). Income inequali y in Cos a Rica: An inqui y owa ds dis ibu ional na ional accoun s. Mas e ’s disse a ion, Pa is School o Economics/École des Hau es É udes en Sciences Sociales. Zwijnenbu g, J. (2019). Unequal dis ibu ions: EG DNA e sus DINA App oach. Ame ican Economic Re iew. Pape s and P oceedings. 109: 296–301.