Po ella, Alysson; Gonçal es, Solange Ledi; de Souza, Ped o H. G. Fe ei a; Fi po,
Se gio
Wo king Pape
Ea nings ola ili y in B azil (2012-2023)
Tex o pa a Discussão, No. 3106
P o ided in Coope a ion wi h:
Ins i u e o Applied Economic Resea ch (ipea), B asília
Sugges ed Ci a ion: Po ella, Alysson; Gonçal es, Solange Ledi; de Souza, Ped o H. G. Fe ei a; Fi po,
Se gio (2025) : Ea nings ola ili y in B azil (2012-2023), Tex o pa a Discussão, No. 3106, Ins i u o de
Pesquisa Econômica Aplicada (IPEA), B asília,
h ps://doi.o g/10.38116/ d3106-eng
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/316168
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 ps://c ea i ecommons.o g/licenses/by/2.5/b /
3106
EARNINGS VOLATILITY IN BRAZIL
(2012-2023)
ALYSSON PORTELLAALYSSON PORTELLA
SOLANGE G. GONÇALVESSOLANGE G. GONÇALVES
PEDRO H. G. FERREIRA DE SOUZAPEDRO H. G. FERREIRA DE SOUZA
SERGIO FIRPOSERGIO FIRPO
3106
B asilia, Ap il 2025
EARNINGS VOLATILITY IN BRAZIL
(2012-2023)
ALYSSON PORTELLA1
SOLANGE L. GONÇALVES2
PEDRO H. G. FERREIRA DE SOUZA3
SERGIO FIRPO4
1. Resea che a Inspe . E-mail: alyssonlp1@inspe .edu.b .
2. Assis an p o esso a he Uni e si y o São Paulo (USP). E-mail: solange.gon-
[email p o ec ed].
3. Resea che a he Ins i u e o Applied Economic Resea ch (Ipea). E-mail: ped o. e -
ei [email p o ec ed].b .
4. P o esso o Economics a Inspe . E-mail: se giop 1@inspe .edu.b .
Discussion Pape
A publica ion o dissemina e he indings o esea ch di ec ly o in-
di ec ly conduc ed by he Ins i u e o Applied Economic Resea ch
(Ipea). Due o hei ele ance, hey p o ide in o ma ion o specialis s
and encou age con ibu ions.
© Ins i u e o Applied Economic Resea ch – ipea 2025
Ea nings ola ili y in B azil (2012-2023) / Alysson Po ella... [e al.].
B asília: Ipea, Ab ., 2025.
38 p. : il. – (Discussion Pape ; n. 3106).
Inclui Bibliog a ia.
1. Desigualdade Sala ial. 2. Dis ibuição de Renda. 3. Vola ilidade
dos Rendimen os. 4. B asil. I. Po ella, Alysson. II. Gonçal es, Solange
L. III. Souza, Ped o H. G. Fe ei a de. IV. Fi po, Se gio. V. Ins i u o de
Pesquisa Econômica Aplicada. VI. Tí ulo.
CDD 331.2
Ficha ca alog á ica elabo ada po Elisangela da Sil a Gomes de Macedo CRB-1/1670
How o ci e:
PORTELLA, Alysson e al. Ea nings ola ili y in B azil (2012-2023).
B asília: Ipea, Ab ., 2025. 38 p. (Discussion Pape , n. 3106). DOI: h p://
dx.doi.o g/10.38116/ d3106-eng
JEL: D31; J31; O15; D63.
DOI: h ps://dx.doi.o g/10.38116/ d3106-eng
Ipea publica ions a e a ailable o ee download in PDF (all) and
ePUB (books and pe iodicals).
Access: h ps://www.ipea.go .b /po al/publicacoes
The opinions exp essed in his publica ion a e o exclusi e espon-
sibili y o he au ho s, no necessa ily exp essing he o icial iews
o he Ins i u e o Applied Economic Resea ch and he Minis y o
Planning and Budge .
Rep oduc ion o his ex and he da a con ained wi hin is allowed
as long as he sou ce is ci ed. Rep oduc ion o comme cial pu -
poses is p ohibi ed.
Fede al Go e nmen o B azil
Minis y o Planning and Budge
O ice Simone Nassa Tebe
A public ounda ion a ilia ed o he Minis y o
Planning and Budge , Ipea p o ides echnical and
ins i u ional suppo o go e nmen ac ions –
enabling he o mula ion o nume ous public
policies and p og ams o B azilian de elopmen –
and makes esea ch and s udies conduc ed by i s
s a a ailable o socie y.
P esiden
LUCIANA MENDES SANTOS SERVO
Di ec o o Ins i u ional De elopmen
FERNANDO GAIGER SILVEIRA
Di ec o o S udies and Policies o he S a e,
Ins i u ions and Democ acy
LUSENI MARIA CORDEIRO DE AQUINO
Di ec o o Mac oeconomic S udies and Policies
CLÁUDIO ROBERTO AMITRANO
Di ec o o Regional, U ban and En i onmen al
S udies and Policies
ARISTIDES MONTEIRO NETO
Di ec o o Sec o al S udies and Policies,
o Inno a ion, Regula ion and In as uc u e
FERNANDA DE NEGRI
Di ec o o Social S udies and Policies
RAFAEL GUERREIRO OSÓRIO
Di ec o o In e na ional S udies
KEITI DA ROCHA GOMES
Chie o S a
ALEXANDRE DOS SANTOS CUNHA
Gene al Coo dina o o P ess and
Social Communica ion
GISELE AMARAL DE SOUZA
Ombudsman: h ps://www.ipea.go .b /Ou ido ia
URL: h ps://www.ipea.go .b
CONTENTS
ABSTRACT
1 INTRODUCTION ......................................................................6
2 DATA AND METHODS ............................................................8
2.1 Da a sou ces ................................................................................8
2.2 Sample selec ion ...................................................................... 10
2.3 Measu ing ea nings ola ili y ................................................... 10
3 MAIN FINDINGS .................................................................... 13
3.1 Va iance o a c pe cen changes ............................................. 13
3.2 Dis ibu ion o a c pe cen changes ........................................ 18
4 HETEROGENEITY ANALYSES .............................................20
4.1 Vola ili y by income le el .......................................................... 20
4.2 Employmen -nonemploymen ansi ions ............................... 22
4.3 Type o employmen ................................................................. 25
4.4 Demog aphic sub-g oups ......................................................... 31
5 CONCLUSION ........................................................................35
REFERENCES ............................................................................ 35
ABSTRACT
This a icle p o ides a comp ehensi e analysis o labo ea nings ola ili y in B azil
be ween 2012 and 2023. Du ing his pe iod, B azil’s economy expe ienced in ense
economic g ow h ollowed by la ge ecessions, allowing us o assess changes in ol-
a ili y o e he business cycle. In addi ion, B azil’s na ional household su ey ollows
indi iduals h oughou an en i e yea , co e ing bo h o mal and in o mal sec o s o he
economy. This o e s he pe ec se ing o s udy ea nings ola ili y in de eloping coun-
ies, a subjec ha has been mos ly in es iga ed in he con ex o ad anced economies.
Apa om documen ing he main ola ili y ends, we also in es iga e how hey a y by
income le el, employmen -nonemploymen ansi ions, and ansi ions be ween o mal
and in o mal sec o s. We also assess how ola ili y a ies by gende , educa ion, ace, and
ma i al s a us. Ou esul s show ha ea nings ola ili y is much highe in B azil han
in ich coun ies, especially among low-wage, in o mal wo ke s. T ansi ions in o and
ou o employmen accoun o a la ge sha e o wage ola ili y le els.
Keywo ds: B azil; wage inequali y; income dis ibu ion; ea nings ola ili y.
DISCUSSION PAPER
6
3106
1 INTRODUCTION
Ea nings ola ili y – he le el o luc ua ions in pay o e ime – has been ex ensi ely
s udied in de eloped coun ies, pa icula ly he Uni ed S a es, bu also in o he economies
(Mo i e al., 2023; Mo i and Zhang, 2018; Cappella i and Jenkins, 2014; Jappelli and
Pis a e i, 2010; A am e al., 2022; Li, La and Sologon, 2021; OECD, 2011). In e es in
his opic a ises no only o i s own sake, bu also due o he links be ween ea nings
ola ili y and inequali y and po e y dynamics, consump ion pa e ns, social mobili y,
economic insecu i y, and income isk (Meghi and Pis a e i, 2011; A anasio and Webe ,
2010; Sho ocks, 1978; Wes e n e al., 2012).
Despi e he ex ensi e li e a u e on ola ili y in high-income coun ies, esea ch on
de eloping economies is limi ed. This a icle conduc s a comp ehensi e analysis o
ea nings ola ili y in B azil, a la ge and highly unequal de eloping coun y. The e a e
se e al easons o belie e ha ola ili y is highe in hese coun ies. Income g ow h is
mo e ola ile in eme ging economies (Aguia and Gopina h, 2007), and he p esence o
la ge in o mal ma ke s likely in luences he le el o ea nings isks in hese economies
(La Po a and Shlei e , 2014; Ulyssea, 2020). High labo ma ke u no e and unemploy-
men a e also p e alen (Ge a d and Gonzaga, 2021). Addi ionally, high inequali y and
low mobili y a e cha ac e is ic ea u es o de eloping coun ies, especially La in Ame ica
(Chancel and Pike y, 2021; Messina and Sil a, 2021; B i o e al., 2022).
Ou analysis elies on panel da a om he Con inuous Na ional Household Sam-
ple Su ey (PNADC). We ocus on yea -o e -yea changes in ea nings, using he a i-
ance o he a c pe cen change as ou main measu e o ola ili y. This measu e can
accoun o changes in ea nings a ising om nonemploymen and has been ex en-
si ely used o documen ends in ola ili y in o he coun ies (Shin and Solon, 2011;
Celik e al., 2012; Ziliak, Ha dy and Bollinge , 2011; Dynan, Elmendo and Sichel, 2012;
Mo i e al., 2023). We calcula e sepa a e es ima es o men and women, as hese
g oups display di e en ajec o ies in high-income coun ies (Mo i e al., 2023).
We assess how ola ili y a ies o e he ea nings dis ibu ion in he i s pe iod and he
ole o ansi ions in o and ou o employmen . Ou da a allow us o in es iga e how ol-
a ili y di e s be ween indi iduals wi h o mal, in o mal, and sel -employed a achmen s
o he labo ma ke , which a e impo an aspec s o labo ma ke in de eloping coun ies.
We also in es iga e he e ogenei y by educa ion, ace, and ma i al s a us.
Ou indings sugges ha ea nings ola ili y in B azil is highe han in he Uni ed
S a es and o he ad anced economies. The ola ili y o male ea nings in B azil was
a ound 0.25 in 2015, compa ed o 0.10 in he Uni ed S a es be o e he G ea Recession
(Ziliak, Hokayem and Bollinge , 2023). In e es ingly, in con as o high-income coun-
ies, he ola ili y o emale ea nings in B azil is g ea e han ha o male ea nings.
DISCUSSION PAPER
DISCUSSION PAPER
7
3106
Howe e , his ela ionship was b ie ly e e sed du ing he peak o he co id-19 pandemic.
As documen ed in o he coun ies, ola ili y is coun e cyclical in B azil, which means
ha i inc eases du ing economic down u ns. This e ec is s onge among men. These
esul s a e obus o se e al changes in speci ica ion and sample de ini ion.
We ind much highe ea nings ola ili y among low-income indi iduals. Vola ili y
eaches he lowes le el among hose who ea n close o he minimum wage, sligh ly
inc easing he ea e . To he bes o ou knowledge, we a e he i s o documen how
ola ili y a ies wi h ea nings. We also decompose labo ma ke ola ili y by ansi ions
in o and ou o employmen . This exe cise e eals ha B azil’s high ola ili y is pa ly
d i en by mo e equen pe iods wi h ze o ea nings. When conside ing he ype o a ach-
men o he labo ma ke , we see ha in o mal and sel -employed wo ke s expe ience
he mos ola ile ea nings. This pa e n is p ima ily d i en by he na u e o in o mal
wo k i sel , a he han equen sec o ansi ions, as wo ke s who hold in o mal jobs
in bo h pe iods ha e simila ola ili y as hose who change om o mal o in o mal o
ice e sa. The he e ogenei y analysis indica es ha one eason o he g ea e ea nings
ola ili y among men is he highe p e alence o sel -employmen .
Ou he e ogenei y analyses by educa ional le els, ace, and ma i al s a us con i m
ha whi e indi iduals and hose wi h highe educa ional a ainmen expe ience lowe
ea nings ola ili y. Howe e , he di e ences be ween ma ied and single indi iduals a e
mo e sub le. We do no obse e signi ican di e ences o men, while he esul s o
ma ied and single women depend on whe he we include indi iduals wi h ze o ea nings.
A subs an ial body o esea ch has in es iga ed he dynamics o ea nings ola il-
i y in high-income coun ies, such as he Uni ed S a es, Eu ope, and Aus alia. Ea lie
esea ch ocused on es ima ing he a iance o ansi o y income using pa ame ic o
nonpa ame ic models (Go schalk e al., 1994; Mo i and Go schalk, 2012), which
equi es long longi udinal panel su eys o adminis a i e da a. This li e a u e has been
complemen ed by mo e desc ip i e and anspa en analyses based on he a iance
o ea nings changes be ween wo pe iods, o en elying on sho e panels such as he
Cu en Popula ion Su ey – CPS (Celik e al., 2012; Dahl, Delei e and Schwabish, 2011;
Shin and Solon, 2011; Mo i e al., 2023; Ziliak, Hokayem and Bollinge , 2023). Ou pape
ollows he la e app oach. Mo e ecen ly, a new line o esea ch has ocused on no only
documen ing he a iance bu also highe momen s o ea nings changes, highligh ing how
hese inno a ions a e a om no mally dis ibu ed, showing asymme y and high ku osis
(Gu enen, Ozkan and Song, 2014; Ho mann and Malac ino, 2019; Gu enen e al., 2021;
A ellano, Blundell and Bonhomme, 2017; Busch e al., 2022).
Ou pape complemen s a small li e a u e on ea nings ola ili y in de eloping coun-
ies. Becca ia e al. (2022) in es iga e income mobili y in se en La in Ame ican coun ies.
DISCUSSION PAPER
8
3106
Con a y o us, hey use household income and ocus on indi idual-le el mobili y mea-
su es. San os and Souza (2007) and A abage and Souza (2019) use da a es ic ed o
he B azilian o mal labo ma ke o es ima e models o he a iance o ansi o y and
pe manen incomes. Engbom e al. (2022) documen long- e m ends in ea nings ol-
a ili y using da a on o mal employmen and da a on o mal and in o mal jobs, bu only
encompassing six B azilian me opoli an egions. Ma inez and Mello (2024) in es iga e
how inc eased ade exposu e a ec s he highe momen s o ea nings isk in B azil using
adminis a i e eco ds, which co e only he o mal sec o o he labo ma ke .
Gomes, Iachan and San os (2020) use PNADC da a o in es iga e ea nings changes
in he o mal and in o mal labo ma ke s bu in es iga e highe momen s among wo k-
e s wi h posi i e ea nings only. We complemen hei analysis by including indi idu-
als wi h ze o ea nings and calcula ing g oss measu es o ola ili y. Thus, we unco e
how ansi ions in o and ou o employmen a e much mo e common in B azil han in
de eloped coun ies and con ibu e decisi ely o he o e all highe le els o ola ili y.
In addi ion, we show how ea nings ola ili y is much highe a lowe income le els, likely
a combina ion o wo se labo ma ke a achmen in he o m o highe in o mali y and
mo e likely ansi ions ou o employmen . This inding has impo an policy implica ions,
indica ing ha means- es ed cash ans e s migh be inadequa e o co e indi iduals wi h
high isk o losing hei employmen and alling in o po e y i hey a e a ge ed oo na owly.
This pape p oceeds in he ollowing way. The nex sec ion discusses he da a and
me hodology used. Sec ion 3 p esen s he main esul s. He e ogenei y analyses based
on employmen ansi ion, ype o employmen , educa ion, ace, and ma i al s a us a e
shown in sec ion 4. Sec ion 5 concludes.
2 DATA AND METHODS
2.1 Da a sou ces
We use da a om he PNADC, B azil’s lagship na ional household su ey, conduc ed
since 2012 by he B azilian Ins i u e o Geog aphy and S a is ics (Ins i u o B asilei o
de Geog a ia e Es a ís ica – IBGE).1 The PNADC is a na ionally ep esen a i e o a ing
panel su ey ha collec s in o ma ion on demog aphics, educa ion, he labo ma ke ,
and o he opics. Households a e in e iewed once pe qua e o i e consecu i e
1. PNADC eplaced wo p e ious su eys: he annual Na ional Household Sample Su ey (Pesquisa
Nacional po Amos a de Domicílios – PNAD), a na ionally ep esen a i e, mul i-pu pose, c oss-sec ional
household su ey, and he Mon hly Employmen Su ey (Pesquisa Mensal do Emp ego – PME), a labo
ma ke su ey wi h a o a ing panel ha co e ed only six me opoli an egions. Bo h su eys coexis ed
wi h he PNADC un il 2015 and 2016, espec i ely, when hey we e discon inued.
DISCUSSION PAPER
DISCUSSION PAPER
15
3106
We also conside labo ma ke ola ili y, which includes indi iduals wi h ze o ea n-
ings in one o bo h pe iods. We di e en ia e be ween hese wo g oups in igu e 2.
The solid lines e e only o indi iduals wi h ea nings in a leas one pe iod, while he
dashed lines include indi iduals wi h ze o ea nings in bo h.
FIGURE 2
Labo ma ke ola ili y – B azil (2013-2023)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
2013.1 2014.1 2015.1 2016.1 2017.1 2018.1 2019.1 2020.1 2021.1 2022.1 2023.1 2024.1
Pe iod
Va iance o a c pe cen changes
Men, ze o ea nings in one pe iod
Men, ze o ea nings in bo h pe iods
Women, ze o ea nings in one pe iod
Women, ze o ea nings in bo h pe iods
Au ho s’ elabo a ion.
Obs.: Labo ma ke ola ili y is measu ed using equa ion (1) o men and women sepa a ely be ween
2013 and 2023, using da a om PNADC. Solid lines include indi iduals wi h a leas one pe iod
o posi i e ea nings. The dashed lines include indi iduals wi h ze o ea nings in bo h pe iods.
We adjus ola ili y by age (quad a ic) and qua e ixed e ec s and im he bo om and op
1% posi i e ea nings each qua e . Labels on he ho izon al axes co espond o he inal yea
o each panel coho .
Labo ma ke ola ili y signi ican ly su passes ea nings ola ili y o bo h gende s
in B azil. The ola ili y o men was a ound 0.8 a he beginning o he pe iod and began
o inc ease du ing 2015. I peaked a he onse o he pandemic, which inaugu a ed a
pe iod o apid luc ua ions wi h swi declines and ises, p esumably mi o ing changes
in employmen a es, a opic we will explo e u he when decomposing ea nings ola il-
i y ac oss job ansi ions. In he las qua e o 2023, labo ma ke ola ili y eached he
same le els as a he beginning o he pe iod. Simila o he indings in Ziliak, Hokayem
and Bollinge (2023) o he Uni ed S a es, labo ma ke ola ili y in B azil exhibi s a
mo e p onounced coun e cyclical pa e n compa ed o ea nings ola ili y. Speci ically,
DISCUSSION PAPER
16
3106
including indi iduals wi h no ea nings esul s in a s onge inc ease in ola ili y du ing
economic down u ns han when we exclude his g oup om ou analysis.
Con a y o ea nings ola ili y, labo ma ke ola ili y is highe o women han
men in B azil, a consequence o gende di e ences in labo ma ke pa icipa ion.
The e was a ma ked downwa d end o women a he beginning o he pe iod,
p obably e lec ing his g oup’s inc easing wo k o ce pa icipa ion a es. This down-
wa d end con inued a e he 2014-2016 ecession, while labo ma ke ola ili y o
men ose sha ply. Bo h g oups expe ienced a pe iod o highe ins abili y du ing he
pandemic. By he end o 2023, labo ma ke ola ili y o men was a ound he same
le el as in 2013, bu women had a modes d op.
Labo ma ke ola ili y is simila o men and women i we include indi iduals wi h
ze o ea nings in bo h pe iods. Because hey ha e ze o a c pe cen changes, ola ili y
dec eases when we include hese indi iduals in he analysis. This is especially ue o
women, as hei pa icipa ion in he wo k o ce is lowe han men. Ne e heless, he
o e all pa e n emains consis en ega dless o he inclusion o hese indi iduals in ou
analysis. Vola ili y declined be ween 2013 and 2015, hen ose amids he ecession,
pla eauing du ing he la e 2010s, ollowed by sudden jumps du ing he co id-19 pan-
demic. Such con as ing u ns did no en ail much cumula i e change when we compa e
2023 o 2013. In e es ingly, women showed sligh ly highe ola ili y a he beginning
o he pe iod, bu male ola ili y ose as e du ing he ecession. Simila ly, we obse e
a as e g ow h in ea nings ola ili y o men i we conside only indi iduals wi h ze o
ea nings in one pe iod.
Ziliak, Hokayem and Bollinge (2023) epo much lowe es ima es o labo ma ke
ola ili y in he Uni ed S a es, which eached 0.4 in he a e ma h o he G ea Recession
o men, much lowe han 0.65, he lowes obse ed alue in B azil a he beginning o
2015. S ill, he same ca ea s men ioned abo e apply o labo ma ke ola ili y. Cappella i
and Jenkins (2014) epo es ima es ha sugges ha ola ili y le els do no change
signi ican ly when impu ed ea nings a e included. In bo h cases, labo ma ke ola ili y
in he Uni ed Kingdom anges om 0.2 o 0.4, much lowe han in B azil. In any case,
di ec compa isons wi h igu es o he Uni ed S a es a e u he complica ed because
we use mon hly a he han annual ea nings. By de aul , he sha e o indi iduals wi h
ze o ea nings in bo h pe iods ends o be highe in ou da a.
We conduc se e al obus ness checks on hese main esul s, all a ailable upon
eques . The esul s do no change i we do no adjus o age. T imming he bo om and
op 5% o ea nings ma ginally educes ola ili y, while no imming ma ginally inc eases
i , bu he e a e no signi ican changes in he esul s, especially in ends. Using in e se
DISCUSSION PAPER
DISCUSSION PAPER
17
3106
p obabili y weigh ing o accoun o a i ion inc eases sligh ly o e all ola ili y, especially
du ing he pandemic, bu ba ely a ec s he main esul s.
Changes in sample selec ion and he de ini ion o ea nings a e mo e consequen ial,
bu he main esul s emain. Res ic ing he sample o wo ke s wi h ac i e pa icipa ion
in he wo k o ce in bo h pe iods signi ican ly educes labo ma ke ola ili y o men and
women. S ill, he e ec s on women a e la ge . Vola ili y becomes smalle o women
han o men du ing mos yea s in his scena io, d opping o a ound 0.5 a he beginning
o he pe iod, hen ising o 0.7 and e u ning o 0.5 a he end o 2023. In any case, he
coun e cyclical pa e n is s ill no iceable. Ei he including one o wo pe iods wi h ze o
ea nings makes almos no di e ence i we exclude indi iduals wi h no pa icipa ion in
he wo k o ce in a leas one pe iod. Ea nings ola ili y is no a ec ed because i equi es
pa icipa ion in he labo ma ke in bo h pe iods. Mo eo e , i is wo h no ing ha he
es ic ion o hose pa icipa ing in he labo ma ke b ings es ima es o labo ma ke
ola ili y in B azil close o ha obse ed in he Uni ed S a es, especially o women.
The mos signi ican de ia ion om ou main esul s occu s when we change he
de ini ion o ea nings and analyze usual a he han e ec i e ea nings. Vola ili y le els
d op conside ably, especially when we conside only indi iduals wi h posi i e ea nings
in bo h pe iods. Albei no en i ely unexpec ed, his disc epancy wa an s u he in es-
iga ion, as i sugges s wo ke s a e qui e accu a e a es ima ing hei a e age long- un
ea nings e en when acing conside able sho - un luc ua ions. This should be s aigh -
o wa d o o mal wage wo ke s, and indeed he PNADC ques ionnai e is designed o
emind esponden s o epo e ec i e wages ne o bonuses and ines. S ill, we would
expec his ask o be conside ably mo e di icul o in o mal and sel -employed wo ke s.
Finally, we also e alua e he e ec o excluding sel -employed wo ke s om ou
es ima es, as is done in some s udies (Shin and Solon, 2011; Ziliak, Ha dy and Bollinge ,
2011; Ziliak, Hokayem and Bollinge , 2023). In his case, all ola ili y measu es decline
in magni ude, bu he coun e cyclical ends do no change. The educ ion in ea n-
ings ola ili y is la ge o men han o women, and hey become simila in magni ude
h oughou he pe iod. This sugges s ha sel -employmen is mo e impo an o men
as a sou ce o ola ili y, some hing we explo e in he nex sec ion. Labo ma ke ol-
a ili y also becomes smalle o men and women. I we conside only one pe iod wi h
ze o ea nings, he educ ion in ola ili y is highe o women, which migh sugges ha
sel -employmen is a mo e empo a y placemen o hem. When ze o ea nings in bo h
pe iods a e conside ed, he educ ion in ola ili y is simila be ween he wo gende s,
and hey con inue o mo e oge he h oughou he pe iod.
DISCUSSION PAPER
18
3106
3.2 Dis ibu ion o a c pe cen changes
To be e unde s and ea nings ola ili y in B azil, we analyze selec ed quan iles o he
dis ibu ion o he a c pe cen changes o men and women, including and excluding ze o
ea nings. Figu e 3 shows a s able pa e n h oughou he pe iod o bo h g oups. Abou
90% o ea nings inno a ions a e wi hin he [−1,1] in e al, and a ound 50% a e close o ze o.
In o he wo ds, 10% o he wo k o ce ypically expe iences uly la ge yea -o e -yea ea nings
luc ua ions. As has al eady been cha ac e ized elsewhe e and in B azil, hese
ea nings inno a ions a e a om ollowing a no mal dis ibu ion (Gomes, Iachan and
San os, 2020; Gu enen e al., 2021; De Na di e al., 2021). De ia ions om his pa e n
occu ed mainly du ing he pandemic. The e was a b ie spike in he le (nega i e) ail o
he dis ibu ion o a c pe cen changes in 2020 and an equally sho -li ed up ick in he igh
(posi i e) ail in 2021. The dis ibu ion o a c pe cen changes is sligh ly mo e comp essed
o women, bu he spikes obse ed du ing he pandemic we e la ge .
FIGURE 3
Quan iles o he a c pe cen change among wo ke s wi h posi i e ea nings – B azil
a) Men b) Women
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1 2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
−2.0
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
2.0
Pe iod
A c pe cen age change
P5 P10 P25 P50 P75 P90 P95
Au ho s’ elabo a ion.
Obs.: These igu es plo he 5 h, 10 h, 25 h, 50 h, 75 h, 90 h, and 95 h pe cen iles o he a c pe cen
change o posi i e ea nings, o men and women sepa a ely be ween 2013 and 2023, using
da a om PNADC. We adjus a c pe cen changes by age (quad a ic) and qua e ixed e ec s
and im he bo om and op 1% posi i e ea nings each qua e . Labels on he ho izon al axes
co espond o he inal yea o each panel coho .
DISCUSSION PAPER
DISCUSSION PAPER
19
3106
Figu e 4 displays he dis ibu ion o a c pe cen changes when we also include indi-
iduals wi h ze o ea nings in bo h pe iods. In his case, he dis ibu ion o inno a ions has
much hea ie ails. Fo bo h men and women, he op and bo om 5% o he dis ibu ion
eached he maximum o ±2 h oughou almos he en i e pe iod. Du ing he pandemic,
shocks leading o no ea nings (hence, an a c pe cen change o -2) occu ed o a leas
10% o men and women. We can also obse e ha he in e qua ile ange (p25-p75) is
mo e comp essed o women han men.
FIGURE 4
Quan iles o he a c pe cen change including indi iduals wi h ze o ea nings in
bo h pe iods – B azil
a) Men b) Women
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1 2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
−2.0
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
2.0
Pe iod
A c pe cen age change
P5 P10 P25 P50 P75 P90 P95
Au ho s’ elabo a ion.
Obs.: These igu es plo he 5 h, 10 h, 25 h, 50 h, 75 h, 90 h, and 95 h pe cen iles o he a c pe cen change
o all indi iduals, o men and women sepa a ely be ween 2013 and 2023, using da a om
PNADC. The sample includes indi iduals wi h ze o ea nings in bo h pe iods. We adjus a c
pe cen changes by age (quad a ic) and qua e ixed e ec s and im he bo om and op 1%
posi i e ea nings each qua e . Labels on he ho izon al axes co espond o he inal yea o
each panel coho .
The esul s shown in igu es 3 and 4 a e mo e ex eme han he es ima es p esen ed
by Shin and Solon (2011) and Cappella i and Jenkins (2014). The la e epo alues
a ound ± 0.5 o he P5 and P95 quan iles o changes in posi i e ea nings in B i ain,
bo h o men and women, ep esen ing a c pe cen changes o 50% in ea nings. Simila
di e ences occu when we also include indi iduals wi h ze o ea nings. In B azil, P10
DISCUSSION PAPER
20
3106
and P90 we e a ound ± 1.0 mos o he ime, meanings ha nea ly 20% o he popu-
la ion expe ience a c pe cen changes la ge han 100%, whe eas his almos ne e
happens in he Uni ed S a es o he Uni ed Kingdom. These con as s show ha he
highe ea nings ola ili y in B azil does no a ise only om mo e likely changes o and
om nonemploymen bu also due o changes in ea nings.
4 HETEROGENEITY ANALYSES
4.1 Vola ili y by income le el
We begin by in es iga ing di e ences in ola ili y le els by ea nings. We classi y wo ke s
by en iles o ea nings in he i s pe iod, while adding an ex a g oup o all indi iduals
wi h ze o ea nings in he i s pe iod (pe cen ile 0). We measu e ola ili y in each g oup
as he sha e o indi iduals wi h a c pe cen changes la ge han 50% in ei he di ec ion.
Figu e 5 shows he esul s o he agg ega ed samples o 2017 and 2018, o men
and women sepa a ely.5 Fo bo h g oups, we obse e highe ola ili y in he lowe ail
o he ea nings dis ibu ion, pla eauing a ound he 30 h pe cen ile, wi h a sligh upwa d
end as we mo e up he dis ibu ion. O e all, a ound 20% o he sample expe ienced
a c pe cen changes la ge han 50%.6 Resul s a e quali a i ely he same i we conside
o he pe iods and h esholds.
5. We p esen esul s o 2017 and 2018 o a oid noise in oduced by he co id-19 pandemic, bu esul s
a e quali a i ely simila o o he pe iods, as no ed.
6. Nea ly 20% o indi iduals wi h ze o ea nings ge a job in he ollowing yea , hence hei a c pe cen
change is 200%. I we es ic he sample o indi iduals wi h ze o ea nings in one pe iod only, hen all o
hose wi h ze o ea nings in he i s pe iod will necessa ily ha e posi i e ea nings one yea la e .
DISCUSSION PAPER
DISCUSSION PAPER
21
3106
FIGURE 5
Sha e o indi iduals wi h a c pe cen changes highe han 50% by ea nings in
he i s pe iod – B azil (2017 and 2018)
a) Men b) Women
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
0
10
20
30
40
50
60
70
80
90
100
Pe cen ile
Sha e o indi iduals (%)
Posi i e ea nings Including ze o ea nings in one pe iod Including ze o ea nings in bo h pe iods
Au ho s’ elabo a ion.
Obs.: Each line shows he sha e o indi iduals wi h a c pe cen changes abo e 50%, nega i e o
posi i e. We conside h ee g oups: hose wi h posi i e ea nings in bo h pe iods; also including
hose wi h ze o ea nings in one pe iod; and also including hose wi h ze o ea nings in bo h
pe iods. Indi iduals wi h posi i e in he i s a e g ouped in o 20 equally-size bins o app oxi-
ma ely 5% o he weigh ed sample, while hose wi h ze o ea nings a e placed in he “ze o” bin.
We also in es iga e he ola ili y o ea nings by analyzing he dispe sion o a c pe -
cen changes o he same g oup o indi iduals using boxplo s ( igu e 6). In his case,
we only plo he esul s o measu es including indi iduals wi h posi i e ea nings and
also indi iduals wi h ze o ea nings in one pe iod. Again, we obse e a much la ge dis-
pe sion in he a c pe cen change among low-wage ea ne s, o men and women alike.
The minimum le el o dispe sion is obse ed be ween he 30 h and 40 h pe cen ile. This
is he posi ion in he ea nings dis ibu ion whe e he minimum wage becomes binding.
A e his pe iod, dispe sion inc eases, bu o mode a e le els when compa ed o he
dispe sion obse ed below he 30 h pe cen ile. The pa e ns obse ed o 2017 and 2018
do no change when compa ed wi h 2013 and 2014 ( esul s no shown).
DISCUSSION PAPER
22
3106
FIGURE 6
Dis ibu ion o a c pe cen changes – B azil (2017 and 2018)
a) Men b) Women
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95100
−2
−1
0
1
2
Pe cen ile
A c pe cen changes
Posi i e ea nings Including ze o ea nings in one pe iod
Au ho s’ elabo a ion.
Obs.: Boxes plo he median, in e quan ile ange, and he whiske s co espond o 1.5 imes he
in e quan ile ange. We conside indi iduals wi h posi i e ea nings and ze o ea nings in one
pe iod. Indi iduals wi h posi i e in he i s a e g ouped in o 20 equally-size bins o app oxi-
ma ely 5% o he weigh ed sample.
4.2 Employmen -nonemploymen ansi ions
We ollow Ziliak, Ha dy and Bollinge (2011) and Cappella i and Jenkins (2014) and
decompose he a iance o a c pe cen change based on employmen ansi ions. We
classi y wo ke s in o ou g oups: employed in bo h pe iods (11), nonemployed in bo h
pe iods (00), employed in he i s pe iod bu nonemployed in he second (10), and
nonemployed in he i s bu employed in he second (01).
Figu e 7 shows he esul s o he decomposi ion in equa ion (2) o men. Panel (a)
shows he o al a iance o a c pe cen changes, as well as how much each componen
con ibu es o i . T ansi ions in o and ou o employmen and ,
espec i ely) accoun o he la ges sha e o ola ili y, as well as o mos changes in he
pe iod. These changes include he inc ease in ola ili y a e 2015 and he luc ua ions
du ing he pandemics. Va iance a ising om hose who emained employed accoun s
o less han 25% o o al ola ili y h oughou mos o he pe iod, e en hough mos
men a e always employed – panel (b).
DISCUSSION PAPER
DISCUSSION PAPER
23
3106
FIGURE 7
Decomposi ion o labo ma ke ola ili y o men – B azil
0.0
0.2
0.4
0.6
0.8
1.0
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
Pe iod
Va iance
To al
P01(2-M)2
P11V11
2
P10(2+M)
P00M2
2
P11(M11-M)
a) Va iance decomposi ion
0
10
20
30
40
50
60
70
80
90
100
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
Pe iod
Pe cen age
P01
P11
P10
P00
100M
11
b) Employmen sha e
Au ho s’ elabo a ion.
Obs.: Panel (a) shows he o al a iance o a c pe cen changes and he espec i e componen s
es ima ed using he decomposi ion in equa ion (2). Panel (b) shows he sha e o each employ-
men ansi ion g oup in each pe iod. The g oups a e: employed in bo h pe iods (11), nonem-
ployed in bo h pe iods (00), o ansi ioned om employmen o non-employmen (10) and
om nonemploymen o employmen (01). We es ic he sample o men only. The sample
includes indi iduals wi h ze o ea nings in bo h pe iods. We adjus a c pe cen changes by
age (quad a ic) and qua e ixed e ec s and im he bo om and op 1% posi i e ea nings
each qua e . Labels on he ho izon al axes co espond o he inal yea o each panel coho .
These esul s pa ially con as wi h hose ound by Ziliak, Ha dy and Bollinge (2011)
and Cappella i and Jenkins (2014) o he Uni ed S a es and he Uni ed Kingdom, espec-
i ely. The o me shows ha ola ili y among con inuously employed wo ke s accoun s
o he majo i y o labo ma ke ola ili y in he Uni ed S a es, especially up o he 1990s.
The la e ind ha ansi ions ou o and in o employmen accoun o a la ge sha e o
labo ma ke ola ili y in he Uni ed Kingdom, bu hei impo ance dec eased in mo e
ecen pe iods and con e ged o le els simila o hose o con inuous wo ke s.
These di e ences be ween B azil and he Uni ed S a es and he Uni ed Kingdom a e
la gely explained by he mo e common ansi ions ou o and in o employmen in B azil
(panel b). While in he Uni ed S a es and he Uni ed Kingdom, be ween 80% and 90% o
male wo ke s emain employed in bo h pe iods, in B azil hese igu es ange om 60%
o 80%. App oxima ely 10% o men ansi ion o and ou o employmen in B azil, while
hese p opo ions a e a ound 5% in he Uni ed S a es and he Uni ed Kingdom.
DISCUSSION PAPER
24
3106
Figu e 8 shows he same esul s o women. They a e somewha simila o men
i we conside he main con ibu o s o labo ma ke ola ili y. Fo women, ansi ions
in o and ou o employmen accoun o mos o he ola ili y, and hei sha e is la ge
han ha obse ed o men. The main con as is o pa icipa ion in he labo ma ke .
The pe cen age o women employed in bo h pe iods is much lowe han ha o men,
anging be ween 40% and 50%. Mo eo e , nea ly 40% o women emain nonemployed
in bo h pe iods, due o hei lowe wo k o ce pa icipa ion a es. Women’s pa icipa ion in
he labo ma ke is highe in he Uni ed S a es and Uni ed Kingdom, app oxima ely 60%
o 70%, while nonpa icipa ion anges om 20% o 25% (Ziliak, Ha dy and Bollinge , 2011;
Cappella i and Jenkins, 2014). T ansi ions in o and ou o employmen a e also highe
o B azilian women in con as o he Uni ed S a es o he Uni ed Kingdom. In B azil, i
is a ound 10%, while in hese coun ies i is always below 10%.
FIGURE 8
Decomposi ion o labo ma ke ola ili y o women – B azil
0.0
0.2
0.4
0.6
0.8
1.0
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
Pe iod
Va iance
a) Va iance decomposi ion
0
10
20
30
40
50
60
70
80
90
100
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
Pe iod
Pe cen age
b) Employmen sha e
To al
P
01
(2-M)2
P11V11
2
P
10
(2+M)
P00M2
2
P
11
(M
11
-M)
P01
P11
P
10
P00
100M
11
Au ho s’ elabo a ion.
Obs.: Panel (a) shows he o al a iance o a c pe cen changes and he espec i e componen s
es ima ed using he decomposi ion in equa ion (2). Panel (b) shows he sha e o each employ-
men ansi ion g oup in each pe iod. The g oups a e: employed in bo h pe iods (11), nonem-
ployed in bo h pe iods (00), o ansi ioned om employmen o non-employmen (10) and
om nonemploymen o employmen (01). We es ic he sample o men only. The sample
includes indi iduals wi h ze o ea nings in bo h pe iods. We adjus a c pe cen changes by
age (quad a ic) and qua e ixed e ec s and im he bo om and op 1% posi i e ea nings
each qua e . Labels on he ho izon al axes co espond o he inal yea o each panel coho .
DISCUSSION PAPER
DISCUSSION PAPER
31
3106
4.4 Demog aphic sub-g oups
Figu e 14 displays he labo ma ke ola ili y among all wo ke s and by h ee educa ional
g oups: indi iduals wi h less han high school, high school g adua es, and college g ad-
ua es. Mo e educa ed indi iduals ace smalle ola ili y in hei ea nings. The a iance
o he a c pe cen changes o male college-educa ed wo ke s is a ound 0.5 h oughou
he pe iod, al hough i became mo e ola ile du ing he pandemic. Among wo ke s who
only comple ed high school, he co esponding igu e was a ound 0.6 in he ea ly 2010s,
hen inc eased o e he decade, peaking a a ound 0.8 du ing he pandemic. Wo ke s
wi h less han high school s a ed a a highe le el (a ound 0.8), bu ollowed a simila
end. In o he wo ds, college-educa ed wo ke s a e be e in income le els (as hey
ypically ea n mo e han less-educa ed wo ke s in he c oss-sec ion) and income s a-
bili y, especially du ing economic down u ns.
FIGURE 14
Labo ma ke ola ili y by educa ion le el – B azil
a) Men b) Women
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1 2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Pe iod
Labo ma ke ola ili y
To al Less han HS High School College
Au ho s’ elabo a ion.
Obs.: We classi y indi iduals based on hei educa ional a ainmen . The black solid line p esen s
esul s o all indi iduals. Labels on he ho izon al axes co espond o he inal yea o each
panel coho .
DISCUSSION PAPER
32
3106
Fo women, labo ma ke ola ili y is also lowe among mo e educa ed wo ke s
( igu e 14), bu di e ences be ween high school d opou s and high school g adua es
a e small, al hough hey widened a bi in mo e ecen yea s. In his case, lowe ola ili y
among high school d opou s likely e lec s dec easing labo ma ke pa icipa ion among
he leas educa ed women, as he e is no labo ma ke ola ili y o indi iduals who a e
ne e employed. College-educa ed women expe ienced labo ma ke ola ili y a ound
0.6 o mos o he pe iod, while ola ili y luc ua ed be ween 0.7 and 0.8 o he o he
educa ional g oups.
The ela i e s abili y o emale labo ma ke ola ili y is d i en by an inc ease in he
componen explained by college and high school g adua es and a decline in he sha e
accoun ed by high school d opou s. This is o a la ge ex en he esul o a la ge sha e o
high school and college g adua es in he popula ion.
Ziliak, Ha dy and Bollinge (2011) also epo di e ences in labo ma ke ola il-
i y in he Uni ed S a es be ween educa ional g oups. Thei es ima es indica e simila
le els o ola ili y o high school and college g adua es, while high school d opou s
display la ge ola ili y. This holds o bo h men and women. Hence, while a high school
diploma seems o p o ide simila p o ec ion agains ea nings isk o a college deg ee in
he Uni ed S a es, in B azil he labo ma ke ola ili y o high school g adua es is la ge
compa ed o college g adua es.
B azil is also ma ked by huge inequali ies be ween Whi e and non-Whi e indi id-
uals, and he wo se ou comes o non-Whi e wo ke s in he labo ma ke a e likely o
mani es also in e ms o highe ea nings ola ili y. Ziliak, Ha dy and Bollinge (2011)
ind ha ola ili y le els o Black men and women in he Uni ed S a es a e la ge han
hei Whi e coun e pa s, especially be o e he 2000s. Figu e 15 p esen s es ima es o
labo ma ke ola ili y by acial g oups. Non-Whi e wo ke s expe ience mo e ola ili y
han Whi es, o men and women alike. Fo men, non-Whi e ola ili y was a ound 0.7
and 0.8 in he ea ly pe iod, while i was a ound 0.6 o Whi es. A e he ecession in
2015, ola ili y inc eases o bo h g oups, bu as e o non-Whi es, eaching a ound 0.9
among hem. Fo Whi es, ola ili y emained below 0.7 be o e he pandemic. Du ing he
pandemic, bo h acial g oups expe imen simila swings, wi h non-Whi e men always
displaying highe ola ili y.
DISCUSSION PAPER
DISCUSSION PAPER
33
3106
FIGURE 15
Labo ma ke ola ili y by acial g oups – B azil
a) Men b) Women
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1 2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
0.0
0.2
0.4
0.6
0.8
1.0
Pe iod
Labo ma ke ola ili y
Whi es Non-Whi es
Au ho s’ elabo a ion.
Obs.: Panel (a) es ic s he sample o men and panel (b) o women. The sample includes indi idu-
als wi h ze o ea nings in bo h pe iods. We adjus a c pe cen changes by age (quad a ic) and
qua e ixed e ec s and im he bo om and op 1% posi i e ea nings each qua e . Labels
on he ho izon al axes co espond o he inal yea o each panel coho .
Panel (b) in igu e 15 shows ha labo ma ke ola ili y is also highe o non-Whi e
han Whi e women, al hough he acial gap is smalle han o men. Bo h g oups display
oughly cons an ola ili y le els p io o he pandemic, wi h small changes du ing he
2015 ecession. The acial gap in ola ili y le els na owed b ie ly du ing he pandemic
bu widened a e wa d.
Finally, we in es iga e labo ma ke ola ili y by ma i al s a us, as amilies a e
isk-pooling o ganiza ions ha help bu e ad e se ea nings shocks. Ma iage o cohabi-
a ion s a us8 ma e s because couples may make in e dependen decisions o s abilize
wel a e, as illus a ed by he “added wo ke e ec ” when spouses ei he en e he labo
o ce o wo k longe hou s o o se ea nings losses esul ing om he in olun a y
unemploymen o he p ima y ea ne (Wes e n e al., 2012).
8. In his sec ion, we e e o ma iage and cohabi a ion in e changeably, ha is, we de ine couples
ega dless o he ela ionship’s legal s a us.
DISCUSSION PAPER
34
3106
Figu e 16 compa es le els and ends in labo ma ke ola ili y be ween g oups
de ined by ela ionship s a us. We p esen ola ili y measu es o indi idual ea nings o
singles and ma ied indi iduals sepa a ely and add a se ies o couples based on he
sum o ea nings o bo h pa ne s.9 Resul s a e simila o single and ma ied men, as
well as couples. Es ima es a e noisie o single men – possibly due o smalle sample
sizes – bu he end is he same as o ma ied men and couples, wi h ising ola ili y
be ween 2015 and he pandemic. Single and ma ied women also ha e e y simila
ola ili y le els and ends.
FIGURE 16
Labo ma ke ola ili y by ma i al s a us – B azil
a) Men b) Women
2013.1 2015.1 2017.1 2019.1 2021.1 2023.1 2013.1 2015.1 2017.1 2019.1 2021.1 2023.1
0.0
0.2
0.4
0.6
0.8
1.0
Pe iod
Labo ma ke ola ili y
Single Ma ied Couples
Au ho s’ elabo a ion.
Obs.: Panel (a) es ic s he sample o men and panel (b) o women. The sample includes indi idu-
als wi h ze o ea nings in bo h pe iods. We adjus a c pe cen changes by age (quad a ic) and
qua e ixed e ec s and im he bo om and op 1% posi i e ea nings each qua e . Labels
on he ho izon al axes co espond o he inal yea o each panel coho .
The e a e some con as s be ween ou esul s and analyses o he Uni ed S a es
in Ziliak, Ha dy and Bollinge (2011). They show ha labo ma ke ola ili y is highe o
unma ied men h oughou he pe iod hey in es iga ed. Unma ied women also display
la ge ea nings luc ua ions, bu less so. In B azil, we see ha he ea nings ola ili y o
ma ied and single men and women a e gene ally simila .
9. By de ini ion, he ola ili y o couples is he same o bo h men and women.
DISCUSSION PAPER
DISCUSSION PAPER
35
3106
5 CONCLUSION
In his s udy, we ha e p o ided a comp ehensi e analysis o ea nings ola ili y in B azil,
a highly unequal de eloping coun y. Ou indings indica e ha ea nings ola ili y in
B azil is signi ican ly highe han in he Uni ed S a es and o he ad anced economies, a
pa e n ha is exace ba ed by he la ge in o mal labo ma ke and high labo u no e .
In pa icula , he ola ili y o male ea nings exceeds ha o emale ea nings, di e ging
om he ypical pa e n obse ed in high-income coun ies, al hough his esul e e ses
when we inco po a e pe iods o ze o ea nings in he analysis. In addi ion, ou esul s
co obo a e p e ious indings on he coun e cyclical na u e o ea nings ola ili y, wi h
la ge income luc ua ions du ing ecessions and he co id-19 pandemic.
Ou da a allowed us no only o documen he g oss measu es o ea nings ola ili y
o he whole popula ion bu also o explo e he nuances in oduced by di e en labo
ma ke a achmen s ha a e mo e common in eme ging economies. We obse ed ha
in o mal and sel -employed wo ke s expe ience he mos signi ican luc ua ions in
ea nings, sugges ing ha he na u e o in o mal wo k con ibu es subs an ially o he
obse ed ola ili y. Mo eo e , we show how low-wage ea ne s a e signi ican ly mo e
exposed o labo ma ke ola ili y. This is ele an in o ma ion o he design o public
sa e y ne s in de eloping coun ies. Fu he mo e, ou he e ogenei y analysis shows
how ea nings ola ili y a ies ac oss educa ional, acial, and ma i al lines, e ealing less
ola ili y among indi iduals wi h highe educa ional a ainmen and Whi e indi iduals,
while ma i al s a us showed mo e nuanced e ec s.
REFERENCES
AGUIAR, M.; GOPINATH, G. Eme ging ma ke business cycles: he cycle is he end.
Jou nal o Poli ical Economy, . 115, n. 1, p. 69-102, Feb. 2007.
ARABAGE, A. C.; SOUZA, A. P. Wage dynamics and inequali y in he B azilian o mal
labo ma ke . EconomiA, . 20, n. 3, p. 153-190, Sep .-Dec. 2019.
ARELLANO, M.; BLUNDELL, R.; BONHOMME, S. Ea nings and consump ion dynamics: a
nonlinea panel da a amewo k. Econome ica, . 85, n. 3, p. 693-734, May 2017.
ATTANASIO, O. P.; WEBER. G. Consump ion and sa ing: models o in e empo al allo-
ca ion and hei implica ions o public policy. Jou nal o Economic Li e a u e, . 48,
n. 3, p. 693-751, Sep . 2010.
AVRAM, S. e al. Household ea nings and income ola ili y in he UK, 2009-2017. The
Jou nal o Economic Inequali y, . 20, n. 2, p. 345-369, 2022.
DISCUSSION PAPER
36
3106
BECCARIA, L. e al. Sho - e m income mobili y in La in Ame ica in he 2000s: in ensi y
and cha ac e is ics. Socio-Economic Re iew, . 20, n. 3, p. 1039-1067, July 2022.
BOLLINGER, C. R. e al. T ouble in he ails? Wha we know abou ea nings non esponse
30 yea s a e Lilla d, Smi h, and Welch. Jou nal o Poli ical Economy, . 127, n. 5,
p. 2143-2185, Oc . 2019.
BRITTO, D. G. e al. In e gene a ionalmobili y in he land o inequali y. Bonn: IZA, Sep .
2022. (Discussion Pape , n. 15611).
BUSCH, C. e al. Skewed idiosync a ic income isk o e he business cycle: sou ces
and insu ance. Ame ican Economic Jou nal: Mac oeconomics, . 14, n. 2, p. 207-242,
Ap . 2022.
CAPPELLARI, L.; JENKINS, S. P. Ea nings and labou ma ke ola ili y in B i ain, wi h a
ansa lan ic compa ison. Labou Economics, . 30, p. 201-211, Oc . 2014.
CELIK, S. e al. Recen ends in ea nings ola ili y: e idence om su ey and adminis-
a i e da a. The BE Jou nal o Economic Analysis & Policy, . 12, n. 2, June 2012.
CHANCEL, L.; PIKETTY, T. Global income inequali y, 1820-2020: he pe sis ence and-
mu a ion o ex eme inequali y. Jou nal o he Eu opean Economic Associa ion, . 19,
n. 6, p. 3025-3062, Dec. 2021.
DAHL, M.; DELEIRE, T.; SCHWABISH, J. A. Es ima es o yea - o-yea ola ili y in ea ning-
sand in household incomes om adminis a i e, su ey, and ma ched da a. Jou nal o
Human Resou ces, . 46, n. 4, p. 750-774, 2011.
DE NARDI, M. e al. Family and go e nmen insu ance: wage, ea nings, and income isks
in he Ne he lands and he U.S. Jou nal o Public Economics, . 193, Jan. 2021.
DYNAN, K.; ELMENDORF, D.; SICHEL, D. The e olu ion o household income ola ili y.
The BE Jou nal o Economic Analysis & Policy, . 12, n. 2, Dec. 2012.
ENGBOM, N. e al. Ea nings inequali y and dynamics in he p esence o in o mali y: he
case o B azil. Quan i a i e Economics, . 13, n. 4, p. 1405-1446, No . 2022. Re ie ed
om: h ps://onlinelib a y.wiley.com/doi/abs/10.3982/QE1855.
FIRPO, S.; PORTELLA, A. L. The labo ma ke in B azil, 2001-2022. IZA Wo ld o Labo ,
Ap . 2024.
GERARD, F.; GONZAGA, G. In o mal labo and he e iciency cos o social p og ams:
e idence om unemploymen insu ance in B azil. Ame ican Economic Jou nal: Eco-
nomic Policy, . 13, n. 3, p. 167-206, Aug. 2021.
DISCUSSION PAPER
DISCUSSION PAPER
37
3106
GOMES, D. B.; IACHAN, F. S.; SANTOS, C. Labo ea nings dynamics in a de eloping
economy wi h a la ge in o mal sec o . Jou nal o Economic Dynamics and Con ol,
. 113, Ap . 2020.
GOTTSCHALK, P. e al. The g ow h o ea nings ins abili y in he US labo ma ke . B ook-
ings Pape s on Economic Ac i i y, . 25, n. 2, p. 217-272, 1994.
GUVENEN, F.; OZKAN, S.; SONG, J. The na u e o coun e cyclical income isk. Jou nal
o Poli ical Economy, . 122, n. 3, p. 621-660, June 2014.
GUVENEN, F. e al. Wha do da a on millions o U.S. wo ke s e eal abou li ecycle ea n-
ings dynamics? Econome ica, . 89, n. 5, p. 2303-2339, Sep . 2021.
HECKSHER, M. D. Inaudí eis: quem deixou de se cobe o pelas pesquisas ele ônicas
e di e enças em elação aos egis os adminis a i os. B asilia: Ipea, June 2021. (No a
Técnica Disoc, n. 95).
HOFFMANN, E. B.; MALACRINO, D. Employmen ime and he cyclicali y o ea nings
g ow h. Jou nal o Public Economics, . 169, p. 160-171, Jan. 2019.
JAPPELLI, T.; PISTAFERRI, L. Does consump ion inequali y ack income inequali y in
i aly? Re iew o Economic Dynamics, . 13, n. 1, p. 133-153, Jan. 2010.
LA PORTA, R.; SHLEIFER, A. In o mali y and de elopmen . Jou nal o Economic Pe -
spec i es, . 28, n. 3, p. 109-126, 2014.
LI, J.; LA, H. A.; SOLOGON, D. M. Policy, demog aphy, and ma ke income ola ili y: wha
shaped income dis ibu ion and inequali y in Aus alia be ween 2002 and 2016? Re iew
o Income and Weal h, . 67, n. 1, p. 196-221, Ma . 2021.
MARTINEZ, T. R.; MELLO, U. T ade shocks and highe -o de ea nings isk in local labo
ma ke s. In e na ional Economic Re iew, . 65, n. 4, No . 2024. Re ie ed om: h ps://
onlinelib a y.wiley.com/doi/10.1111/ie e.12703.
MEGHIR, C.; PISTAFERRI, L. Ea nings, consump ion and li e cycle choices. In: CARD,
D.; ASHENFELTER, O. (Ed.). Handbook o labo economics. Ams e dam: Else ie , 2011,
. 4, p. 773-854.
MESSINA, J.; SILVA, J. Twen y yea s o wage inequali y in La in Ame ica. The Wo ld
Bank Economic Re iew, . 35, n. 1, p. 117-147, Feb. 2021.
MOFFITT, R. e al. Reconciling ends in US male ea nings ola ili y: esul s om su ey and
adminis a i e da a. Jou nal o Business & Economic S a is ics, . 41, n. 1, p. 1-11, 2023.
DISCUSSION PAPER
38
3106
MOFFITT, R.; GOTTSCHALK, P. T ends in he ansi o y a iance o male ea nings: me h-
ods and e idence. Jou nal o Human Resou ces, . 47, n. 1, p. 204-236, 2012.
MOFFITT, R.; ZHANG, S. Income ola ili y and he PSID: pas esea ch and new esul s.
AEA Pape s and P oceedings, . 108, p. 277-280, May 2018.
OECD – ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT. Ea n-
ings ola ili y: causes and consequences. In: OECD – ORGANISATION FOR ECONOMIC
CO-OPERATION AND DEVELOPMENT. OECD Employmen Ou look 2011. Pa is: OECD
Publishing, 2011. Re ie ed om: h ps://www.oecd.o g/en/publica ions/oecd-employ-
men -ou look-2011_empl_ou look-2011-en.h ml.
OSORIO, R. G. Sob e a mon agem e a iden i icação dos painéis da PNAD Con ínua.
Me cado de T abalho: conjun u a e análise, B asilia, n. 73, p. 69-81, Ap . 2022.
SANTOS, A. L.; SOUZA, A. P. Ea nings inequali y in B azil: is i pe manen o ansi o y?
B azilian Re iew o Econome ics, . 27, n. 2, p. 259-284, No . 2007.
SHIN, D.; SOLON, G. T ends in men’s ea nings ola ili y: wha does he panel s udy o
income dynamics show? Jou nal o Public Economics, . 95, n. 7-8, p. 973-982, Aug. 2011.
SHORROCKS, A. Income inequali y and income mobili y. Jou nal o Economic Theo y,
. 19, n. 2, p. 376-393, Dec. 1978.
ULYSSEA, G. In o mali y: causes and consequences o de elopmen . Annual Re iew
o Economics, . 12, p. 525-546, 2020.
WESTERN, B. e al. Economic insecu i y and social s a i ica ion. Annual Re iew o
Sociology, . 38, p. 341-359, 2012.
ZILIAK, J. P.; HARDY, B.; BOLLINGER, C. Ea nings ola ili y in Ame ica: e idence om
ma ched CPS. Labou Economics, . 18, n. 6, p. 742-754, Dec. 2011.
ZILIAK, J. P.; HOKAYEM, C.; BOLLINGER, C. R. T ends in ea nings ola ili y using linked
adminis a i e and su ey da a. Jou nal o Business & Economic S a is ics, . 41, n. 1,
p. 12-19, 2023.
Ipea – Ins i u e o Applied Economic Resea ch
PUBLISHING DEPARTMENT
Head o he Publishing Depa men
Ae omilson T ajano de Mesqui a
Assis an s o he Head o he Depa men
Ra ael Augus o Fe ei a Ca doso
Samuel Elias de Souza
Supe ision
Ana Cla a Escó cio Xa ie
E e son da Sil a Mou a
Typese ing
Ande son Sil a Reis
Augus o Lopes dos San os Bo ges
C is iano Fe ei a de A aújo
Daniel Al es Ta a es
Danielle de Oli ei a Ay es
Leona do Hideki Higa
Vanessa Viei a
Co e design
Aline C is ine To es da Sil a Ma ins
G aphic design
Aline C is ine To es da Sil a Ma ins
The manusc ip s in languages o he han Po uguese
published he ein ha e no been p oo ead.
Ipea – B asilia
Se o de Edi ícios Públicos Sul 702/902, Bloco C
Cen o Emp esa ial B asília 50, To e B
CEP: 70390-025, Asa Sul, B asilia-DF
DISCUSSION PAPER
Ipea´s mission
To enhance s a e decision-making and public deba e.
