Can o e, C is iano; Mum az, Ha oon; Fe oni, Filippo; Theophilopoulou, Angeliki
Wo king Pape
A ail o labo supply and a ale o mone a y policy
Quade ni - Wo king Pape DSE, No. 1210
P o ided in Coope a ion wi h:
Uni e si y o Bologna, Depa men o Economics
Sugges ed Ci a ion: Can o e, C is iano; Mum az, Ha oon; Fe oni, Filippo; Theophilopoulou, Angeliki
(2025) : A ail o labo supply and a ale o mone a y policy, Quade ni - Wo king Pape DSE, No.
1210, Alma Ma e S udio um - Uni e si à di Bologna, Dipa imen o di Scienze Economiche (DSE),
Bologna,
h ps://doi.o g/10.6092/unibo/amsac a/8501
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/331369
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-nc/4.0/
ISSN 2282-6483
A ail o labo supply
and a ale o mone a y policy
C is iano Can o e
Ha oon Mum az
Filippo Fe oni
Angeliki Theophilopoulou
Quade ni - Wo king Pape DSE N°1210
A ail o labo supply and a ale o mone a y policy∗
C is iano Can o e
Sapienza Uni e si y o Rome
Filippo Fe oni
Uni e si y o Bologna
Ha oon Mum az
Queen Ma y, Uni e si y o London
Angeliki Theophilopoulou
B unel Uni e si y London
Sep embe 2, 2025
Abs ac
We s udy he in e ac ion be ween mone a y policy and labo supply decisions a he
household le el. We unco e e idence o he e ogeneous esponses and a s ong coun e -
cyclicali y o hou s wo ked in he le ail o he income dis ibu ion ollowing a mone a y
policy shock in he U.S. Speci ically, while agg ega e hou s and labo ea nings decline a e
a mone a y igh ening, indi iduals a he bo om o he income dis ibu ion inc ease hei
hou s wo ked. Mo eo e , his posi i e labo supply esponse is quan i a i ely signi ican ,
subs an ially dampening he decline in agg ega e hou s wo ked. We show ha he empi -
ical pa e ns a e consis en wi h a s anda d one-asse HANK model ea u ing endogenous
labo supply. The model e eals ha s ong income e ec s a he bo om o he dis ibu ion
can accoun o he obse ed coun e cyclical labo esponses, highligh ing how labo supply
adjus men s ac as an addi ional ma gin h ough which households smoo h consump ion.
Compa ing his speci ica ion o a model wi h a homogeneous labo supply, we ind ha
labo supply he e ogenei y educes he agg ega e MPC and a enua es he ansmission o
mone a y policy h ough agg ega e demand. As a esul , he ou pu cos o disin la ion is
lowe in economies whe e poo e households can lexibly adjus hei labo e o , easing he
ade-o aced by he cen al bank.
Keywo ds: Mone a y policy, Household Su ey, FAVARs, HANK.
JEL classi ica ion: E52, E32, C10
∗We would like o hank Vale io Pie oni o he lengh y discussions and sugges ions abou he HANK analysis
we also hank Michele And eolli, Guido Asca i, Saleem Bahaj, Ch is ian Baye , Gadi Ba le y, Flo in O. Bilbiie,
Da ide Debo oli, Luca Fo na o, Luca Gambe i, Nezih Gune , Ch is Huck eld , Ma hias Klein, Guido Lo enzoni,
Leona do Melosi, Ca lo Pizzinelli, Rica do Reis, Kje il S o esle en, Dan Sulli an, Paolo Su ico, Gianluca Violan e,
h ee anonymous e e ees and pa icipan s a nume ous con e ences and semina s o commen s and sugges ions.
1
Non Technical Summa y
The pape asks a simple ques ion wi h a su p ising answe : when he cen al bank aises
in e es a es, do people change how much hey wo k, and does ha change di e by income?
Looking a U.S. su ey da a a e unexpec ed a e hikes, he au ho s ind ha o al hou s
and ea nings in he economy all, as s anda d s o ies p edic . Bu among people a he
bo om o he income dis ibu ion, hou s wo ked ac ually ise. These wo ke s a e also less
likely o sepa a e om hei jobs and hei beha io is mo e sensi i e o in e es a e mo es
han ha o middle- and highe -income households. Because low- and mode a e-income
wo ke s accoun o a meaning ul sha e o o al hou s, his ex a e o pa ially o se s he
o e all decline in labo inpu .
Why would hose wi h he leas esou ces wo k mo e when he economy cools? The
pape poin s o a s aigh o wa d in ui ion. Mone a y igh ening lowe s eal wages and
aises he cos o se icing deb . Households wi h li le sa ings and igh budge s ha e a
ha de ime smoo hing hei spending by dipping in o asse s o bo owing. Ins ead, hey
smoo h by supplying mo e labo . In o he wo ds, he ”income e ec ” o eeling poo e
domina es he usual ”subs i u ion e ec ” ha would make people wo k less when wages
so en. The au ho s also a gue ha his pa e n e lec s people’s choices a he han i ms’
selec i e hi ing o scheduling. The esul holds in panel da a, emains when ocusing on ull-
ime wo ke s, and is accompanied by alling wages alongside ising hou s a he bo om–signs
ha supply, no demand, is doing mos o he wo k.
To es he in e p e a ion, he pape builds a s anda d mac o model ha allows house-
holds o di e and le s each choose how much o wo k. Calib a ed o ma ch ealis ic a ia ion
in income, asse s, and bo owing limi s, he model ep oduces he da a’s key ea u e: a e
a con ac iona y mone a y shock, lowe -income households inc ease hei labo e o while
highe -income households educe i . In he model, he mechanism comes om s ong income
e ec s among cons ained amilies, d i en by lowe eal wages and highe deb paymen s.
These di e ences ma e o he b oade s o y o how mone a y policy mo es he economy.
When poo e households can lex hei hou s, hey ely less on cu ing spending o adjus ,
so he hi o o e all demand is smalle han in models ha assume e e yone beha es he
same. The au ho s show ha his cushions he ou pu decline ha ypically accompanies
disin la ion. Measu ed by he sac i ice a io– he cumula i e ou pu loss pe pe cen age poin
educ ion in in la ion o e he i s yea – he cos is abou one- hi d lowe in he e sion o
he model whe e labo supply can a y ac oss households ( o example, oughly 0.67 a he
han 1.02 unde a calib a ion wi h s ong income e ec s).
The upsho is ha he ” ail” o he income dis ibu ion changes he ” ale” o mone a y
policy. I cen al banks igno e how low-income households adjus hei wo k, hey isk
o e s a ing he g ow h cos o b inging in la ion down and mis eading he ade-o s in ol ed
in aising a es.
2
1 In oduc ion
Do people adjus how much hey wan o wo k when he cen al bank’s mone a y policy
s ance shi s? Mo e speci ically, does an in e es a e hike induce indi iduals o wo k mo e
o ewe hou s? And does his e ec di e ac oss households wi h di e en le els o income
(o ea nings)?
The as li e a u e on he he e ogeneous e ec s o mone a y policy has ocused on he
in e - empo al channel ha a ec s he consump ion and sa ings plans o households (Bil-
biie (2008), Aucle (2019), Cloyne, Fe ei a and Su ico (2020), Kaplan, Moll and Violan e
(2018)). Howe e , changes in consump ion plans induced by a ia ion in a es also in luence
he in a empo al alloca ion be ween consump ion and leisu e; i.e., a household’s desi ed
supply o labo depends on how each indi idual can subs i u e consump ion wi h wo king
ime and/o compensa e wi h di e en sou ces o income. In s anda d models, he lowe
wage a es induced by a con ac iona y mone a y policy ha e wo e ec s on households’
labo supply: a subs i u ion e ec ha educes how much households would p e e o wo k
and an income e ec ha inc eases i .
The majo i y o he heo e ical mac o li e a u e, assumes no o negligible income e ec s
on he labo supply.1I is o en hough ha income e ec s a e small because –being sho -
li ed– mone a y policy shocks do no ha e la ge e ec s on li e ime income, which is wha
ma e s o an op imizing wo ke -consume .2Mo eo e , mone a y policy is adi ionally
iewed as a ec ing labo demand h ough he ex ensi e ma gin and ha ing li le e ec on
labo supply.3
The scope o his pape is o e isi his channel and s udy he ansmission mecha-
nism o mone a y policy o he labo supply decisions a he household le el.4Fi s , we
o e no el empi ical e idence on he e ec o mone a y policy on hou s wo ked a a mo e
g anula , disagg ega ed le el. To do his, we s udy he e ec s o unexpec ed shi s in he
mone a y policy s ance on he amoun o hou s wo ked by households wi h di e en income
le els using su ey da a o he U.S. We ind ha indi iduals a he bo om o he income
dis ibu ion inc ease hei hou s wo ked ollowing a mone a y policy igh ening, in con as
wi h con en ional mac oeconomic heo y. A he same ime, agg ega e hou s and wages
ac oss he whole dis ibu ion decline. This adjus men occu s h ough bo h he in ensi e
1E.g. Gal´ı, Sme s and Wou e s (2012); Dy da and Ped oni (2022); Wol (2021); Aucle and Rognlie (2020);
amongs o he s.
2Howe e , mos o he empi ical e idence used o suppo his assump ion ocuses on o he shocks and no on
mone a y policy shocks. See he li e a u e e iew sec ion o mo e de ails.
3Fo example, quo ing he Fede al Rese e Chai man Je ome Powell on a speech on No embe 30 2022 ”Policies
o suppo labo supply a e no he domain o he Fed: Ou ools wo k p incipally on demand.”
4The consequences o mone a y policy ac ions on he labo ma ke dynamics a e no only o in e es in academic
cycles. Policymake s ha e exp essed conside able in e es in labo ma ke ou comes ac oss he whole spec um o
he popula ion and in pa icula in low- and mode a e-income communi ies. E.g. in a Jackson Hole speech on
Augus 27, 2020, J. Powell said in un eiling he new Fed s a egy ha “ou e ised s a emen emphasizes ha
maximum employmen is a b oad-based and inclusi e goal. This change e lec s ou app ecia ion o he bene i s
o a s ong labo ma ke , pa icula ly o many in low- and mode a e-income communi ies.”
3
and ex ensi e ma gins o labo supply, bu wi h impo an he e ogenei y ac oss he income
dis ibu ion. In pa icula , low-income indi iduals end o inc ease hei hou s wo ked and
exhibi lowe sepa a ion a es ollowing a mone a y con ac ion. Mo eo e , hei esponse is
mo e sensi i e o in e es a e a ia ions compa ed wi h o he pe cen iles o income in he
popula ion. As he labo supplied by low- and mode a e-income households ( he ail o labo
supply) ep esen s bo h a non-negligible sha e o he ola ili y and a ele an p opo ion o
hou s wo ked in he agg ega e, his esponse is also quan i a i ely ele an om a mac o
pe spec i e.
The coun e cyclicali y o hou s wo ked in he le ail o he income dis ibu ion obse ed
in he da a is an equilib ium ou come esul ing om he in e ac ion o households’ labo sup-
ply o ces and i ms’ labo demand ac o s and consis en wi h mul iple explana ions. Fo
example, as he ecession induced by a con ac iona y mone a y policy inc eases he p ob-
abili y o becoming unemployed, households wi h limi ed income sou ces ha e incen i es o
wo k mo e. Simila ly, indi iduals who a e close o hei bo owing limi s may need o wo k
mo e hou s o mee hei deb obliga ions when in e es a es ise. Supply-side explana ions
sugges ha when lacking bu e sa ings o non-labo income sou ces, households wi h low-
and mode a e-incomes ha e less oom o maneu e du ing ough economic imes, and by
a ying hei labo supply hey can smoo h consump ion along he business cycle. Al e na-
i ely, on he demand side, i ms may lay o empo a y o pa - ime wo ke s and adjus he
labo ’s inpu by u ilizing mo e o hei exis ing labo o ce inducing selec ion in he sample.
While i is e y di icul o isola e he dominan channel esponsible o ou empi ical ind-
ings and a combina ion o hese s o ies is mo e likely, se e al pieces o e idence sugges ha
selec ion is no he dominan o ce in his con ex ; i.e. he esul s ca y o e when using
he panel dimensions o ou su ey da a and o when isola ing he esponse o only ull- ime
employed wo ke s. Finally, he e idence o alling wages alongside ising hou s wo ked a
he bo om o he income dis ibu ion sugges s ha labo supply o ces- a he han labo
demand shi s-play a dominan ole in d i ing his pa e n.
I is he e o e na u al o ask whe he he labo supply beha io obse ed among low-
income indi iduals can be heo e ically a ionalized. Ano he way o assess whe he he
empi ical pa e ns a e p ima ily d i en by labo supply a he han labo demand o ces
is o s udy hem in a s uc u al model. I a s anda d he e ogeneous-agen amewo k wi h
endogenous labo supply-abs ac ing om he e ogenei y in labo demand-can ep oduce he
coun e cyclical esponses o hou s wo ked a he bo om o he income dis ibu ion, his
would p o ide s ong suppo o a supply-side in e p e a ion o he da a. The second
con ibu ion o he pape is hus o explo e hese mechanisms in heo y and assess hei
implica ions o he ansmission ( he ale) o mone a y policy.
We s a by p o iding a simple in ui ion o he mechanism a wo k. Bo owing cons ain s
and limi ed consump ion smoo hing (A h eya, Owens and Schwa zman (2017)) a e likely
o d i e s onge income e ec s on labo supply decisions o households wi h low income
4
and limi ed asse s. In pa icula , cons ained o hand- o-mou h (HTM) households ace a
igh e in e empo al budge cons ain and a wedge in hei Eule equa ion, making hei
ma ginal u ili y o consump ion highly sensi i e o income luc ua ions. By analyzing how
bo owing cons ain s and he cu a u e o he u ili y unc ion wi h espec o consump ion
a ec he op imal choices o consump ion and leisu e, we show ha mone a y policy shocks
can gene a e an inc ease in labo supply among cons ained households, d i en by income
e ec s. This mechanism di ec ly links he in e empo al and in a empo al choices o house-
holds, highligh ing he impo ance o he e ogenei y in bo owing cons ain s and ma ginal
p opensi ies o wo k.
We hen u n o a quan i a i e analysis using a s anda d one-asse he e ogeneous-agen
New Keynesian (HANK) model wi h endogenous labo supply and nominal p ice igidi ies.
We show ha his ”o - he-shel ” HANK model, calib a ed o ma ch plausible ea u es o
household he e ogenei y, is able o ep oduce he key labo supply pa e ns we unco e in
he da a: ollowing a con ac iona y mone a y policy shock, households in he lowe pa
o he income dis ibu ion inc ease labo supply, while labo supply declines among highe -
income households. Wi h his model, we can also decompose he labo supply esponse in o
i s unde lying channels. This e eals ha he coun e cyclical labo supply a he bo om o
he income dis ibu ion is p ima ily d i en by income e ec s-speci ically, he decline in eal
wages and he inc ease in deb epaymen bu dens ollowing a mone a y igh ening. We hen
sys ema ically s udy how he s eng h o hese he e ogeneous labo supply esponses a ies
ac oss di e en model calib a ions by al e ing he elas ici y o in e empo al subs i u ion
(EIS) and he bo owing limi . To u he quan i y he mac oeconomic implica ions o he -
e ogeneous labo supply, we compa e his baseline model o a simila HANK economy whe e
labo supply is homogeneous ac oss households, as in Aucle , Rognlie and S aub (2024).
To ensu e compa abili y, we calib a e bo h models o ma ch he labo supply esponse o
he median agen ype, and examine di e ences in he s eady s a e and in he esponses o
mone a y policy shocks.
We ind ha allowing o he e ogeneous labo supply has quan i a i ely signi ican im-
plica ions o mone a y ansmission. In pa icula , he s eady-s a e agg ega e ma ginal
p opensi y o consume (MPC) is sys ema ically lowe in models wi h endogenous, he e o-
geneous labo supply han in compa able models whe e labo supply is homogeneous. This
di e ence is especially p onounced a low alues o he elas ici y o in e empo al subs i-
u ion, whe e income e ec s a e s onge and cons ained households ely mo e on labo
e o o bu e shocks. This addi ional adjus men ma gin dampens he agg ega e consump-
ion esponse o mone a y policy and, c ucially, educes he eal cos o disin la ion o he
mone a y au ho i y. We quan i y his e ec by compu ing he sac i ice a io, de ined as he
cumula i e pe cen age ou pu loss pe cumula i e pe cen age poin educ ion in in la ion
o e he i s yea ollowing a con ac iona y mone a y shock. Ac oss di e en calib a ions
o he EIS, we ind ha he sac i ice a io is sys ema ically lowe in he model wi h he -
5
e ogeneous labo supply. Fo ins ance, unde a low EIS (high income e ec ), he sac i ice
a io alls om 1.02 in he homogeneous labo supply model o 0.67 in he he e ogeneous
labo supply one model, a 35% educ ion in he ou pu cos o disin la ion. This esul
a ises because low-income households inc ease labo e o in esponse o he shock, pa ially
o se ing he decline in consump ion and mi iga ing he con ac ion in agg ega e demand.
F om a policy pe spec i e, his implies ha ailing o accoun o he e ogenei y in labo
supply may lead cen al banks o o e es ima e he ou pu cos s o achie ing disin la ion and
misjudge he ade-o s in ol ed in mone a y igh ening.
The pape is o ganized as ollows: he nex subsec ion discusses he exis ing li e a u e.
Sec ion 2desc ibes he da a and he empi ical s a egy and p esen s ou empi ical e idence.
Sec ion 3p esen s a s uc u al model ha accoun s o his e idence and in es iga es he im-
plica ion o he ansmission o mone a y policy. Finally, sec ion 4p o ides some concluding
ema ks.
1.1 Rela ed Li e a u e
This pape con ibu es o he li e a u e on mone a y policy and household he e ogenei y.
While mos empi ical wo k has ocused on balance shee composi ion and he he e ogenei y
in MPCs ollowing mone a y shocks (Cloyne e al. (2020), Aucle (2019)), we ins ead s udy
how such shocks a ec labo supply decisions ac oss households. By examining he e ogeneous
labo supply esponses in HANK models, we also highligh hei implica ions o agg ega e
MPCs.
Kehoe, Lopez, Pas o ino and Salgado (2020) and Ami -Ahmadi, Ma hes and Wang
(2021) documen he e ogenei y in he esponses o hou s wo ked and unemploymen ac oss
U.S. demog aphic g oups. The o me inds ha labo supply is less cyclical o olde and
college-educa ed wo ke s, while he la e shows la ge a ia ion in unemploymen esponses.
We complemen hese s udies by so ing households by income bins a he han demog aphic
ai s and ocusing on he in ensi e ma gin o labo supply.
G a es, Huck eld and Swanson (2023) s udy he e ec o mone a y policy on he labo
ma ke lows and ind ha a mone a y policy igh ening induces an inc ease in he ac ion o
labo o ce non-pa icipan s epo ing ha hey wan a job and an inc ease in he numbe o
dis inc job sea ch me hods by unemployed indi iduals. Bo h hese ma gins o adjus men s
a e consis en wi h an inc ease in he labo supply o non-employed indi iduals. These esul s
a e in line and complemen a y wi h ou indings on he inc ease o hou s wo ked o wo ke s
wi h low o mode a e income (bo h cu en ly employed o coming om non-employmen )
ollowing a mone a y policy igh ening.
Del Can o, G igsby, Qian and Walsh (2025) also s udy he dis ibu ional e ec s o he US
mone a y shocks using mon hly VARs and da a om he CPS, bu hei ocus is no ma i e
a he han posi i e.
Se e al pape s use adminis a i e da a o s udy he he e ogeneous e ec s o mone a y
6
policy on labo ma ke ou comes. Fo Scandina ian coun ies Ambe g, Jansson, Klein and
Rogan ini-Picco (2022), Ande sen, Johannesen, Jø gensen and Peyd ´o (2021) and Holm,
Paul and Tischbi ek (2021)) ocus on labo income and cap u e combined e ec s on bo h
he ex ensi e and in ensi e ma gins. Coglianese, Olsson and Pa e son (2025) analyze ad-
minis a i e da a om Sweden and show ha unemploymen esponses o mone a y shocks
a y ac oss he ea nings dis ibu ion, ocusing on labo ma ke ansi ions. Hube and Sa -
ignac (2024) ind ha in F ance, mos o he a ia ion in labo income o he bo om hal o
he dis ibu ion s ems om he ex ensi e ma gin, while B oe , K ame and Mi man (2022)
documen he e ogeneous unemploymen isk in Ge many, wi h low-income wo ke s acing
mo e p o-cyclical sepa a ion a es. Howe e , none o hese s udies can disen angle hou s
wo ked om wages, as we do he e. Ou con ibu ion is o iden i y a dis inc ansmission
channel: he he e ogeneous esponse o hou s wo ked o a mone a y policy shock. Mo eo e ,
unlike mos o hese s udies (excep B oe e al. (2022)), we use da a a mon hly equency,
which allows us o exploi a longe ime-se ies dimension o iden i y he ansmission o
mone a y policy shocks.
Mo i a ed by ou empi ical indings o he U.S., Das, Hambu , Hellwig and Sp ay (2025)
s udy he e ec s o mone a y policy on hou s wo ked using adminis a i e da a om Aus-
alia. Le e aging high- equency iden i ica ion and indi idual-le el income and hou s da a,
hey ind ha labo supply esponses a e s onge among low-income and low-liquidi y in-
di iduals. Thei esul s con i m ha income e ec s play a key ole in shaping labo supply
eac ions o in e es a e shocks.
As discussed in he in oduc ion, mac oeconomic models o en assume negligible income
e ec s on labo supply. Howe e , he empi ical e idence suppo ing his iew a ely ocuses
on business cycle shocks. Mos es ima es come om idiosync a ic income shocks, such
as lo e y winnings. Fo example, Cesa ini, Lindq is , No owidigdo and ¨
Os ling (2017)
use Swedish adminis a i e da a and ind modes income e ec s, while Goloso , G abe ,
Mogs ad and No go odsky (2023), using U.S. da a, a gue ha labo supply esponses o
lo e y winnings a e sizable and no negligible.
F om a heo e ical pe spec i e, we con ibu e o he li e a u e on mic o-le el he e ogene-
i y in New Keynesian models. Mos exis ing wo k ocuses on he consump ion channel o
mone a y policy while abs ac ing om labo supply he e ogenei y (e.g., Aucle (2019),
Aucle e al. (2024), Baye , Bo n and Lue icke (2024), Bilbiie (2024)). A no able excep-
ion is A h eya e al. (2017), who emphasize he ole o labo supply decisions and ma ginal
p opensi ies o wo k in shaping he e ec s o iscal ans e s. To ou knowledge, we a e
he i s o s udy his channel in he con ex o mone a y policy. Simila ly, Gue ie i and
Lo enzoni (2017) explo e how di e en u ili y calib a ions a ec labo supply esponses o
c edi shocks in a he e ogeneous-agen model wi h incomple e ma ke s.
Impo an ly, while ou empi ical and heo e ical esul s highligh he ele ance o he -
e ogeneous labo supply, inco po a ing his ea u e in o HANK models p esen s challenges-
7
Figu e 3: Impulse esponses o a mone a y policy shock.
0 20 40
-0.5
0
0.5
1
1.5
0 20 40
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0 20 40
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
No es: This igu e depic s he impulse esponses o a mone a y policy igh ening shock. The le panel epo s
he esponses o hou s wo ked o he i s quin ile o he ea ning dis ibu ion, he cen al panel he agg ega e
measu e o hou s wo ked using he CPS da a, and igh panel he al e na i e agg ega e measu e o hou s wo ked
which excludes i s quin ile o he ea ning dis ibu ion. Da k (ligh ) ed a eas 68 (90)% con idence se s.
The ex en o which his milde con ac ion in labo supply ansla es in o less ampli ica-
ion o o he a iables (especially in la ion) is less clea . Fo answe ing he la e we need o
cons uc an hypo he ical coun e ac ual economy wi hou he le ail o labo supply. The
empi ical model does no allow o un such coun e ac uals. The s uc u al model p esen ed
in sec ion 3can shed some ligh on his poin .
Composi ion E ec s A po en ial conce n abou he empi ical e idence p esen ed ea lie
is ha he obse ed inc ease in hou s wo ked among low-income indi iduals ollowing a
mone a y igh ening may e lec composi ion e ec s. Fo ins ance, i pa - ime o low-hou
wo ke s a e mo e likely o exi employmen du ing down u ns, a e age hou s could ise
mechanically e en i indi idual labo supply emains unchanged. To add ess his conce n,
we i s es ic he sample o ou analysis o ull- ime wo ke s.11 Figu e 4displays he
esponse o hou s wo ked o di e en income le els. Remo ing pa - ime wo ke s does no
in alida e ou main indings and hou s inc ease a he le ail o he ea nings dis ibu ion
a e a mone a y con ac ion.
Up o now ou empi ical analysis is cons uc ed using a epea ed c oss-sec ion which e en
con olling o pa - and ull- ime wo ke s migh s ill be p one o composi ion e ec s. To
ule hose ou , we le e age he panel dimension o he CPS cons uc ed by he Kansas Ci y
11Fo his exe cise we use he Kansas Fed ex ac o he CPS by se ing he a iable l de ail76 equal o ei he 1
o 2.
14
Figu e 4: Responses o hou s o ull- ime employees
6 m hs
2 yea s
No es: This igu e depic s he impulse esponses o a mone a y policy igh ening shock. Da k (ligh ) ed a eas 68
(90)% con idence se s.
Fed (h ps://cps.kansasci y ed.o g/) and ack changes in hou s wo ked a he indi idual
le el. Speci ically, we compu e he change in hou s be ween mon h and + 12 and use
i s a e age wi hin income g oups as he dependen a iable in a FAVAR amewo k. This
app oach mi iga es composi ion conce ns inhe en in c oss-sec ional a e ages.
The op panel o Figu e 5shows he g ow h in hou s wo ked six mon hs ( ed) and wo
yea s (blue) pos -shock, ac oss he income dis ibu ion. Despi e wide con idence in e als,
he esul s suppo ou main inding: low-income indi iduals inc ease hei hou s wo ked
in esponse o con ac iona y policy. The bo om panel o Figu e 5 epo s employmen
ou lows12 o e he same ho izon. While ou lows ise o middle- and high-income g oups,
hey decline o low-income indi iduals, indica ing s onge job a achmen in he lowe ail
o he dis ibu ion. These indings con i m ha bo h in ensi e and ex ensi e labo supply
ma gins espond o mone a y shocks, bu in income-dependen ways. In he nex sec ions,
12See oo no e 8and Appendix A.1 o de ails.
15
Figu e 5: Dis ibu ion o esponses o he g ow h a e o hou s wo ked and o employmen a e
mone a y policy igh ening.
6 m hs
2 yea s
No es: The op panel depic s esponse o he g ow h a e o hou s wo ked g ow h six mon hs ( ed) and wo
(blue) yea s a e he shock using he panel e sion o he CPS. The bo om panel depic s esponse o he
ou lows om employmen six mon hs ( ed) and wo (blue) yea s a e he shock using he panel e sion o he
CPS. Shaded a eas (solid lines) 68 (90)% con idence se s.
we in e p e hese pa e ns h ough he lens o heo y.
Supply and demand ac o s The coun e cyclical inc ease in labo supply among low-
income wo ke s e lec s he in e play be ween household labo supply and i m-le el labo
demand. Fo middle- and high-income indi iduals, bo h eal wages and hou s wo ked decline
16
ollowing a mone a y igh ening, consis en wi h a le wa d shi in labo demand. In con-
as , among low-income wo ke s, we obse e ising hou s despi e alling eal wages-indica ing
ha income e ec s on labo supply domina e in his g oup. This pa e n o nega i e co-
mo emen be ween wages and hou s is sugges i e o supply-side o ces a play, as discussed
in Ka z and Mu phy (1992), whe e such como emen ypically e lec s mo emen s along a
downwa d-sloping labo demand cu e in esponse o shi s in labo supply. The esponses
o eal wages a di e en income pe cen iles a e epo ed in he Appendix D.3 igu e D.8.
Sec o s and educa ion We u he explo e he he e ogenei y behind his pa e n by
disagg ega ing esponses by indus y and educa ion.13 Low-wage wo ke s a e la gely em-
ployed in sec o s such as wholesale and e ail ade, leisu e and hospi ali y, and educa ion
and heal h se ices (Appendix Figu e D.3). Wi hin hese sec o s, we obse e clea posi i e
esponses o hou s wo ked a e a mone a y con ac ion, especially a he bo om o he wage
dis ibu ion (Appendix Figu e D.4).
Educa ional a ainmen also plays a ole. Low-income, non-college-educa ed indi iduals
exhibi a la ge inc ease in labo supply a e mone a y igh ening han hei college-educa ed
pee s sugges ing ha educa ion mode a es he s eng h o income e ec s (Appendix Figu e
D.7).
3 Labo supply and he e ogenei y
We now u n o a heo e ical model o a ionalize ou empi ical indings and assess hei
implica ions o he mone a y policy ansmission mechanism. We s a om a gene al
and s ylized amewo k ha cla i ies he no el link be ween he e ogeneous labo supply and
mone a y policy. Speci ically, cons ained agen s acing igh e inancial condi ions end o
sac i ice leisu e and inc ease labo supply o sus ain hei consump ion in esponse o a
decline in income.
Ou empi ical e idence, demons a ing ha hou s wo ked inc ease among households
in he lowe pa o he income dis ibu ion a e a mone a y policy igh ening, aligns wi h
se e al po en ial heo ies. While he obse ed equilib ium ou come e lec s bo h supply- and
demand-side ac o s, he in e se mo emen o wages and hou s wo ked–wages alling while
hou s inc ease–poin s p edominan ly owa d a labo supply esponse d i en by income e ec s.
Fo his eason, in his sec ion we ocus he heo e ical analysis on labo supply he e ogenei y
while assuming a s anda d labo demand side, modeled h ough a ep esen a i e i m wi h
homogeneous labo demand.
3.1 Income e ec s
We begin by analyzing he household’s p oblem in pa ial equilib ium, aking labo income
as gi en, o highligh how he s eng h o he income e ec on indi idual labo supply is
13Fo mo e de ails see appendix D.2.
17
shaped by wo key o ces: he cu a u e o he u ili y unc ion and he igh ness o he
bo owing cons ain .14
A key ea u e in mac oeconomic models de e mining household labo supply is he in-
a empo al op imali y condi ion go e ning he ade-o be ween consump ion and leisu e.
Le H deno e hou s wo ked a ime ,C consump ion, and U(C , H ) he household’s u ili y
unc ion. Wi h w as he eal wage a e, his op imali y condi ion is:
−Uh(C , H )
Uc(C , H )=w ,(4)
whe e Uhand Uca e pa ial de i a i es wi h espec o hou s and consump ion, espec i ely.
Households also ace an in e empo al consump ion decision summa ized by he Eule
equa ion (abs ac ing om unce ain y o now):
Uc(C , H )
Uc(C +1, H +1)=β(1 + )(1 + ω ),(5)
whe e is he eal in e es a e, a ec ed di ec ly by mone a y policy, and ω ep esen s a
wedge a ising om bo owing cons ain s o o he inancial ic ions. Typically, cons ained
households ace a posi i e wedge (ω >0), e lec ing a highe ma ginal u ili y o cu en
consump ion.
Following a con ac iona y mone a y policy shock ha aises , he household’s in a em-
po al op imali y condi ion de e mines how hou s wo ked adjus in esponse o changes in
income and wages. The s eng h o his labo supply adjus men c ucially depends on he
cu a u e o he u ili y unc ion o e consump ion. We assume p e e ences o he o m:
U(C , H ) = C1−1/σ
1−1/σ −φH1+ν
1 + ν,
whe e σdeno es he elas ici y o in e empo al subs i u ion and ν he in e se o he F isch
elas ici y o labo supply.
A lowe σ(co esponding o a mo e conca e u ili y in consump ion) implies ha ma ginal
u ili y eac s mo e s ongly o income changes, ampli ying he labo supply esponse when
consump ion declines. Households acing a nega i e income shock will hen supply mo e
labo o s abilize hei u ili y. Con e sely, a highe σ la ens he u ili y cu e, making
households less sensi i e o luc ua ions in income, and hus dampening he adjus men o
hou s wo ked.15
Bo owing cons ain s ( ia ω in he Eule equa ion) u he magni y his e ec by o c-
ing some households o adjus labo supply mo e agg essi ely in esponse o changes in
cu en income and in e es a es. The bo owing limi di ec ly a ec s he s eng h o his
labo supply channel. A igh e bo owing cons ain (o highe equilib ium deb le els)
ampli ies he sensi i i y o labo supply o mone a y policy shocks: when he eal in e es
14We hank an anonymous e e ee o sugges ing how o s uc u e his sec ion.
15See Bilbiie (2008), who shows, in a wo-agen NK model, how σa ec s he sign o he esponse o he labo
supply o hand- o-mou h agen s.
18
a e ises, highe deb epaymen s educe disposable income, s eng hening he income e -
ec . Cons ained households, unable o smoo h consump ion h ough bo owing, espond
by inc easing hei labo supply o o se he highe inancial bu den.
In Appendix F, we illus a e how hese o ces in e ac in a simple wo-agen (bo owe -
sa e ) economy whe e bo owe s ace igh e bo owing limi s and highe impa ience ela i e
o sa e s. We show analy ically how bo owe s’ labo supply esponse a e an in e es a e
inc ease depends di ec ly on he EIS (dec easing in σ) and hei bo owing limi (inc easing
in he ne deb posi ion).
Thus, bo h he EIS and bo owing cons ain s c ucially shape he ex en o which he
labo supply o cons ained households becomes coun e cyclical ollowing mone a y policy
shocks, o e ing a clea heo e ical in e p e a ion in e ms o labo supply o ou empi ical
indings.
In he nex sec ion, we mo e o a gene al equilib ium analysis and conside a s anda d
one-asse he e ogeneous agen s New Keynesian model wi h s agge ed p ice se ing.
3.2 He e ogeneous Labo Supply in HANK
The pu pose o his sec ion is o demons a e ha ou empi ical e idence aligns wi h he
implica ions o he e ogeneous labo supply esponses in a s anda d HANK model. To
do so, we use he one-asse HANK model wi h endogenous labo supply, as in Aucle ,
Ba d´oczy, Rognlie and S aub (2021).16 The goal is o examine how di e en calib a ions o
he model a ec he beha io o labo supply ac oss he income and weal h dis ibu ion, and
o compa e he agg ega e implica ions o he mone a y ansmission mechanism wi h and
wi hou he e ogeneous labo supply esponses in he model. The main esul s a e as ollows:
(i) An o - he-shel HANK model wi h he e ogeneous labo supply and homogeneous labo
demand can eplica e ou empi ical indings on he labo supply esponse ac oss he income
dis ibu ion. (ii) The main d i e o he esponse a he lowe end o he dis ibu ion is
an income e ec , s emming om alling wages and ising deb epaymen s. Mo eo e , he
s eng h o his esponse is ampli ied by he cu a u e o he u ili y unc ion in consump ion
and he igh ness o he bo owing limi , u he highligh ing he impo ance o income
e ec s. (iii) The p esence o labo supply he e ogenei y lowe s he eal cos o disin la ion,
making i easie o he cen al bank o achie e i s in la ion a ge .
3.2.1 One-asse HANK
The model ea u es a he e ogeneous agen s sec o simila o McKay, Nakamu a and S einsson
(2016), coupled wi h a s anda d New Keynesian supply-side block.
Households. The e is a uni mass o ex-an e iden ical households who di e ex-pos by
hei labo p oduc i i y e(“skill”) and asse holdings a. Fo no a ional simplici y, we use
16The only di e ence wi h Aucle e al. (2021) is ha we abs ac om go e nmen spending, gi en he ocus
on mone a y policy he e. Resul s a e no a ec ed by his choice.
19
he subsc ip i o deno e household-le el ou comes, ins ead o w i ing hem explici ly as
unc ions o s a e a iables.
Skill e ollows a ime-in a ian disc e e Ma ko chain wi h nes a es, E={e1, e2, . . . , ene},
and exogenous ansi ion p obabili ies P(e′, e). This in oduces cyclical income isk in he
model. The s a iona y dis ibu ion o Pis deno ed by πe. A e age labo p oduc i i y
R1
0ei, di is in a ian and no malized o 1.
We assume ha P(e′, e) disc e izes a log AR(1) p ocess:
log ei =ρelog ei −1+σeϵi ,
wi h no mal inno a ions ϵi ∼ N(0,1), and we use he Rouwenho s me hod o disc e iza-
ion.
Households can eely choose he numbe o hou s h hey wo k, subjec o an addi i ely
sepa able u ili y cos o wo king. Consump ion, sa ings, and labo choices c,a, and ha e
he solu ion o he household’s u ili y maximiza ion p oblem, cha ac e ized by he ollowing
Bellman equa ion:
V (ei , ai −1) = max
ci ,hi ,ai (c1−σ−1
i
1−σ−1−φh1+ν
i
1 + ν+βE V +1 (ei +1, ai ))
s. . ci +ai = (1 + )ai −1+w ei hi −τ ¯τ(ei ) + d ¯
d(ei )
ai ≥a.
Households ecei e an hou ly wage w , pay axes, and ecei e di idends om i m own-
e ship acco ding o incidence ules ¯τ(e) and ¯
d(e).17
Fi ms. A compe i i e inal goods i m assembles i s ou pu using a CES p oduc ion unc-
ion Y =R1
0y
1
µ
j djµ
, gi ing ise o a s anda d CES demand sys em o he con inuum o
in e media e goods. These in e media es, in u n, a e supplied by monopolis s who p oduce
using only labo , such ha yj =Z hj , and a e subjec o quad a ic p ice adjus men cos s
a la Ro embe g (1982):
ψ (pj , pj −1) = µ
µ−1
1
2κlog pj
pj −12
Y .
He e, Z deno es agg ega e p oduc i i y ha may be ime- a ying. In a symme ic equilib-
ium, g oss in la ion 1 + π =P
P −1e ol es acco ding o he Phillips cu e:
log(1 + π ) = κw
Z
−1
µ+1
1 + +1
Y +1
Y
log(1 + π +1),
and di idends equal ou pu ne o labo and p ice adjus men cos s: d =Y −w H −ψ .
Policy. Mone a y policy se s he nominal a e on bonds acco ding o a s anda d Taylo
ule:
i = ∗+ϕπ +ϵ ,
17This implies ha skill ede e mines no only a household’s income pe hou wo ked, bu also he amoun o
lump-sum axes she mus pay and he di idends she ecei es, bo h o which a e dis ibu ed p opo ionally o e.
20
whe e ∗
is he economy’s long- un “na u al a e” and ϵ he mone a y policy shock. The
eal in e es a e in pe iod is de e mined by he p e iously se nominal a e and in la ion
so ha :
1 + =1 + i −1
1 + π
.
The iscal au ho i y, in u n, issues a cons an amoun o go e nmen bonds Beach
pe iod and collec s he lump-sum axes τ al eady men ioned abo e. Since he e a e no
o he spending i ems (abs ac ing om go e nmen spending), i chooses he ax a e o
co e i s in e es a e paymen s e e y pe iod:
τ = B.
Ma ke Clea ing. In an equilib ium, he ollowing ma ke clea ing condi ions mus hold:
•Asse ma ke :
B=Z1
0
ai di,
•Labo ma ke :
Y =H =Z1
0
ei hi di,
•Goods ma ke :
Y =Z1
0
ci di −ψ .
These in u n imply ha agg ega e household sa ings equal go e nmen -p o ided liquid-
i y, labo demand equals supply in e iciency uni s, and he inal good is used o consump ion
and p ice adjus men cos s.18
Calib a ion The model is sol ed using he sequence-space Jacobians app oach pionee ed
by Aucle e al. (2021).19 Fo echnical de ails, we e e he eade o hei pape . The
calib a ion is also mos ly aken om Aucle e al. (2021) and summa ised in Table 1. The
main di e ence om hei calib a ion is ha he e we also choose B o a ge a pe cen age
o HTM agen s in s eady s a e o 20%. This is o ensu e a compa ison ac oss di e en
models and/o calib a ions keeping ixed he s eady s a e p opo ion o cons ained agen s.
The o he di e ence is ha , in he baseline calib a ion, we allow o some bo owing in
equilib ium (a=−0.5).20 This calib a ion implies an income-weigh ed agg ega e MPC in
s eady s a e equal o 12.5%.
18As is well known, p ice adjus men cos s don’ ma e in linea ized models.
19See Baye and Lue icke (2020) o an al e na i e solu ion me hod.
20The implied βand φa e almos iden ical o hei s, while ou calib a ion equi es a lowe amoun o liquidi y
compa ed o hei model, whe e hey se B= 5.6 and ob ain MPC = 11% and HTM = 17%. All he esul s
p esen ed in his sec ion do no change subs an ially i we use hei o iginal calib a ion.
21
Pa ame e Value Ta ge
Households
βDiscoun ac o 0.98 = 0.005
φDisu ili y o labo 0.78 H= 1
σEIS .5
νIn e se F isch 2
aBo owing cons ain -0.5
ρeAu oco ela ion o ea nings 0.966
σeC oss-sec ional s d o log ea nings 0.5
Fi ms
µS eady-s a e ma kup 1.2
κSlope o Phillips cu e 0.1
Policy
BBond supply 3.84 HTM = 0.2
ϕTaylo ule coe icien on in la ion 1.5
Mone a y Policy Shock
ρϵPe sis ence 0.61
σϵS anda d De ia ion 0.0025
Disc e iza ion
nePoin s in Ma ko chain o e7
naPoin s on asse g id 500
¯aUppe limi on asse g id 150
Un a ge ed S eady S a e
MPC % agg ega e Ma ginal P opensi y o Consume 12.5%
Table 1: One-asse HANK calib a ion
Ra ionalizing he empi ical e idence Mapping he disc e e income s a es o hei
posi ions in he s eady-s a e dis ibu ion e eals ha bo h consump ion and labo supply, on
a e age, inc ease wi h income–consis en wi h empi ical e idence. Howe e , poo e house-
holds nea he bo owing cons ain end o wo k mo e han iche ones due o s onge
income e ec s and igh e liquidi y.21
Figu e 6shows he impulse esponses o a mone a y policy igh ening, assuming ha ϵ
ollows an AR(1) p ocess wi h pe sis ence ρϵ= 0.61 and s anda d de ia ion σϵ= 0.0025. The
uppe panels display he dynamics o he in e es a e shock, in la ion, and agg ega e ou pu .
The lowe panels highligh he e ogeneous labo supply and consump ion esponses ac oss
income g oups. Poo e households (e.g., P0 2, P2 11, and P11 34) inc ease labo supply
ollowing he shock, consis en wi h a dominan income e ec d i en by igh bo owing
cons ain s. In con as , highe -income households educe labo supply, e lec ing s anda d
subs i u ion e ec s. Consump ion alls ac oss all g oups bu declines mo e sha ply o lowe -
income households, as expec ed. These esul s demons a e ha an o - he-shel HANK
21See appendix G.1 o de ails.
22
Figu e 6: Impulse esponses o a 25 basis poin s Mone a y Policy Shock
23
hou s wo ked, while weal hie households educe hem. Impo an ly, he model allows us
o decompose he labo supply esponses in o unde lying channels, and we ind ha he
coun e cyclical beha io a he bo om o he income dis ibu ion is p ima ily d i en by
income e ec s. We sys ema ically s udy how hese he e ogeneous labo supply esponses
a y ac oss di e en calib a ions, changing he elas ici y o in e empo al subs i u ion and
he bo owing limi , and compa e hem o a HANK model wi h homogeneous labo supply.
Ou quan i a i e analysis shows ha he e ogenei y in labo supply beha io al e s he
agg ega e MPC implied by he model. In pa icula , s onge coun e cyclical labo supply
esponses a he bo om o he dis ibu ion imply a lowe MPC ela i e o he HANK model
whe e e e ybody supplies he same amoun o hou s, highligh ing how labo supply ac as
an addi ional ma gin h ough which households can smoo h consump ion. This addi ional
adjus men ma gin also has subs an ial e ec s on he dynamics o he model. Impo an ly
i educes he sac i ice a io- he cumula i e ou pu loss pe poin o in la ion educ ion-
aced by he mone a y au ho i y. Ac oss model calib a ions, we ind ha inco po a ing
he e ogeneous labo supply leads o a sys ema ically lowe sac i ice a io, wi h educ ions
o up o 35%, implying ha disin la iona y policies may en ail smalle ou pu cos s han
p e iously es ima ed i his ma gin is accoun ed o .
The e o e, igno ing labo supply he e ogenei y isks o e s a ing households’ ulne abili y
o mone a y policy shocks and mischa ac e izing bo h he dis ibu ional and agg ega e e ec s
o policy in e en ions.
30
Re e ences
Ambe g, N., Jansson, T., Klein, M. and Rogan ini-Picco, A.: 2022, Fi e Fac s abou he
Dis ibu ional Income E ec s o Mone a y Policy Shocks, Ame ican Economic Re iew:
Insigh s .
Ami -Ahmadi, P., Ma hes, C. and Wang, M.-C.: 2021, Wha Does Mone a y Policy Do To
Di e en People?, Mimeo.
Ande sen, A. L., Johannesen, N., Jø gensen, M. and Peyd ´o, J.-L.: 2021, Mone a y policy
and inequali y, Economics Wo king Pape s 1761, Depa men o Economics and Business,
Uni e si a Pompeu Fab a.
A h eya, K., Owens, A. and Schwa zman, F.: 2017, Does edis ibu ion inc ease ou pu ?
he cen ali y o labo supply, Quan i a i e Economics 8(3), 761–808.
Aucle , A.: 2019, Mone a y Policy and he Redis ibu ion Channel, Ame ican Economic
Re iew 109(6), 2333–2367.
Aucle , A., Ba d´oczy, B., Rognlie, M. and S aub, L.: 2021, Using he sequence-space
jacobian o sol e and es ima e he e ogeneous-agen models, Econome ica 89(5), 2375–
2408.
URL: h ps://onlinelib a y.wiley.com/doi/abs/10.3982/ECTA17434
Aucle , A. and Rognlie, M.: 2020, Inequali y and Agg ega e Demand, Mimeo.
Aucle , A., Rognlie, M. and S aub, L.: 2024, The in e empo al keynesian c oss, Jou nal
o Poli ical Economy 132(12), 4068–4121.
URL: h ps://doi.o g/10.1086/732531
Bai, J. and Ng, S.: 2002, De e mining he Numbe o Fac o s in App oxima e Fac o Models,
Econome ica 70(1), 191–221.
Banbu a, M., Giannone, D. and Reichlin, L.: 2007, Bayesian a s wi h la ge panels, CEPR
Discussion Pape s 6326.
Ba igozzi, M., Lippi, M. and Luciani, M.: 2021, La ge-dimensional Dynamic Fac o Mod-
els: Es ima ion o Impulse-Response Func ions wi h I(1) coin eg a ed ac o s, Jou nal o
Econome ics 221(2), 455–482.
Baue , M. D. and Swanson, E. T.: 2023, A Reassessmen o Mone a y Policy Su p ises and
High-F equency Iden i ica ion, NBER Mac oeconomics Annual 37(1), 87–155.
URL: h ps://ideas. epec.o g/a/ucp/macann/doi10.1086-723574.h ml
31
Baye , C., Bo n, B. and Lue icke, R.: 2024, Shocks, ic ions, and inequali y in us business
cycles, Ame ican Economic Re iew 114(5), 1211ˆa47.
URL: h ps://www.aeaweb.o g/a icles?id=10.1257/ae .20201875
Baye , C. and Lue icke, R.: 2020, Sol ing he e ogeneous agen models in disc e e ime wi h
many idiosync a ic s a es by pe u ba ion me hods, Quan i a i e Economics 11, 1253–
1288.
Bilbiie, F.: 2008, Limi ed asse ma ke s pa icipa ion, mone a y policy and (in e ed) ag-
g ega e demand logic, Jou nal o Economic Theo y 140(1), 162–196.
Bilbiie, F.: 2020, The new keynesian c oss, Jou nal o Mone a y Economics 114, 90–108.
Bilbiie, F. O.: 2024, Mone a y policy and he e ogenei y: An analy ical amewo k*, The
Re iew o Economic S udies p. dae066.
URL: h ps://doi.o g/10.1093/ es ud/ dae066
Bilbiie, F. O., Monacelli, T. and Pe o i, R.: 2013, Public deb and edis ibu ion wi h
bo owing cons ain s, The Economic Jou nal 123(566), F64–F98.
B oe , T., K ame , J. and Mi man, K.: 2022, The Cu ious Incidence o Mone a y Policy
Shocks Ac oss he Income Dis ibu ion, Wo king Pape Se ies 416, S e iges Riksbank
(Cen al Bank o Sweden).
Cano a, F. and Fe oni, F.: 2022, Mind he gap! S ylized dynamic ac s and s uc u al
models, Ame ican Economic Jou nal: Mac oeconomics o hcoming.
Cesa ini, D., Lindq is , E., No owidigdo, M. J. and ¨
Os ling, R.: 2017, The e ec o weal h
on indi idual and household labo supply: E idence om swedish lo e ies, Ame ican
Economic Re iew 107(12), 3917–46.
Cloyne, Fe ei a, C. and Su ico, P.: 2020, Mone a y Policy when Households ha e Deb : New
E idence on he T ansmission Mechanism, Re iew o Economic S udies 87(1), 102–129.
Coglianese, J., Olsson, M. and Pa e son, C.: 2025, Mone a y policy and he labo ma ke :
A quasi-expe imen in sweden, Ame ican Economic Re iew .
Das, M., Hambu , J., Hellwig, K. and Sp ay, J.: 2025, Labo supply e ec s o mone a y
policy: E idence om aus alian mo gage holde s, mimeo.
De Gio gi, G. and Gambe i, L.: 2017, Business cycle luc ua ions and he dis ibu ion o
consump ion, Re iew o Economic Dynamics 23, 19 – 41.
Del Can o, F., G igsby, J., Qian, E. and Walsh, C.: 2025, A e in la iona y shocks eg essi e?
a easible se app oach, The Qua e ly Jou nal o Economics .
32
Dy da, S. and Ped oni, M.: 2022, Op imal iscal policy in a model wi h uninsu able idiosyn-
c a ic shocks, Re iew o Economic S udies .
Fo ni, M. and Gambe i, L.: 2014, Su icien in o ma ion in s uc u al VARs, Jou nal o
Mone a y Economics 66(C), 124–136.
Gal´ı, J., Sme s, F. and Wou e s, R.: 2012, Unemploymen in an es ima ed new keynesian
model, NBER Mac oeconomics Annual 26, 329–360.
Ge ke, R., Giesen, S., Lozej, M. and Ro ge , J.: 2024, On household labou supply in
s icky-wage HANK models, Technical epo .
Ge le , M. and Ka adi, P.: 2015, Mone a y policy su p ises, c edi cos s, and economic
ac i i y, Ame ican Economic Jou nal: Mac oeconomics 7(1), 44–76.
Goloso , M., G abe , M., Mogs ad, M. and No go odsky, D.: 2023, How ame icans espond
o idiosync a ic and exogenous changes in household weal h and unea ned income*, The
Qua e ly Jou nal o Economics 139(2), 1321–1395.
URL: h ps://doi.o g/10.1093/qje/qjad053
G a es, S., Huck eld , C. and Swanson, E. T.: 2023, The labo demand and labo supply
channels o mone a y policy, mimeo.
Gue ie i, V. and Lo enzoni, G.: 2017, C edi C ises, P ecau iona y Sa ings, and he Liq-
uidi y T ap, The Qua e ly Jou nal o Economics 132(3), 1427–1467.
Holm, M. B., Paul, P. and Tischbi ek, A.: 2021, The ansmission o mone a y policy unde
he mic oscope, Jou nal o Poli ical Economy 129(10), 2861–2904.
Hube , P. and Sa ignac, F.: 2024, The he e ogeneous e ec s o mone a y policy on labo
income: Disen angling he ex ensi e and in ensi e ma gins, Jou nal o Poli ical Economy
Mac oeconomics 2(4), 687–719.
URL: h ps://doi.o g/10.1086/732469
Kaplan, G., Moll, B. and Violan e, G. L.: 2018, Mone a y policy acco ding o hank, Ame -
ican Economic Re iew 108(3), 697–743.
Ka z, L. F. and Mu phy, K. M.: 1992, Changes in ela i e wages, 1963ˆa1987: Supply and
demand ac o s, The Qua e ly Jou nal o Economics 107(1), 35–78.
Kehoe, P. J., Lopez, P., Pas o ino, E. and Salgado, S.: 2020, On he Dynamic E ec s o
Mone a y Policy Wi h He e ogeneous Agen s, Mimeo.
McKay, A., Nakamu a, E. and S einsson, J.: 2016, The powe o o wa d guidance e isi ed,
Ame ican Economic Re iew 106(10), 3133–58.
33
Me ens, K. and Ra n, M. O.: 2013, The Dynamic E ec s o Pe sonal and Co po a e Income
Tax Changes in he Uni ed S a es, Ame ican Economic Re iew 103(4), 1212–47.
Mi anda-Ag ippino, S. and Ricco, G.: 2021, The ansmission o mone a y policy shocks,
Ame ican Economic Jou nal: Mac oeconomics 13(3), 74–107.
Pijoan-Mas, J.: 2006, P ecau iona y sa ings o wo king longe hou s?, Re iew o Economic
Dynamics 9(2), 326–352.
Ro embe g, J. J.: 1982, S icky p ices in he uni ed s a es, Jou nal o Poli ical Economy
90(6), 1187–1211.
S ock, J. H. and Wa son, M. W.: 2008, Wha ’s new in econome ics - ime se ies, Lec u e 7,
Na ional Bu eau o Economic Resea ch, Inc.
Wol , C.: 2021, The Missing In e cep : A Demand Equi alence App oach, Na ional Bu eau
o Economic Resea ch WP29558.
34
Online Appendix
A Household le el da a
A.1 US: CPS Su ey
Fo he US, indi idual le el da a is ob ained om he cu en popula ion su ey (CPS) o
each mon h om 1985 o 2019. In he benchma k case we use he uni o m CPS ou going
o a ion g oup ex ac s c ea ed by CEPR da a. Ou benchma k measu e o hou s is he
a iable hou slw (Hou s las week, all jobs). We use he a iable w as ou measu e o eal
hou ly wage. This is he ecommended consis en eal wage a iable ac oss he sample we
employ. Fo de ails see he CEPR FAQ.
Figu e A.1: Cha ac e is ics along he wage dis ibu ion. A e age ac oss he sample.
Hou s
<20
>20 and <=40
>40 and <=60
>60 and <=80 >80
wage dis ibu ion
0
20
40 College
<20
>20 and <=40
>40 and <=60
>60 and <=80 >80
wage dis ibu ion
0
0.2
0.4
0.6
0.8
Male
<20
>20 and <=40
>40 and <=60
>60 and <=80 >80
wage dis ibu ion
0
0.2
0.4
0.6
Whi e
<20
>20 and <=40
>40 and <=60
>60 and <=80 >80
wage dis ibu ion
0
0.2
0.4
0.6
0.8
Age
<20
>20 and <=40
>40 and <=60
>60 and <=80 >80
wage dis ibu ion
0
20
40
35
Figu e A.2: P opo ion o indi iduals in di e en indus ies.
Indus y by wage
<20 >20 and <=40 >40 and <=60 >60 and <=80 >80
wage dis ibu ion
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
ag
min
con
man
ade
ans
in o
in
p o
heal h
un
o he
public
No es: The p opo ions a e a e aged o e he sample. ag is Ag icul u e, o es y, ishing and hun ing. min is
mining, con is cons uc ion, man is manu ac u ing, ade is wholesale and e ail ade, ans is anspo and
u ili ies, in o is in o ma ion, in is inancial ac i i ies, p o is p o essional and business se ices, heal h is educa ion
and heal h se ices, un is leisu e and hospi ali y, o he is o he se ices and public deno es public adminis a ion.
Figu e A.1 p o ides in o ma ion ega ding he cha ac e is ics o he ea nings dis ibu ion.
Responden s in he igh ail o he dis ibu ion end o be olde , be e educa ed, a e likely
o wo k longe hou s and be whi e and Male. Figu e A.2 shows indus y o employmen
ac oss he wage dis ibu ion. A he le ail, indus ies such as wholesale and e ail ade,
heal h, leisu e and manu ac u ing a e impo an .
As discussed abo e, we also employ he longi udinally ma ched e sion o CPS p o ided by
he Kansas Fed. We employ he a iable wagepe h clean82 as ou measu e o hou ly ea nings.
This se ies is de la ed by CPI. The change in hou s is cons uc ed as he di e ence in he
log o hou s82 and hou s82 m12 whe e he o me deno es a consis en se ies cons uc ed
36
using ac ual o usual hou s and he la e is he 12 mon h lag o his a iable. We cons uc
he a e o ansi ion om employmen using he labou ma ke s a us a iable ml 76 and
he 12 mon h lag o his a iable ml 76 m12. As discussed in he ex , we impu e ea nings
o indi iduals ha a e no employed o do no epo ea nings da a. To do his we eg ess
ea nings on expe ience and indi idual cha ac e is ics including gende , occupa ion, indus y
o employmen , geog aphical cha ac e is ics (US s a e, me opoli an indica o ) and ace.
The i ed alues om his eg ession a e used o ob ain impu ed ea nings.The coe icien s
o he eg ession a e shocked and used o p oduce p edic ed alues. These a e assigned
o indi iduals wi h missing ea nings da a using p edic ed mean ma ching. We p oduce 5
eplica es, wi h he inal impu ed da a aken o be he mean ac oss hese eplica ions.
A.2 Compa ison wi h agg ega e da a
The op panel o Figu e A.3 compa es agg ega e ac ual hou s om he CPS o a measu e o
mon hly hou s om he Bu eau o Labou s a is ics (a e age weekly hou s o p oduc ion and
non-supe iso y employees). The CPS da a cap u es he main mo emen s in he agg ega e
da a ai ly well.
Figu e A.3: Compa ison o su ey based o al hou s (blue) wi h agg ega e (o ange).
1990 1995 2000 2005 2010 2015
-0.025
-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015 Annual G ow h o hou s wo ked
Cons uc ed om CPS
AWHNONAG
37
B P io s and es ima ion o he FAVAR
The FAVAR model is de ined by he ollowing equa ions:
Xi =ci+biτ+ ΛiF +ξi (B.1)
Y =c+
P
X
j=1
βjY −j+u (B.2)
co (u ) = Σ = A0A′
0(B.3)
Whe e Y =R
F
| {z }
N×1
,R deno es he 1 yea in e es a e, i= 1,2, .., M deno es he
c oss-sec ional dimension o he panel da a-se Xi while = 1,2, .., T is he dimension.
As desc ibed in Ba igozzi e al. (2021), he ac o s can be consis en ly es ima ed using a
p incipal componen s (PC) es ima o . In pa icula , he ac o loadings a e es ima ed ia
PC analysis o he i s di e enced da a ∆Xi . Wi h hese in hand, he ac o s a e es ima ed
as ˆ
F =1
Mˆ
Λ′˜
X . He e,Λ is he ma ix o ac o loadings, ˜
X is gi en by (x1 , x2 , ..., xM )
whe e xi =Xi −ˆci−ˆ
biτNo e ha Ba igozzi e al. (2021) desc ibe a p ocedu e o check i
he i h se ies con ains a linea end and ha ˆ
biis di e en om ze o.
Gi en he es ima ed ac o s, he VAR in equa ions B.2 is es ima ed using a Bayesian
me hods.
B.1 P io s
Deno e he a coe icien s as B= ec ([β1, β2, .., βP, c]). We ollow Banbu a, Giannone and
Reichlin (2007) and use a Na u al Conjuga e p io implemen ed ia dummy obse a ions.
The p io s a e implemen ed by he dummy obse a ions yDand xD ha a e de ined as:
yD=
diag(γ1s1...γnsn)
κ
0N×(P−1)×N
diag (s1...sn)
..............
0EX×N
, xD=
JP⊗diag(s1...sn)
κ0NP×EX
..............
0N×(NP)+EX
..............
0EX×NP IEX ×1/c
(B.4)
whe e JP=diag(1,2, ..., P), γ1 o γndeno e he p io mean o he pa ame e s on he
i s lag ob ained by es ima ing indi idual AR(1) eg essions, s1 o snis an es ima e o he
a iance o he endogenous a iables ob ained indi idual AR(1) eg essions, κmeasu es he
igh ness o he p io on he au o eg essi e VAR coe icien s, and cis he igh ness o he
p io on he emaining eg esso s. We se κ= 0.2 and c= 1000. We also implemen p io s
on he sum o coe icien s (see Banbu a e al. (2007)). The dummy obse a ions o his
p io a e de ined as:
˜yD=diag (γ1µ1...γnµn)
τ,˜xD=(11×P)⊗diag(γ1µ1...γnµn)
τ0N×EX (B.5)
whe e µiis he sample a e age o he i h a iable. As in Banbu a e al. (2007) we se
τ= 10κ. The o al numbe o dummy obse a ions is TD.
38
B.2 MCMC algo i hm
Banbu a e al. (2007) show ha pos e io dis ibu ion can be w i en as:
g(Σ|Y)∼iW ¯
Σ, TD+2+T−K(B.6)
g(B|Σ, Y )∼N¯
B, Σ⊗X′
∗X∗−1(B.7)
whe e iW deno es he in e se Wisha dis ibu ion, Kdeno es he numbe o eg esso s in
each equa ion o he VAR model. No e ha Y∗=
Y
yD
˜yD
and X∗=
X
xD
˜xD
,Xcollec s
he eg esso s, and
˜
B=X′
∗X∗−1X′
∗Y∗
¯
B= ec ˜
B
¯
Σ = Y∗−X∗˜
B′Y∗−X∗˜
B
Pos e io d aws can be easily gene a ed by d awing Σ om he ma ginal dis ibu ion in B.6
and hen b om he condi ional dis ibu ion in equa ion B.7. We se he numbe o d aws
o 21,000 wi h a bu n-in o 1,000. We e ain e e y second d aw a e he bu n-in pe iod.
C IV Iden i ica ion
Fo a gi en d aw o B, Σ and u , we ob ain he i s column o A0by using he p ocedu e
p oposed by Me ens and Ra n (2013). We assume ha he ins umen is ele an and
exogenous:
co (m , ε1 ) = α
co m , ε−
= 0
whe e ε1 deno es he s uc u al shock o in e es ha is o de ed i s o con enience, while
ε−
ep esen all emaining shocks and ε =ε1 ε−
. Re-w i ing he ele ance and exo-
genei y condi ions in ec o o m:
E(m ε′
) = [α0] (C.1)
E(m ε′
A′
0)=[α0]A′
0(C.2)
E(m u′
) = αa0(C.3)
whe e a0is a (1 ×R) ec o co esponding o he i s ow o A′
0(hence i s column o
A0). An es ima e o E(m u′
) =
E(m u′
1 )
E(m u′
2 )
.
.
E(m u′
N )
′
can be easily ob ained by using a linea
39
Figu e D.5: P opo ion o indi iduals wi h educa ion less han high school (LTHS), high school
(HS), some college (somecollege), college and ad anced deg ees (ad anced).
Educa ion Ac oss he wage dis ibu ion
<5
<10
<15
<20
<25
<30
<35
<40
<45
<50
<55
<60
<65
<70
<75
<80
<85
<90
<95
wage dis ibu ion
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
LTHS
HS
Somecollege
college
ad anced
No es: The p opo ions a e a e aged o e he sample.
46
Figu e D.6: Employees by le el o educa ion in each indus y o indi iduals ea ning below he
20 h pe cen ile o wage.
No es: P opo ion o indi iduals wi h educa ion less han high school (LTHS), high school (HS), some college
(somecollege), college and ad anced deg ees (ad anced). The p opo ions a e a e aged o e he sample.
47
Figu e D.7: IRF o hou s a he 6m h ho izon o a mone a y con ac ion
48
D.3 Response o eal wages
Figu e D.8 epo s he esponses o eal wages bo h in he agg ega e and ac oss di e en
pe cen iles o he income dis ibu ion. We ind ha o indi iduals in he middle and uppe
pa s o he dis ibu ion, eal wages and hou s wo ked bo h decline ollowing a mone a y
igh ening. This pa e n is consis en wi h a educ ion in labo demand.
In con as , o indi iduals in he lowe end o he dis ibu ion, eal wages also decline,
bu hou s wo ked inc ease. This di e gence sugges s ha , in his g oup, an inc ease in
labo supply is he dominan o ce d i ing he esponse. The o e all pic u e hus poin s
o he e ogeneous labo ma ke adjus men mechanisms ac oss he income dis ibu ion, wi h
labo supply playing a mo e p ominen ole o lowe -income wo ke s.
Figu e D.8: Responses o eal wages a e mone a y policy igh ening.
0 50
-2.5
-2
-1.5
-1
-0.5
0
0 50
-3
-2
-1
0
0 50
-2.5
-2
-1.5
-1
-0.5
0
0 50
-3
-2
-1
0
0 50
-3
-2
-1
0
0 50
-3
-2
-1
0
6 mon hs
2 yea s
No es: The op panels show he esponse o he eal wages a di e en quin ile o he wage dis ibu ion and he
agg ega e eal wages ( igh mos panel). The bo om panel displays he dis ibu ion o he esponses o eal wages
six mon hs ( ed) and wo (blue) yea s a e he shock. Shaded a eas (solid lines) 68 (90)% con idence se s.
49
E Sample o Indi iduals Employed Con inuously o Th ee Mon hs
Figu e E.1: IRF o hou s o Indi iduals Employed Con inuously o Th ee Mon hs
6 m hs
2 yea s
We es ic he sample o wo ke s employed o h ee mon hs by se ing he a iables
ml 76,ml 76 m1, and ml 76 m2 equal o 1. Fo his exe cise we use he Kansas Fed
ex ac . Figu e E.1 shows ha Hou s o his coho go up a he le ail o he Wage
dis ibu ion a e a con ac iona y mone a y policy shock.
50
F TANK model wi h Bo owe s and Sa e s
A simple way o cha ac e ize he household he e ogenei y and e ain analy ical ac abili y is
o conside a wo agen s model whe e one ype o agen s a e ne bo owe (indexed wi h B)
and he o he ne sa e (indexed wi h S), as in Bilbiie, Monacelli and Pe o i (2013).27 The
key ea u es o his class o models a e ha bo owe s a e mo e impa ien han sa e s, ha e
no access o go e nmen bonds, and can bo ow up o a limi . Following Bilbiie (2020) and
McKay e al. (2016) we conside an economy whe e he mone a y au ho i y can e ec i ely
choose he eal a e which makes he algeb a simple .
The key equa ions o he log-linea ized model a e epo ed in able F.1.
Log-linea ized Condi ions
1: Labo Supply S νˆ
HS
= ˆw −σ−1ˆcS
2: Eule S ˆcS
= ˆcS
+1| −σˆ
R −ˆ
Π +1|
3: Labo Supply B νˆ
HB
= ˆw −σ−1ˆcB
4: Budge cons ain B ˆcB
γ+¯
D(ˆ
R −1−ˆ
Π ) = ˆw +ˆ
HB
5: Phillips Cu e ˆ
Π =βE ˆ
Π +1 +κˆw
6: Agg ega e C ˆc =λγˆcB
+ (1 −λγ)ˆcS
7: Agg ega e B ˆc =ˆ
H =λˆ
HB
+ (1 −λ)ˆ
HS
8: Taylo Rule ˆ
R =ˆ
Π +1| +ϵm
Table F.1: Log-linea ized Condi ions o Sa e s/Bo owe s model
This sec ion ollows a di e en no a ional con en ion ha he one used o he HANK
model in he main ex . Small case le e s ep esen eal a iables while capi al le e s ep-
esen nominal a iables o hou s wo ked. A ’ha ’ on op o he a iable deno es pe cen age
de ia ions om he s eady s a e. The assump ion ha bo owe s discoun mo e u u e con-
sump ion, βB< βS=β, implies ha hey become ne bo owe in equilib ium wi h he
bo owing limi ( ¯
D) always binding.28 γis a s eady s a e pa ame e which cap u es he
consump ion inequali y be ween bo owe s and sa e s, i.e. γ=cB/c = 1 + ¯
D(β−1) <1.
No ice ha when ¯
D= 0 →γ= 1 and he model is iden ical o he one wi h Sa e s and
HTM consume s.29 In his model, a ac ion o agen s a e no on he Eule equa ion and
canno op imize in e empo ally. Impo an ly, and di e en ly om a model wi h HTM ha
a e no ne bo owe s, a change in he nominal a e will ha e an impac no only on he
27See hei pape o de ails on he model de i a ion. Rela i e o hem we simpli y i u he by abs ac ing
om go e nmen deb , expendi u e, and edis ibu ion conce ns. We ollow Bilbiie (2020) and assume ha he e
is a p oduc ion subsidy ha induces ma ginal cos p icing which implies ha he s eady s a e o ma ginal cos s is
1 which simpli ies subs an ially he s eady s a e and he log-linea ized condi ions.
28No e ha in equa ion (4) in able F.1 ¯
Dis e ec i ely di ided by he s eady s a e o o al income/consump ion.
Bu his is =1 one in his simple se up.
29In ha se up, Bilbiie (2008) showed ha σ < 1 is he condi ion o he hou s o hand- o-mou h agen s o
inc ease ollowing a decline in hei income.
51
ime consump ion and labo supply decision bu also on he + 1 decisions because he
deb epaymen s a + 1 depend on he ime in e es a es.
Unde mild condi ions, bo owe s ha e an incen i e o inc ease hei labo supply a e
an in e es a e hike; his is o malized in he ollowing p oposi ion.
P oposi ion 1 Unde SADL (λ < 1
1+ν(1+ ¯
Dκ)) and σ < 1+ ¯
Dκ
γ, a a e hike a ime induces
an inc eases in he bo owe s labo supply bo h a ime and + 1.
The p oo is on he nex page. A he co e o his esul , we ha e ha consump ion and
he labo supply o he bo owe s mo e in opposi e di ec ions. This can be app ecia ed when
combining he ime + 1 op imal esponse o bo owe s in e ms o consump ion and labo
supply a e a mone a y policy shock which a e30
ˆ
HB
+1 =¯
D(νλγσ −1 + λ)
(νλ −1 + λ)(1 + γνσ)ϵm
,
ˆcB
+1 =νσ ¯
D
(νλ −1 + λ)(1 + γνσ)ϵm
.
Combining he la e wo equa ions we ha e ha
ˆcB
+1 =νσ
νλγσ −1 + λˆ
HB
+1.
The nume a o is posi i e. No ice ha i he condi ions o P oposi ion 1hold, we ha e
λ < 1
1+ν(1+ ¯
Dκ)<1
1+νσγ , which implies ha (1 −λ−νσλγ)>0; hence he denomina o is
nega i e.
De i a ion o P oposi ion 1Recall ha ϵm
+j= 0 o j > 0 and ϵm
= 0. This implies
ha om +2 onwa d he economy is back o s eady s a es and all quan i ies a e ze o. This
means also ha ˆ
R +j=ˆ
Π +j+1| +j o j > 0, which implies ha
ˆcS
+1 = 0
The sa e labo supply becomes
νˆ
HS
+1 = ˆw +1
Using he bo owe s BC we ha e
ˆcB
+1γ+¯
D(ˆ
R −ˆ
Π +1) = ˆw +1 +ˆ
HB
+1
ˆcB
+1γ+¯
D(ˆ
Π +1| +ϵm
−ˆ
Π +1) = (1 + ν)ˆ
HB
+1 + 1/σˆcB
+1
1/σˆcB
+1 =1 + ν
γσ −1ˆ
HB
+1 −¯
D
γσ −1ϵm
30The ime op imal esponses a e analogous bu mo e in ol ed. To ease he no a ion we only discuss he ime
+ 1 decisions.
52
no ice ha in absence o shocks in + 1 ˆ
Π +1 =ˆ
Π +1| . Combining he o labo supply
condi ions we ha e
νˆ
HS
+1 +σ−1ˆcS
+1 =νˆ
HB
+1 +σ−1ˆcB
+1
νˆ
HS
+1 =νˆ
HB
+1 +1 + ν
γσ −1ˆ
HB
+1 −¯
D
γσ −1ϵm
νˆ
HS
+1 =1 + γνσ
γσ −1ˆ
HB
+1 −¯
D
γσ −1ϵm
Combining he agg ega e condi ions we ha e
λγˆcB
+1 + (1 −λγ)ˆcS
+1 =λˆ
HB
+1 + (1 −λ)ˆ
HS
+1
λγσ 1 + ν
γσ −1ˆ
HB
+1 −¯
D
γσ −1ϵm
=λˆ
HB
+1 + (1 −λ)1 + γνσ
ν(γσ −1) ˆ
HB
+1 −¯
D
ν(γσ −1)ϵm
ˆ
HB
+1 (1 + ν)λγσ
γσ −1−λ−(1 −λ)1 + γνσ
ν(γσ −1)=¯
Dλγσ
γσ −1−(1 −λ)¯
D
ν(γσ −1)ϵm
ˆ
HB
+1 (1 + ν)νλγσ −λν(γσ −1) −(1 −λ)(1 + γνσ)
ν(γσ −1) =νλγσ −(1 −λ)
ν(γσ −1) ¯
Dϵm
ˆ
HB
+1 νλ(γσ +νγσ −γσ + 1) −(1 −λ)(1 + γνσ)
ν(γσ −1) =νλγσ −(1 −λ)
ν(γσ −1) ¯
Dϵm
ˆ
HB
+1
(νλ −1 + λ)(1 + γνσ)
ν(γσ −1) =νλγσ −(1 −λ)
ν(γσ −1) ¯
Dϵm
which yield o
ˆ
HB
+1 =νλγσ −1 + λ
(νλ −1 + λ)(1 + γνσ)¯
D ϵm
(F.1)
This implies ha bo owe consump ion a ime + 1 is
ˆcB
+1 =σ(1 + ν)
γσ −1ˆ
HB
+1 −σ¯
D
γσ −1ϵm
=σ(1 + ν)
γσ −1
νλγσ −1 + λ
(νλ −1 + λ)(1 + γνσ)¯
D ϵm
−σ¯
D
γσ −1ϵm
=ϵm
σ¯
D
γσ −1(1 + ν)(−1 + λ(1 + νγσ)) −(νλ −1 + λ)(1 + γνσ)
(νλ −1 + λ)(1 + γνσ)
=ϵm
σ¯
D
γσ −1−(1 + ν) + (1 + ν)λ(1 + νγσ)−νλ(1 + γνσ) + (1 −λ)(1 + γνσ)
(νλ −1 + λ)(1 + γνσ)
=ϵm
σ¯
D
γσ −1−1−ν+1+γνσ
(νλ −1 + λ)(1 + γνσ)
=ϵm
νσ ¯
D
(νλ −1 + λ)(1 + γνσ)
53
and wages
ˆw +1 =νˆ
HB
+1 + 1/σˆcB
+1
=ννλγσ −1 + λ
(νλ −1 + λ)(1 + γνσ)¯
D ϵm
+ϵm
ν¯
D
(νλ −1 + λ)(1 + γνσ)
=ϵm
¯
Dν νλγσ −1 + λ+ 1
(νλ −1 + λ)(1 + γνσ)
=ϵm
¯
Dνλ
νλ −1 + λ
and in la ion
ˆ
Π +1 =βˆ
Π +2| +1 +κˆw +1 =ϵm
¯
Dνλκ
νλ −1 + λ
Now, we a e in a posi ion o sol e o ime . Sol ing he Eule equa ion o wa d we ha e
ˆcS
=−σϵm
F om he NKP we ha e an exp ession o oday in la ion
ˆ
Π =βˆ
Π +1| +κˆw =ϵm
¯
Dνλκβ
νλ −1 + λ+κˆw
Using he bo owe s BC we ha e
γˆcB
+¯
D(ˆ
R −1−ˆ
Π ) = ˆw +ˆ
HB
γˆcB
−ϵm
¯
D¯
Dνλκβ
νλ −1 + λ−¯
Dκν ˆ
HB
−¯
Dκ/σˆcB
=νˆ
HB
+ 1/σˆcB
+ˆ
HB
which yields o
ˆcB
=σ(1 + ν(1 + ¯
Dκ))
γσ −1−¯
Dκ ˆ
HB
+¯
D2σνλκβ
e1e0
ϵm
whe e e1=γσ −1−¯
Dκ and e0=νλ −1 + λ. Combining he labo supply decision we ha e
νˆ
HS
+σ−1ˆcS
=νˆ
HB
+σ−1ˆcB
νˆ
HS
−ϵm
=νˆ
HB
+1 + ν+ν¯
Dκ
γσ −1−¯
Dκ ˆ
HB
+¯
D2νλκβ
(γσ −1−¯
Dκ)(νλ −1 + λ)ϵm
which yields o
ˆ
HS
=1 + νγσ
ν(γσ −1−¯
Dκ)ˆ
HB
+¯
D2νλκβ +e0e1
νe0e1
ϵm
whe e e1=γσ −1−¯
Dκ and e0=νλ −1 + λ. Combining he agg ega e condi ions we ha e
λγˆcB
+1 + (1 −λγ)ˆcS
+1 =λˆ
HB
+1 + (1 −λ)ˆ
HS
+1
λγ σ(1 + ν(1 + ¯
Dκ))
γσ −1−¯
Dκ ˆ
HB
+¯
D2σνλκβ
e1e0
ϵm
+ (1 −λγ)[−σϵm
]
=λˆ
HB
+1 + (1 −λ)1 + νγσ
ν(γσ −1−¯
Dκ)ˆ
HB
+¯
D2νλκβ +e0e1
νe0e1
ϵm
54
ˆ
HB
λγνσ(1 + ν(1 + ¯
Dκ))
ν(γσ −1−¯
Dκ)−λν(γσ −1−¯
Dκ)
ν(γσ −1−¯
Dκ)−(1 −λ)(1 + νγσ)
ν(γσ −1−¯
Dκ)
=ϵm
σ(1 −λγ)−γλ ¯
D2σνλκβ
e1e0
+ (1 −λ)¯
D2νλκβ +e0e1
νe0e1
Focusing on e ms inside he le hand side squa e b acke we ha e
ν−1e−1
1νλ(σγ +σγν(1 + ¯
Dκ)) −λν(γσ −1−¯
Dκ)−(1 −λ)(1 + νγσ)
=ν−1e−1
1νλ(1 + σγν)(1 + ¯
Dκ)−(1 −λ)(1 + νγσ)
=ν−1e−1
1(1 + σγν)(νλ(1 + ¯
Dκ)−1 + λ)
Focusing on e ms inside he igh hand side squa e b acke we ha e
(νe0e1)−1νσ(1 −λγ)e0e1−νγλ(¯
D2σνλκβ) + (1 −λ)( ¯
D2νλκβ +e0e1)
= (νe0e1)−1νσ(1 −λγ)e0e1−νγλ(¯
D2σνλκβ) + (1 −λ)¯
D2νλκβ + (1 −λ)e0e1
= (νe0e1)−1e0e1(νσ(1 −λγ)+1−λ)−νγλσ ¯
D2νλκβ + (1 −λ)¯
D2νλκβ
= (νe0e1)−1νσe0e1+e0e1(1 −λ−νσλγ) + ¯
D2νλκβ(1 −λ−νσλγ)
= (νe0e1)−1νσe0e1+ (e0e1+¯
D2νλκβ)(1 −λ−νσλγ)
Rea anging e ms we ha e
ˆ
HB
=νσe0e1+ (e0e1+¯
D2νλκβ)(1 −λ−νσλγ)
(1 + σγν)(νλ(1 + ¯
Dκ)−1 + λ)(νλ −1 + λ)ϵm
(F.2)
whe e e1=γσ−1−¯
Dκ and e0=νλ−1+λ. While s ill no e y ac able we can de i e a se
o su icien condi ions such ha he la e exp ession becomes posi i e. These condi ions
a e
1. λ < 1
1+ν(1+ ¯
Dκ)
2. σ < 1+ ¯
Dκ
γ
Condi ion 1 implies ha e1= [νλ(1+ ¯
Dκ)−1+λ]<0; his implies also ha (νλ−1+λ)<0.
Thus, i condi ion 1 holds, he denomina o is posi i e. Condi ion 2 implies ha e0=
(σγ −1−¯
Dκ)<0; his implies also ha e0e1>0. Mo eo e , i condi ion 2 hold, i is he
case ha
λ < 1
1 + ν(1 + ¯
Dκ)<1
1 + νσγ
The la e implies ha (1 −λ−νσλγ)>0. The e o e also he nume a o is posi i e. These
condi ions imply also ha he coe icien in (F.1) is posi i e.
55
Figu e G.12: Labo Supply Policy Func ions wi h a= 0 and σ= 0.5
Figu e G.13: Consump ion policy unc ion wi h a= 0 and σ= 0.5
62
Figu e G.14: Impulse esponses o a 25 basis poin s Mone a y Policy Shock wi h a= 0 and
σ= 0.5
Figu e G.15: Labo Supply Policy Func ions wi h a= 0 and σ= 0.25
63
Figu e G.16: Consump ion policy unc ion wi h a= 0 and σ= 0.25
Figu e G.17: Impulse esponses o a 25 basis poin s Mone a y Policy Shock wi h a= 0 and
σ= 0.25
64
G.3 IRFs decomposi ions HANK
Figu e G.18: Decomposi ion o consump ion Response in HANK.
No es: No e: Pe cen de ia ion in agg ega e labo hou s om s eady s a e, decomposed in o ma ginal channels:
in e es a e, di idends, axes, and wages.
65
Figu e G.19: Decomposi ion o consump ion Responses by Income Bin — HANK.
No es: No es: Each panel shows pe cen de ia ion in labo hou s om s eady s a e, decomposed in o ma ginal
channels: in e es a e, di idends, axes, and wages.
66
G.4 HANK s HANK-HomL
Figu e G.20: Impac esponse o labo supply ac oss income s a es in HANK s he agg ega e
impac esponse in HANK-HomL
67
Figu e G.21: Di ec s indi ec e ec s o mone a y policy on agg ega e consump ion in HANK s
HANK-HomL o baseline σ.
68
Table 2in he main ex highligh s ha accoun ing o he e ogenei y in labo supply has
quan i a i ely signi ican implica ions, no ably lowe ing he economy’s a e age MPC. This,
in u n, a ec s he ansmission o mone a y policy. To illus a e his we compa e he e he
dynamic esponses o a con ac iona y mone a y policy shock in he HANK and HANK-
HomL models. Figu e G.22 epo s agg ega e impulse esponses o in e es a es, in la ion,
ou pu , wages, di idends, and he sha e o hand- o-mou h households. Agg ega e p ice and
di idend dynamics a e b oadly simila ac oss models, as expec ed gi en hei iden ical i m
and policy s uc u e. Howe e , wo no able di e ences eme ge. Fi s , he HTM sha e ises
mo e and emains ele a ed longe in HANK, sugges ing ha lexible labo supply ampli ies
inequali y in down u ns.Second, agg ega e consump ion declines by 16% less on impac in
HANK compa ed o HANK-HomL. This mu ed esponse e lec s he addi ional ma gin o
adjus men p o ided by he e ogeneous labo supply, which allows low-income households o
bu e shocks h ough wo k e o . We explo e he dis ibu ional esponses ha d i e his
esul nex .
Figu e G.23 shows ha , by cons uc ion, labo supply esponses a e almos iden ical o
he median agen ype (34–66%). A he bo om o he dis ibu ion, agen s inc ease hei
labo supply, consis en wi h ou empi ical e idence. In con as , uppe -middle and op-
income households educe hei labo supply mo e in he HANK model compa ed o HANK-
HomL. This highligh s ha he le ail o he labo supply dis ibu ion is esponsible o he
mu ed ampli ica ion o mone a y policy on agg ega e demand. Figu e G.24 shows how hese
labo esponses ansla e in o indi idual consump ion dynamics. The consump ion decline
is mos se e e a he bo om o he dis ibu ion in bo h models, bu he d op is smalle and
less pe sis en in he HANK model. A he op o he dis ibu ion, consump ion esponses
a e e y simila ac oss models.
Ano he way o quan i y he ole o di e en households ac oss he income dis ibu ion in
shaping he agg ega e e ec s o mone a y policy is o compu e hei espec i e con ibu ions
o he o e all impac on consump ion and labo supply, measu ed in basis poin de ia ions.
Figu e G.25 p esen s he esul s.
69
Figu e G.22: Impulse Responses o Co e Mac oeconomic Va iables: HANK-HomL s HANK
Figu e G.23: Labo Supply Responses Ac oss he Income Dis ibu ion
70
Figu e G.24: Consump ion Responses Ac oss he Income Dis ibu ion
Figu e G.25: Con ibu ion o each income bin o he agg ega e esponse.
(a) Hou s
(b) Con ibu ion o each income bin o he ag-
g ega e esponse.
(c) Consump ion
71
