Low pass-through from inflation expectations to income growth expectations: Why people dislike inflation

Hajdini, Ina e al.
Wo king Pape
Low pass- h ough om in la ion expec a ions o income
g ow h expec a ions: Why people dislike in la ion
IDB Wo king Pape Se ies, No. IDB-WP-1672
P o ided in Coope a ion wi h:
In e -Ame ican De elopmen Bank (IDB), Washing on, DC
Sugges ed Ci a ion: Hajdini, Ina e al. (2025) : Low pass- h ough om in la ion expec a ions o
income g ow h expec a ions: Why people dislike in la ion, IDB Wo king Pape Se ies, No. IDB-
WP-1672, In e -Ame ican De elopmen Bank (IDB), Washing on, DC,
h ps://doi.o g/10.18235/0013365
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/309192
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/3.0/igo/
Low Pass- h ough om In la ion Expec a ions
o Income G ow h Expec a ions:
Why People Dislike In la ion
Ina Hajdini
Edwa d S. Kno ek II
John Lee
Ma hieu Pedemon e
Robe Rich
Raphael Schoenle
WORKING PAPER No IDB-WP-1672
In e -
A
me ican De elopmen Bank
Depa men o Resea ch and Chie Economis
Janua y 2025
* Fede al Rese e Bank o Cle eland
** Mo ning Consul
*** In e -Ame ican De elopmen Bank
**** B andeis Uni e si y
Low Pass- h ough om In la ion Expec a ions
o Income G ow h Expec a ions:
Why People Dislike In la ion
Ina Hajdini*
Edwa d S. Kno ek II*
John Lee **
Ma hieu Pedemon e***
Robe Rich*
Raphael Schoenle****
In e -Ame ican De elopmen Bank
Depa men o Resea ch and Chie Economis
Janua
y
2025
Ca aloging-in-Publica ion da a p o ided by he
In e -Ame ican De elopmen Bank
Felipe He e a Lib a y
Low pass- h ough om in la ion expec a ions o income g ow h expec a ions:
why people dislike in la ion / Ina Hajdini, Edwa d S. Kno ek II, John Lee ,
Ma hieu Pedemon e, Robe Rich, Raphael Schoenle.
p. cm. — (IDB Wo king Pape Se ies ; 1672)
Includes bibliog aphical e e ences.
1. In la ion (Finance)-Uni ed S a es. 2. Income dis ibu ion-E ec o in la ion
on-Uni ed S a es. 3. Income dis ibu ion-Ma hema ical models-Uni ed S a es.
I. Hajdini, Ina. II. Kno ek, Edwa d S. III. Lee , John. IV. Pedemon e, Ma hieu. V.
Rich, Robe . VI. Schoenle, Raphael. VII. In e -Ame ican De elopmen Bank.
Depa men o Resea ch and Chie Economis . VIII. Se ies.
IDB-WP-1672
h p://www.iadb.o g
Copy igh © 2025 In e -Ame ican De elopmen Bank ("IDB"). This wo k is subjec o a C ea i e
Commons license CC BY 3.0 IGO (h ps://c ea i ecommons.o g/licenses/by/3.0/igo/legalcode). The
e ms and condi ions indica ed in he URL link mus be me and he espec i e ecogni ion mus be
g an ed o he IDB.
Fu he o sec ion 8 o he abo e license, any media ion ela ing o dispu es a ising unde such license
shall be conduc ed in acco dance wi h he WIPO Media ion Rules. Any dispu e ela ed o he use o
he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he
Uni ed Na ions Commission on In e na ional T ade Law (UNCITRAL) ules. The use o he IDB's name
o any pu pose o he han o a ibu ion, and he use o IDB's logo shall be subjec o a sepa a e
w i en license ag eemen be ween he IDB and he use and is no au ho ized as pa o his license.
No e ha he URL link includes e ms and condi ions ha a e an in eg al pa o his license.
The opinions exp essed in his wo k a e hose o he au ho s and do no necessa ily e lec he iews o
he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun ies hey ep esen .
Abs ac
Using a la ge, na ionally ep esen a i e su ey o US consume s, we es ima e a causal
20 pe cen pass- h ough om in la ion expec a ions o income g ow h expec a ions o he
a e age consume , wi h conside able he e ogenei y in pass- h ough associa ed wi h socio-
demog aphic ac o s. The esul s also indica e ha highe in la ion expec a ions cause an in-
c ease in consume s’ likelihood o sea ch o highe -paying jobs bu do no change he likeli-
hood o asking o a aise, sugges ing ha consume s ecognize signi ican wage igidi y wi h
hei cu en employe . In a calib a ed sea ch-and-ma ching model, we ind ha demand and
supply shocks combined wi h incomple e pass- h ough p oduce a s ong nega i e ela ionship
be ween expec ed in la ion and expec ed u ili y. Taken oge he , he su ey esul s and model
analysis p o ide a labo ma ke accoun o why people dislike in la ion.
JEL classi ica ions: E31, E24, E71, C83
Keywo ds: In la ion, Wage-p ice spi al, Expec a ions, Randomized con olled ial
The andomized con olled ial is egis e ed a he AER RCT Regis y (#AEARCTR-0009062). The iews
exp essed he e a e solely hose o he au ho s and do no necessa ily e lec he iews o he Fede al Rese e Bank o
Cle eland o he Fede al Rese e Sys em. We hank Osca A ce, Alex Bick (discussan ), Ma k Bils, Oli ie Coibion,
Julia Co onado, Jon Faus , Juan He eño, Olena Kos yshyna, Emiliano Lu ini (discussan ), Ay¸segül ¸Sahin, Maa en
an Rooij (discussan ), Michael Webe and semina pa icipan s a a ious ins i u ions o aluable commen s and
discussions. We hank Ca oline Smi h om Mo ning Consul o he wo k ielding he expe imen s.

1 In oduc ion
The apid economic eco e y in he Uni ed S a es om he COVID-19-induced ecession was
cha ac e ized by he highes in la ion a es seen in he las o y yea s. These high in la ion ead-
ings we e accompanied by inc eases in in la ion expec a ions and s ong wage gains in igh labo
ma ke s, aising conce ns abou po en ial eedback in o expec a ions o o he mac oeconomic ag-
g ega es, in pa icula in he labo ma ke (e.g., Cu in (2022); Blancha d (1986)).1Howe e , disen-
angling he causal e ec o in la ion expec a ions on income g ow h expec a ions is challenging
because hese concep s should be ela ed in gene al equilib ium.2Mo e gene ally, while he li -
e a u e on expec a ions o ma ion has made p og ess in examining how expec a ions espond o
in o ma ion ea men s, i has made less p og ess in unde s anding how indi iduals pe cei e he
ela ionship be ween di e en expec ed a iables.
This pape sheds new ligh on hese issues by in es iga ing he causal ela ionship be ween in-
la ion expec a ions and income g ow h expec a ions, and how hose expec a ions a ec labo ma -
ke decisions, in he con ex o a andomized con olled ial (RCT) o a la ge, na ionally ep esen-
a i e su ey o he US popula ion. Th ee indings eme ge. Fi s , in la ion expec a ions causally
a ec income g ow h expec a ions bu pass- h ough om he o me o he la e is a less han
one- o -one, on he o de o 20 pe cen . Second, highe in la ion expec a ions inc ease he p oba-
bili y ha consume s will sea ch o a new job ha pays mo e bu do no a ec he likelihood ha
hey will nego ia e o a highe wage wi h hei cu en employe . This inding is consis en wi h
consume s’ ecogni ion o subs an ial nominal wage igidi y wi h hei cu en employe .3Thi d, a
canonical sea ch-and-ma ching model calib a ed o i ou empi ical indings shows ha low pass-
h ough om expec ed in la ion o expec ed income g ow h is consis en wi h consume s’ belie s
ha highe u u e in la ion will educe hei expec ed u ili y. Taken oge he , he su ey esul s
and model analysis o malize a labo ma ke channel unde lying consume s’ a e sion o in la ion.
Ou empi ical indings p ima ily come om a su ey expe imen ielded by he decision in-
elligence company Mo ning Consul in Ma ch 2022, a a ime when in la ion expec a ions and
in la ion conce ns we e s a ing o ise o no able le els, and be o e in la ion had clea ly begun
1See Lo enzoni and We ning (2023) o a heo e ical analysis on he wage-p ice spi al in he con ex o a New
Keynesian model.
2See, o example, We ning (2022) o a discussion on he challenges ela ed o pinning down he pass- h ough
om in la ion expec a ions o cu en in la ion.
3The ecen inding o Jäge e al. (2023) ha wo ke s w ongly ancho hei belie s abou ou side op ions on hei
cu en wage speaks o he ole ha pe cei ed nominal wage igidi y plays o wo ke s’ income g ow h expec a ions.
1
o u n back down.4The embedded expe imen al module consis ed o ou pa s. The i s pa
elici ed in la ion expec a ions and income g ow h expec a ions o e he nex 12 mon hs p io o
any expe imen al ea men s.5The second pa consis ed o an RCT ha allowed us o p o ide in-
o ma ion o esponden s on wo key objec s, in la ion o income g ow h, o de e mine he causal
ela ionship be ween in la ion expec a ions and income g ow h expec a ions. In pa icula , we
andomly assigned in o ma ion ea men s o six g oups: one con ol g oup; one placebo g oup;
h ee g oups ha ecei ed di e en in o ma ion on in la ion; and one g oup ha ecei ed in o -
ma ion on wage g ow h, which is he p ima y sou ce o income g ow h o mos consume s.
Following he ea men s, he hi d pa o he expe imen e-elici ed in la ion expec a ions
and income g ow h expec a ions. This expe imen al s ep allows us o measu e how consume s’
pos e io expec a ions o in la ion and income g ow h eac o in o ma ion ea men s while condi-
ioning on hei p io expec a ions. Speci ically, he esul ing exogenous, expe imen ally induced
a ia ion in pos e io in la ion expec a ions hen allows us o es ima e he causal impac on income
g ow h expec a ions. This pass- h ough es ima e is causal as i conside s exogenous a ia ion o
in la ion o income, while allowing esponden s o o m hei own men al models abou he p e-
cise ansmission mechanism as highligh ed in (And e e al.,2022a). We ind ha a 1.0 pe cen age
poin inc ease in in la ion expec a ions inc eases income g ow h expec a ions, bu only by 0.2 pe -
cen age poin —implying an expec ed dec ease in eal income g ow h o 0.8 pe cen age poin .
The e is, howe e , conside able a ia ion in pass- h ough associa ed wi h socio-demog aphic
cha ac e is ics. While he ex en o pass- h ough is high and s a is ically signi ican o highe -
income esponden s, i is low and s a is ically insigni ican o lowe -income esponden s. This
inding is consis en wi h he o me g oup belie ing i is be e p o ec ed om inc eases in ex-
pec ed in la ion han he la e g oup. We also ind a la ge pass- h ough poin es ima e o male
esponden s han o emale esponden s. This esul is consis en wi h e idence ha highligh s
di e en cha ac e is ics in he labo ma ke o women and men. Fo ins ance, Biasi and Sa sons
(2022) ind ha in he Uni ed S a es, women engage less equen ly in pay nego ia ions, whe eas
Ca d, Ca doso, and Kline (2016) ind ha , in Po ugal, women a e less likely o wo k a i ms
whe e wo ke s ha e high ba gaining powe . I is impo an o no e, howe e , ha pass- h ough
emains incomple e and is well below one- o -one in all cases.
4We also pe o med a pilo in Janua y 2022 as well as a ollow-up exe cise in Sep embe 2022 ha con i ms he
Ma ch esul s.
5We an obus ness exe cises wi h di e en p io ques ion wo dings o mi iga e any conce n abou pa icula
wo ding, inding no s a is ical di e ence depending on he speci ic p io . This exe cise includes using he poin
es ima e ques ion o he NY Fed Su ey o Consume Expec a ions ins ead o he ques ion ielded by Mo ning Consul .
2
Finally, he ou h pa o ou su ey asks esponden s abou he likelihood o pu suing di e -
en labo ma ke ac ions o e he ollowing yea o inc ease hei incomes and po en ially o se
he e ec s o in la ion. Exploi ing he exogenous a ia ion in belie s once again o es ima ion
pu poses, we ind ha highe in la ion expec a ions mode a ely inc ease he pe cei ed likelihood
ha an indi idual applies o ano he job paying a highe wage.6Howe e , highe in la ion ex-
pec a ions do no inc ease he pe cei ed likelihood o wo o he labo ma ke ac ions: wo king
longe hou s o asking o a aise om a cu en employe . These esul s sugges ha consume s’
men al models (see, o example, And e e al. (2022a) o a gene al s udy o subjec i e models)
inco po a e he belie ha he e is a high deg ee o nominal wage igidi y associa ed wi h hei
cu en employe .
In e p e ing ou indings s uc u ally, h ough a model, can p o ide u he economic insigh ,
which, in pa icula , can help unde s and why people may dislike in la ion. We show how his
conclusion can a ise by adap ing a ela i ely s anda d New Keynesian model wi h sea ch-and-
ma ching in labo ma ke s as in Mo ensen and Pissa ides (1994), acing ou he expec ed u ili y
implica ions o demand and supply shocks. A cen al inding om his model se up is ha wage
igidi y s ands ou in cap u ing a labo ma ke channel explana ion why people dislike in la ion.
While cu en wo k by A ouzi e al. (2024) and Gue ei o e al. (2024) p o ides u he mic o-
ounded modeling ad ances in his con ex , ou modeling exe cise also simply gauges he ex en
o which a canonical model can i ou empi ical ac s.
The model ea u es se e al ic ions. Mo i a ed by he obse a ion ha he p o ision o pub-
licly a ailable in o ma ion mo es consume s’ expec a ions which con as s wi h a ull-in o ma ion
a ional expec a ions iew o he wo ld, we allow o s icky in o ma ion in he in la ion expec a-
ions o ma ion p ocess, simila o Mankiw and Reis (2002). In a no el in e p e a ion o how in o -
ma ion s ickiness can play ou , we calib a e he deg ee o in o ma ion s ickiness o be consis en
wi h he es ima ed e ec ha new in o ma ion om ea men s has on ou esponden s’ in la ion
expec a ions.7In addi ion, ma ching ou su ey indings equi es sluggish wage adjus men . We
model wage igidi y as in equen nominal wage enego ia ion in a Cal o (1983) ashion, cali-
b a ed o ma ch ou es ima e o empi ical pass- h ough as a momen .8Finally, o cap u e he im-
pac o in la ion expec a ions on labo ma ke ac ions, we assume ha wo ke s who canno enego-
6Pilossoph and Ryngae (2022) ind ha highe in la ion expec a ions a e co ela ed wi h he likelihood ha
wo ke s will sea ch o o he jobs in he sho e m.
7We ind ha he deg ee o in o ma ion s ickiness is abou 0.28.
8We would no e ha , in con as o he expe imen in ou su ey, i is impossible wi hin he model se ing o isola e
he causal e ec o in la ion expec a ions on income g ow h expec a ions.
3
ia e hei wages and who apply o o he jobs due o highe in la ion expec a ions gene a e an ou -
side con ac wi h ce ain y. This wage-push ac o pu s upwa d p essu e on hei nominal wage
wi h he cu en employe , wi h an elas ici y ha we calib a e o ma ch ou empi ical indings.
Gi en his se up, ou model analysis hen highligh s he esponses o key mac oeconomic a i-
ables o a posi i e demand shock and a posi i e (ad e se) supply shock ha a e mean o b oadly
cap u e he p e ailing in la iona y dis u bances in he US economy a he ime o ou su ey in
ea ly 2022. A cen al inding ha eme ges is ha nominal wage igidi y plays a c ucial ole in
d i ing he dynamics o mac oeconomic a iables wi hin he model. When we subjec he model
o an in la iona y demand shock, his igidi y causes a decline in eal wages ela i e o a coun e -
ac ual o ull pass- h ough om in la ion expec a ions o expec ed nominal wage g ow h. When
we subjec he model o an in la iona y supply shock, s icky wages empe he mo emen s in eal
wages compa ed o he coun e ac ual o ull pass- h ough. In bo h cases, he esponses o eal
wages unde impe ec pass- h ough help o ampli y he luc ua ions in ou pu and consump ion,
gene a ing addi ional ola ili y in he wake o he o iginal shock. Mo eo e , he model p edic s
ha g ea e wage igidi y p oduces a s onge nega i e ela ionship be ween in la ion expec a-
ions and expec ed u ili y ega dless o whe he we look a supply o demand shocks. This la e
esul is pa icula ly impo an because i iden i ies a labo ma ke channel ha can explain why
consume s dislike in la ion.9
The es o he pape is o ganized as ollows. Sec ion 2discusses wo k ela ed o ou pape .
Sec ions 3and 4p o ide a de ailed desc ip ion o ou expe imen and i s implemen a ion, espec-
i ely. Sec ion 5explains ou iden i ica ion s a egy and p esen s he main empi ical indings.
Sec ion 6gi es a b ie o e iew o he model, ou calib a ion s a egy, and he mac oeconomic
implica ions o he model. Sec ion 7concludes.
2 Li e a u e Re iew
Ou pape is mos ela ed o a se ies o pape s ha s udy he issue o public a i udes abou
in la ion, speci ically why consume s and i ms associa e highe in la ion expec a ions wi h lowe
ou pu and well-being. Fo example, Shille (1997) and Candia, Coibion, and Go odnichenko
(2020) p o ide e idence consis en wi h ou esul s, hough ha e idence is non-causal. O he
s udies, such as Sa ignac e al. (2021), look a he ela ionship be ween i ms’ in la ion expec a-
9Following a one- ime exogenous shock occu ing in he p esen pe iod, ealized in la ion hpe iods ahead co-mo es
wi h cu en expec a ions abou in la ion hpe iods ahead, in he p esence o in o ma ion s ickiness. The e o e, wi hin
he con ex o he model, we e e o he wo a iables in e changeably.
4
Responden s hen selec om h ee op ions, illing in he pe cen ages i hey selec (1) o (3),
while (2) is coded as ze o:
1. Inc ease by %;
2. S ay abou he same; and
3. Dec ease by %.
Ou pos e io in la ion expec a ions ques ion uses he ollowing wo ding:
“In he nex yea , do you hink p ices in gene al will inc ease, dec ease, o s ay abou he same?”
I esponden s’ answe s indica ed an expec ed inc ease o dec ease, hen hey we e subsequen ly
asked o p o ide a quan i a i e pe cen age esponse.
As no ed in he abo e design desc ip ion, his ques ion is pu posely sligh ly di e en om he
p io in la ion expec a ions ques ion, by asking di ec ly abou p ices and by i s ocus on p ices in
gene al a he han he p ices o which consume s a e exposed. We expec ha answe s o his
ques ion will no be iden ical o he indi ec measu e o in la ion expec a ions. Ne e heless, e-
sponses should be s ongly posi i ely co ela ed, which allows us o cap u e he (po en ial change
in) pos e io belie s a e an in o ma ion ea men . In e ms o he in e p e a ion o he esul s in
he es o he pape , all exe cises in e ms o in la ion will conside his ques ion as he pos e io .
While he ICIE ques ion is used o measu e esponden s’ p io s, i is he sys ema ic de ia ion be-
ween he ea ed g oups and he con ol g oup in e ms o his agg ega e in la ion ques ion ha is
ou main ou come o in e es . The p io only se es as a con ol a iable o measu e he in o ma-
ion se o he esponden s. In ac , as shown in Appendix E, ou esul s a e no a ec ed i we selec
he canonical NY Fed in la ion expec a ions ques ion o elici he p io in la ion expec a ions.
Ou second p io ques ion elici s income g ow h expec a ions. The second ques ion is he ol-
lowing:
“Do you expec you income o inc ease, dec ease, o s ay abou he same o e he nex 12 mon hs?”
The ques ion comes wi h he same op ions as in he p e iously desc ibed pos e io ques ion. I e-
sponden s indica ed hey expec hei income o inc ease o dec ease, hen hey we e subsequen ly
asked o p o ide a quan i a i e pe cen age esponse.
Ou pos e io income g ow h expec a ions ques ion uses he ollowing wo ding:
“Be ween Decembe 2022 and Decembe 2023, do you expec you income o inc ease, dec ease, o s ay
abou he same o e he nex 12 mon hs?”
Compa ed o he p io ques ion on income g ow h expec a ions, his ques ion mainly di e s
11

in i s e e ence o a ixed ime pe iod. This pe iod pa ially o e laps wi h he p e ious income
g ow h ques ion, so we expec ed a posi i e co ela ion wi h he p e ious ques ion gi en he o e -
lap as well as he ac ha many wages a e adjus ed in equen ly and a a pa icula ime o he
yea . In Sec ion 6, we conside hese pe iods o model and e alua e ou esul s.
Ques ions abou labo ma ke decisions ollow he elici a ion o all hese pos e io expec a-
ions, asking consume s:
“How likely a e you o do he ollowing o inc ease you income o e he nex h ee mon hs?”
We asked esponden s o p o ide answe s o h ee ac ions, choosing om he esponse se e y
likely,somewha likely,somewha unlikely, e y unlikely, o hey do no know. The ac ions we asked o
a e:
• Apply o a job(s) ha pays mo e
• Wo k longe hou s
• Ask o a aise
In addi ion o hese ac ions, an open-ended answe op ion eco ds any u he possibili ies ha
su ey esponden s migh o e .
5 Empi ical Analysis
This sec ion uses he expec a ions elici ed h ough he RCT o es ima e he causal impac o
in la ion expec a ions on income g ow h expec a ions as well as on he sho - e m plans a ound
labo ma ke decisions. Th ee main indings eme ge. Fi s , he pass- h ough o in la ion expec-
a ions o income g ow h expec a ions is posi i e and s a is ically signi ican bu less han uni y.
Second, es ima ed pass- h ough a ies ac oss esponden demog aphic cha ac e is ics, wi h some
e idence o s a is ically signi ican di e ences. Thi d, while highe in la ion expec a ions cause
consume s o epo a mode a ely highe p obabili y ha hey will sea ch o a highe -paying job,
hey do no inc ease he pe cei ed p obabili y o wo king mo e hou s o asking o a aise om a
cu en employe .
5.1 In la ion Expec a ions and Income G ow h Expec a ions
The analysis akes h ee s eps. Fi s , we e i y ha ou “pos e io ” ques ions cap u e in o ma-
ion simila o ha o he baseline p io ques ions. This inding alida es he choices o ques ion
wo ding agains he backd op o he design conside a ions ou lined abo e in he expe imen al
12
desc ip ion. Second, we es ablish which ea men s a ec he pos e io belie s. Las , we use he
esul s om he ea men s o in e he causal e ec o in la ion expec a ions on income g ow h
expec a ions, which yields ou main indings.
In he i s s ep, we es ima e wo speci ica ions ha ela e p io belie s o pos e io belie s. Fo
in la ion expec a ions, we es ima e he ollowing speci ica ion:
EihπPos e io
pi=α+βEihπP io
pi+εi(1)
whe e EihπP io
pideno es esponden i’s p io in la ion expec a ions om he ICIE ques ion and
EihπPos e io
pideno es he pos e io gene al p ice g ow h expec a ions in he nex yea . Fo income
g ow h expec a ions, we es ima e he ollowing speci ica ion:
EihπPos e io
yi=α+βEihπP io
yi+εi(2)
whe e EihπPos e io
yideno es he pos e io expec a ions o income g ow h be ween Decembe 2022
and Decembe 2023 and EihπP io
yideno es he p io income g ow h expec a ions o e he nex 12
mon hs.
When we es ima e eg essions in (1) and (2) o he ull sample o esponden s and he con-
ol g oup (a e winso izing 2.5 pe cen o he highes and lowes esponses o emo e ex eme
ou lie s), we ind ha he esponses o ou pos e io ques ions a e sys ema ically ela ed o he
esponses o he p io ques ions. Table 1 epo s he es ima ion esul s. As columns (1) and (4) in
Table 1show, we ind posi i e and s a is ically signi ican co ela ions be ween he p io and pos-
e io belie s o bo h in la ion expec a ions and income g ow h expec a ions o he ull sample.
This inding in pa icula alida es he choices o ques ion wo ding agains he backd op o he
design conside a ions ou lined abo e in he expe imen al desc ip ion.
Ou second s ep in es iga es he p ope ies o ou ea men s and hei e ec on he pos e io
in la ion expec a ions and pos e io income g ow h expec a ions. In he case o in la ion expec a-
ions, we es ima e he ollowing speci ica ion:
13
EihπPos e io
pi=α+βEihπP io
pi+
6
∑
j=2
γj
p×Tj
i+
6
∑
j=2
θi
p×Tj
i×EihπP io
pi+εi(3)
We es ima e a simila eg ession o income g ow h expec a ions:
EihπPos e io
yi=α+βEihπP io
yi+
6
∑
j=2
γj
y×Tj
i+
6
∑
j=2
θj
y×Tj
i×EihπP io
yi+εi(4)
whe e Tj
iis a dummy a iable ha is equal o 1 i esponden i ecei ed ea men jand 0 o he -
wise. The con ol g oup j=1 is he e e ence g oup.
Reg ession speci ica ions (3) and (4) ela e he pos e io belie o he p io belie and each o he
ea men s. Ideally, i he ea men ep esen s new in o ma ion o he esponden , hen p o iding
ha in o ma ion will elici a esponse and mo e he pos e io away om he p io . I ea men
jis e ec i e, hen we should expec a nega i e coe icien o θj
pand θj
yas he p io will ha e a
educed ole in explaining he pos e io o he ea ed g oup compa ed o he con ol g oup.
To es ima e speci ica ions (3) and (4), we un wo ypes o eg essions. Fi s , we conduc Hube -
obus eg essions, and second, we un immed eg essions, wi h he la e d opping 5 pe cen
o he bigges changes be ween indi iduals’ p io and pos e io belie s. Bo h ypes o eg essions
aim o emo e he in luence o ou lie s, especially hose ha display ex eme e isions. D awing
upon he common p ac ice in su ey analysis ( o example, Coibion, Go odnichenko, and Ropele
(2020b)), we iew he Hube - obus eg essions as ou p e e ed speci ica ion, wi h he immed
eg essions se ing mainly as a obus ness check.13
The es ima ion o (3) and (4)— epo ed in columns 2-3 and 5-6, espec i ely, in Table 1—shows
h ee esul s.14 Fi s , he e is a high co ela ion o he pos e io s wi h he p io s as in he abo e
i s s ep, e en a e con olling o ou lie s. Fo in la ion expec a ions, we ind ha a 1 pe cen age
poin inc ease in he p io belie s o he con ol g oup inc eases he pos e io belie s by a ound 0.51
pe cen age poin . This co ela ion be ween p io and pos e io o he con ol g oup is simila o
he one ound in o he simila household expe imen s, such as Coibion e al. (2019) o Coibion,
13Appendix Bimplemen s a hi d quan ile eg ession app oach, wi h esul s epo ed in Table 9.
14Figu e 6in Appendix Cshows he dis ibu ion o he p io and pos e io and Figu e 7in Appendix Cshows he
dis ibu ion o he pos e io o each ea men g oup. We obse e ounding (see Binde (2017)) in pa icula a ze o
as in o he su eys (43 pe cen o he p io , 32 pe cen o he pos e io ; see And ade, Gau ie , and Mengus (2023) o
p ope ies o ze o answe s). Hajdini e al. (2024) desc ibe in mo e de ail he dis ibu ion o he p io .
14
Go odnichenko, and Webe (2022), who ind a co ela ion o 0.54 and 0.66 using a dis ibu ional
ques ion as p io . This esul con i ms ha he ICIE measu e is a good p io o agg ega e in la-
ion expec a ions. In he case o income g ow h expec a ions, he co ela ion is e en highe and
associa es he same 1 pe cen age poin inc ease in p io belie s wi h an inc ease in he pos e io
belie s o be ween 0.78 and 0.96 pe cen age poin s.
Second, in e ms o he e ec o he ea men s, ou esul s show ha all o he ea men s o
in la ion expec a ions ha e a s a is ically signi ican e ec on he pos e io excep o he placebo,
as column 2 indica es. Mo eo e , he es ima ed coe icien s on he in e ac ed ea men and p io
a e nega i e, indica ing ha consume s who ecei e one o he ea men s place less weigh on
hei p io belie s. Column 3 shows simila esul s when we explici ly d op esponden s who
make ex eme changes be ween hei p io and pos e io belie s (o e 50 pe cen age poin s).
The magni ude o he es ima ed e ec s a ies ac oss ea men s. In pa icula , while he p io
in e ac ed wi h he ea men abou he Fede al Rese e’s in la ion a ge is nega i e and s a is-
ically signi ican , he coe icien is an o de o magni ude smalle compa ed o hose epo ed
o he p io in e ac ed wi h ea men s 3-5. As p e iously no ed, he p io in e ac ed wi h he
placebo does no gene a e a meaning ul e ec on he pos e io belie s compa ed o he con ol
g oup. These esul s a e no d i en by ou lie s.15
Thi d, in con as o in la ion expec a ions, he eg ession esul s show ha he ea men s ha e
li le e ec on he pos e io belie s o income g ow h expec a ions. Tha is, he e is a high co e-
la ion be ween he p io and pos e io belie s, meaning ha mos esponden s do no e ise hei
answe s. As a esul , he Hube - obus eg essions ail o un wi h he s anda d uning ac o due
o he small numbe o ou lie s ha can be d opped. When we use he minimum uning alue
o achie e con e gence, he esul s in column 5 indica e ha he ea men s gene ally exe li le
in luence on he pos e io belie s. Howe e , he same conclusion a ises o he immed eg es-
sions in which we elimina e esponden s who epo ed ex eme absolu e changes be ween hei
p io and hei pos e io belie s a o abo e he 95 h pe cen ile (10 pe cen age poin s). As shown
in column 6, we ind li le e ec om he in o ma ion ea men s, o he han he wage in la ion
ea men , on esponden s’ pos e io belie s o income g ow h expec a ions.
O e all, he esul s in Table 1sugges ha he in o ma ion ea men s ha e a g ea e e ec on
15As a obus ness check using o he echniques, Table 9in Appendix Bcon i ms hese esul s using quan ile
eg essions. Figu es 8and 9in Appendix Cplo he dis ibu ion o p io s and pos e io s and hei ela ionship wi h
he con ol g oup. We obse e big di e ences be ween he con ol g oup and ea men s 3, 4, and 5. The change in
he slope is smalle bu s a is ically signi ican o ea men 2. The con ol g oup and he placebo ha e a e y simila
dis ibu ion, wi h small di e ences ha a e i ele an in e ms o he magni ude and he dis ibu ion o he esponses.
15
in la ion expec a ions han on income g ow h expec a ions. The e idence o s ong p io s o in-
come g ow h expec a ions is consis en wi h he iew ha consume s a e e y a en i e o hei in-
come ajec o ies, which, as in Webe e al. (2023), makes hei o ecas s less esponsi e o in o ma-
ion ea men s abou agg ega e a iables. In he case o in la ion expec a ions, howe e , he ind-
ings sugges ha esponden s a e subjec o some ype o in o ma ion ic ions as all ea men s
con ain public in o ma ion. In ac , e en hough in la ion was high a he ime o he expe imen
and salien because o ele a ed news co e age and he no able impac o in la ion on consume s’
budge s, he esul s sugges ha consume s we e no ully in o med abou p ice de elopmen s.
While a de ailed in es iga ion in o in o ma ion ic ions is beyond he scope o his pape , he
obse ed ea men e ec s o e some insigh s in o how hese ic ions mani es in consume s’ in-
la ion expec a ions. F om ou ea men abou he Fed’s in la ion a ge , we see unce ain y abou
he Fed’s objec i es, a poin s udied in Coibion e al. (2020a). F om he SPF ea men , we see ha
he e is unce ain y abou he in la ion ou look. Mo eo e , he ac ha consume s con inue o pu
some weigh on hei p io s, e en a e he eceip o his in o ma ion, sugges s ha hey ace slug-
gish o cos ly in la ion expec a ions o ma ion, as in Coibion and Go odnichenko (2015). Finally,
while pas in la ion can a ec expec a ions in many ways, he ac ha i a ec s expec a ions o e
12 mon hs indica es o e -ex apola ion, as in Angele os, Huo, and Sas y (2021).
16

Table 1: E ec s o T ea men s on Expec a ions
(1) (2) (3) (4) (5) (6)
EihπPos e io
piEihπPos e io
piEihπPos e io
piEihπPos e io
yiEihπPos e io
yiEihπPos e io
yi
EihπP io
pi0.262*** 0.506*** 0.490***
(0.026) (0.006) (0.020)
EihπP io
yi0.775*** 0.775*** 0.960***
(0.048) (0.056) (0.010)
T2: Ta ge 0.126 -0.382 -0.292 -0.081
(0.138) (0.395) (0.296) (0.104)
T3: Wages 0.771*** -0.540 -0.445* 0.146
(0.153) (0.385) (0.256) (0.108)
T4: CPI 0.586*** -0.547 -0.271 -0.048
(0.150) (0.395) (0.277) (0.112)
T5: SPF 0.720*** -0.429 -0.147 -0.049
(0.149) (0.409) (0.338) (0.106)
T6: Placebo 0.498*** 0.482 -0.439 -0.182*
(0.148) (0.403) (0.274) (0.106)
T2 x P io -0.023*** -0.053* -0.116 -0.003
(0.008) (0.028) (0.081) (0.015)
T3 x P io -0.213*** -0.036 -0.037 -0.029*
(0.013) (0.028) (0.087) (0.017)
T4 x P io -0.258*** -0.065** -0.171* 0.013
(0.011) (0.027) (0.092) (0.013)
T5 x P io -0.281*** -0.084*** -0.061 0.005
(0.011) (0.030) (0.085) (0.016)
T6 x P io -0.008 -0.026 -0.103 0.006
(0.008) (0.026) (0.085) (0.015)
Cons an 5.667*** 1.343*** 4.223*** 0.925*** 0.925*** 0.274***
(0.337) (0.098) (0.291) (0.185) (0.217) (0.075)
Reg ession OLS Hube T immed OLS Hube T immed
Obse a ions 1,072 5,892 6,373 1,074 6,622 6,335
R-squa ed 0.236 0.786 0.432 0.604 0.555 0.922
No es: The able shows es ima es o equa ions 1and 2 ha ela e p io s and pos e io s, as well as es ima es o equa ions 3
and 4 ha gauge he e ec o ea men s and hei in e ac ion wi h p io belie s.
In he hi d and inal s ep, ou analysis uses in o ma ion om he e ec i e ea men s in Ta-
ble 1 o de i e an ins umen ha can be used o in e he causal e ec o in la ion expec a ions
on income g ow h expec a ions. Speci ically, we cons uc he ins umen o expec ed in la ion,
EihπPos e io
pi, using he ollowing speci ica ion:
17
EihπPos e io
pi=




∑j=2,4,5 γj
p×Tj
i+∑j=2,4,5 θj
p×Tj
i×EihπP io
pii Ti=2,4,5
0i Ti=1,6
(5)
whe e we exclude he ea men p o iding in o ma ion on wage in la ion (T3) because he e-
po ed esul s indica e i di ec ly a ec s income g ow h expec a ions. Based on he es ima ion
esul s om he Hube eg ession and he immed eg ession, we hen apply he ele an coe i-
cien s in column 2 and column 3 o o m an ins umen o each eg ession model. This app oach
is simila in spi i o he one in Coibion e al. (2019) ha uses he p io as an ins umen . In his
case, i only uses he a ia ion o he ea ed g oup. Any conce n abou he e ec o he p io ,
such as p iming because o he ques ion, on he pass- h ough is sha ed by bo h ea ed and con-
ol g oups. Because mul iple ea men s a e a ailable o us, we weigh hem acco ding o hei
impo ance in a ec ing he pos e io .16
This iden i ica ion s a egy is alida ed by a combina ion o ac o s ela ed o ou su ey de-
sign and he es ima ed e ec s o in o ma ion ea men s on expec a ions. Fi s , he assignmen
o in o ma ion ea men s o he esponden s in he su ey is andom. Second, we only use a -
ge ed, ca e ully wo ded ea men s con aining in o ma ion abou in la ion o o m he ins umen
o in la ion expec a ions. Thi d, and in line wi h he indings o o he RCT wo k on in la ion ex-
pec a ions, we ind ha p o iding people wi h publicly a ailable in o ma ion ea men s—e en
a a ime when in la ion was pa icula ly salien — ends o mo e hei belie s, hus in alida ing
ull-in o ma ion a ional expec a ions. Fou h, he esul s in Table 1demons a e ha he in la ion
ea men s in he i s s age only change he pos e io belie s o in la ion expec a ions bu do no
ha e an e ec on income g ow h expec a ions, which se es as a es o exclusion es ic ions in
he ins umen a ion. Mo eo e , ou inding ha in la ion- ela ed in o ma ion ea men s only a -
ec in la ion expec a ions is consis en wi h he heo e ical indings in Angele os and Lian (2023)
ha in o ma ion ic ions a enua e gene al equilib ium in e ence.
In e ms o income g ow h expec a ions, we conside bo h OLS and ins umen al- a iable (IV)
eg essions o he pos e io belie o income g ow h expec a ions on he p io belie o income
g ow h expec a ions and he pos e io belie o in la ion expec a ions, whe e he ins umen is de-
ined by equa ion (5). As p e iously discussed, he ins umen cap u es he exogenously induced
16Coibion, Go odnichenko, and Ropele (2020b) use he pas in la ion ea men as an ins umen . Un o una ely, we
do no ha e he ime se ies dimension ha hey ha e o gene a e enough p edic i e powe o he ins umen .
18
a ia ion in expec ed in la ion gene a ed om he assigned in o ma ion ea men (s).17 Because
only h ee o he ea men s a e used in cons uc ing he ins umen , he sample size o he e-
g essions is smalle compa ed o hose in Table 1.
Table 2: E ec o In la ion Expec a ions on Income G ow h Expec a ions
(1) (2) (3)
EihπPos e io
yiEihπPos e io
yiEihπPos e io
yi
EihπPos e io
pi0.085*** 0.203*** 0.168***
(0.014) (0.069) (0.045)
EihπP io
yi0.674*** 0.636*** 0.624***
(0.025) (0.033) (0.033)
Cons an 0.109 -0.805 -0.563*
(0.101) (0.521) (0.332)
Reg ession OLS IV IV
Sample All Hube T immed
F- es 120.584 572.491
Obse a ions 5,525 5,525 5,322
R-squa ed 0.558 0.539 0.538
No es: This able shows esul s om OLS and IV eg essions o he pos e-
io o income g ow h expec a ions on he p io o income g ow h expec a-
ions and he pos e io o in la ion expec a ions. Columns (2) and (3) use
IV, ins umen ing wi h
EiπPos e io
p. Column (2) uses he ins umen con-
s uc ed om he eg ession in (1) wi h Hube weigh s, whe eas column (3)
uses he ins umen cons uc ed om he immed eg ession in (1). The
es ima es o γj
pand θj
p, whe e j={2,4,5}, o bo h Hube and immed e-
g essions a e epo ed in Table 1. Robus s anda d e o s a e in pa en heses.
Table 2 epo s he esul s and highligh s a key empi ical inding o ou pape . Speci ically, we
documen a mode a e posi i e causal ela ionship om in la ion expec a ions o income g ow h
expec a ions ha e lec s only pa ial pass- h ough. As shown in column 1, he OLS eg ession
indica es ha in la ion expec a ions exhibi a e y low co ela ion wi h income g ow h expec a-
ions. Howe e , as shown in column 2, he Hube IV eg ession yields a no ably highe coe icien .
In pa icula , he es ima e implies ha a 1 pe cen age poin inc ease in in la ion expec a ions in-
c eases expec ed income g ow h by 0.2 pe cen age poin .18 The immed IV eg ession in column
3 shows a sligh ly lowe pass- h ough es ima e o 0.17, bu i is wi hin one s anda d de ia ion o
17As a obus ness check, we ha e also cons uc ed ins umen s by demog aphic g oups, such as by gende , allowing
o coe icien he e ogenei y in he γi
pand θi
p. We ind ha subsequen esul s a e no a ec ed.
18This pass- h ough di e s ma kedly om a co ela ion o 0.37 in he aw da a, as shown in Table 8in Appendix B.
This di e ence highligh s he impo ance o es ima ing a causal ela ionship as we do based on ou RCT.
19
he es ima e in column 2. Mo eo e , he ins umen displays a ela i ely high F- es s a is ic.
Looking mo e closely a he Hube IV eg ession, which is ou p e e ed speci ica ion, he e-
sul s sugges ha pass- h ough is conside ably lowe han one- o-one.19 Viewed di e en ly, he
same 1 pe cen age poin inc ease in in la ion expec a ions implies a 0.8 pe cen age poin educ ion
in expec ed eal income g ow h. A key akeaway om his inding is ha i sugges s consume s
associa e inc eases in expec ed in la ion wi h a ma ked decline in expec ed eal income g ow h
and o e s one eason o an a e sion o in la ion. Ou subsequen analysis will explo e how he
e ec o expec ed in la ion on eal income may in luence he labo ma ke ac ions o consume s
and u he shape hei a i udes owa d in la ion.
Finally, we show ha dis inc demog aphic cha ac e is ics a e associa ed wi h di e en de-
g ees o pass- h ough om in la ion expec a ions o income g ow h expec a ions. To do so, we
sepa a e ou sample based on he gende o su ey esponden s and hei sel - epo ed annual
income (less han $50,000, be ween $50,000 and $100,000, and mo e han $100,000). We epo OLS
and IV eg ession esul s in Table 3.
19In Table 17 in Appendix D, we calcula e he pass- h ough o each o he ea men s indi idually, a he han
combining hem as in Table 2. Each o he in la ion ea men s p oduces e y simila es ima es, poin ing o incomple e
pass- h ough in each ea men , wi h he magni udes simila o he main esul o 0.2.
20
o highe deg ees o nominal wage igidi y. Second, he mechanism we p opose o cap u e he
ela ionship be ween in la ion expec a ions and labo ma ke ac ions has a negligible e ec on he
mac oeconomic dynamics o he model; on a e age, consume s’ e o s o inc ease hei wages due
o highe in la ion expec a ions do no imp o e hei u ili y, eal wage, o consump ion. O e all,
we iew he lessons coming om his modeling exe cise as helping us u he unde s and why
consume s dislike cu en and u u e in la ion.
6.1 A Sea ch-and-Ma ching Model
We employ a New Keynesian model ea u ing a Mo ensen and Pissa ides (1994) ype o sea ch-
and-ma ching ic ions in labo ma ke s. We u he inco po a e a igh - o-manage ea u e as de-
eloped in T iga i (2006), whe e i ms and wo ke s ba gain o e nominal wages and hen wo ke s
gua an ee o supply he labo hou s demanded by i ms a he ba gained wage.23 A ma ched
i m-wo ke pai nego ia es wages in equen ly in a Cal o ashion. Finally, as in Ch is o el and
Kues e (2008), we accoun o i ms’ ixed cos s o main aining a job.24
The economy in he model is composed o ep esen a i e amilies ha make op imal decisions
on behal o hei membe s wi h espec o consump ion and one-pe iod iskless bond holdings.
The e a e h ee ypes o i ms: labo goods i ms p oduce a homogeneous labo in e media e
good; wholesale s use he labo good as an in e media e o p oduce di e en ia ed goods and ace
Cal o p ice igidi y; and e aile s bundle he di e en ia ed goods in o a homogeneous consump-
ion baske sold o households and he go e nmen . Mone a y policy se s he nominal in e es
a e ollowing a Taylo ule, and go e nmen spending is exogenous. Because hese pa s o he
model a e s anda d in he li e a u e and a e no cen al o ou pape , we desc ibe hem in mo e
de ail in Appendix F.
We now lay ou some key ea u es o he labo ma ke because hey di ec ly connec he model
wi h ou empi ical indings p esen ed in Sec ion 5. The ma ching p ocess be ween wo ke s and
23Fo ou pu poses, he igh - o-manage (RTM) amewo k di e s om, o ins ance, “e icien ba gaining" (EB),
whe e labo supply always equals labo demand. The ad an age o he RTM o e EB is ha i gene a es mo e ealis ic
mo emen s in in la ion dynamics, which acili a es ma ching he model-implied pass- h ough wi h he empi ical
es ima es. On he o he hand, RTM can igge luc ua ions in labo hou s ha a e la ge han wha is obse ed in he
da a. The inc eased a iabili y in labo hou s is a pa icula ly impo an limi a ion ha we e u n o below, especially
because ou empi ical esul s sugges ha consume s do no expec o inc ease hei hou s when hey aise hei
in la ion expec a ions. See de Walque e al. (2009) o an ins uc i e e iew o such ensions in his g oup o models.
24The RTM amewo k can coun e ac ually dampen he esponse o employmen in he ex ensi e ma gin, and, as
shown in Ch is o el and Kues e (2008), he p esence o a ixed cos ampli ies he esponse o unemploymen o e he
business cycle.
27

labo i ms is go e ned by a Cobb-Douglas unc ion:
m =σmuξ
1−ξ
(7)
whe e m a e ma ches o med in pe iod ;u is unemploymen ; a e acancies; ξ∈[0,1]is he
elas ici y o ma ching wi h espec o unemploymen ; and σm>0 is ma ching e iciency. Ma ches
become p oduc i e in he ollowing pe iod, so employmen in he ex ensi e ma gin e ol es ac-
co ding o
n = (1−µ)n −1+m −1(8)
whe e µ∈[0,1]is he employmen sepa a ion a e. Labo ma ke igh ness is de ined as:
θ =
u
(9)
Then, he p obabili ies ha a acancy is illed and ha an unemployed wo ke ma ches wi h a
i m a e, espec i ely,
q =m
,s =m
u
(10)
To ma ch ou indings in Table 1 ha p o iding an indi idual a ea men consis ing o pub-
licly a ailable in o ma ion a ime has an e ec on ou esponden s’ in la ion expec a ions, we
assume ha in la ion expec a ions a e subjec o s icky in o ma ion, such ha :
e
E ˆ
π +h= (1−λ)E ˆ
π +h+λe
E −1ˆ
π +h, o any h≥1 (11)
whe e E is he ull-in o ma ion a ional expec a ions ope a o , λ∈[0,1]deno es he p obabili y
ha ou agen s do no upda e hei in o ma ion se in pe iod , and ˆ
π is in la ion in log-linea
de ia ion om i s s eady-s a e alue.
To ma ch Fac 1, we assume ha agen s in he economy ace nominal wage igidi ies. I a
wo ke is no sepa a ed om employmen , she can ba gain he nominal wage o W∗
+1in pe iod
( +1)wi h p obabili y (1−γ)∈[0,1]. In con as , he nominal wage o he γsha e o wo ke s
who canno ba gain pa ially adjus s o pas in la ion such ha W +1=W (ew
πζw
¯
π1−ζw), whe e
ζw∈[0,1]deno es ime- a ying wage indexa ion o pas in la ion and ew
is a newly in oduced
wage-push ac o explained u he in he subsequen pa ag aph. In ou se up, di e en combi-
na ions o he nominal wage s ickiness pa ame e , γ, gene a e di e en le els o model-implied
28
pass- h ough om in la ion expec a ions o nominal wage g ow h expec a ions. This model ea-
u e allows us o s udy he mac o implica ions o Fac 2 and o a coun e ac ual scena io o uni
pass- h ough.
Finally, o ma ch Fac 3 one would ideally wan o inco po a e on- he-job sea ch, which is a -
ec ed p ima ily by in la ion expec a ions. Howe e , o simplici y pu poses, we abs ac om
o mally modeling ha channel in he p esen pape . Ins ead, we in oduce a wage-push ac o ,
ew
. The wage-push ac o a ec s he nominal wage only i he wo ke canno ba gain he wage
o W∗
+1and i cap u es he ollowing idea: in he case o no ba gaining, we assume ha , due o
highe in la ion expec a ions, he wo ke applies o ano he job wi h some p obabili y and is able
o gene a e an ou side con ac wi h ce ain y, which is used o pu upwa d p essu e on he nom-
inal wage wi h he cu en employe .25 The wage-push ac o is assumed o be pe sis en and o
be a ec ed by in la ion expec a ions as ollows
ˆ
ew
=ρwˆ
ew
−1+¯
eπE ˆ
π +1(12)
whe e ˆ
ew
is he wage-push ac o in log de ia ions om i s s eady-s a e alue; ¯
eπis he elas ici y
be ween in la ion expec a ions and he wage-push ac o ; and ρw∈[0,1)is he pe sis ence in he
wage-push ac o .
Fo wo ke s who ba gain in a gi en pe iod, he nominal wage is se acco ding o Nash ba -
gaining,
W∗
=a gmaxW (VE
− VU
)η (J )1−η (13)
whe e VE
and VU
deno e, espec i ely, he alue o employmen and unemploymen o a wo ke ;
J is he ma ke alue o a labo i m ma ched o a wo ke ; and η is he ime- a ying ba gaining
powe o wo ke s.26
6.2 Calib a ion
Ou calib a ion o he model aims o cap u e US labo ma ke ends a ound he ime o ou su ey
in ea ly 2022 while also ma ching ou h ee empi ical indings. In e ms o s eady-s a e alues, we
25The wage-push ac o plays a ole simila o ha ing wi hin-qua e job- o-job ansi ions wi h a ime- a ying
ansi ion p obabili y ha is only a ec ed by in la ion expec a ions. Wi hin-pe iod job- o-job ansi ions wi h cons an
p obabili y ha e been inco po a ed in K usell e al. (2017). Ano he in e p e a ion would be o ha e a non-ba gaining
wo ke ’s nominal wage indexed o a base, ixed eal wage g ow h ha is g ea e han 1, along wi h indexa ion o pas
in la ion. Time a ia ion in his case would only be induced by in la ion expec a ions.
26Unde EB, op imal nominal wages sa is y η J = (1−η )(VE
− VU
). In ou case o an RTM amewo k, he op imal
nominal wage condi ion is η δW
J = (1−η )δF
(VE
− VU
), whe e δW
and δF
deno e, espec i ely, he ne ma ginal
bene i s om an inc ease in he wage o he wo ke and he i m. See Ch is o el and Kues e (2008) o mo e de ails.
29
se he unemploymen and acancy a es o hei espec i e qua e ly ealiza ions in 2021:IV o 4.2
pe cen and 7 pe cen . The sepa a ion a e in he s eady s a e is se o 4.1 pe cen , ma ching he
qua e ly sepa a ion a e in 2021:IV. Table 5summa izes hese choices. Due o high labo ma ke
igh ness hese choices imply ha in he s eady s a e he p obabili y o inding a job is e y high (s
= 93.52 pe cen ), whe eas he likelihood ha a i m inds a wo ke is e y low (q = 0.27 pe cen ).
Table 5: Pa ame e s
Va iable Value Desc ip ion
u4.2 pe cen Unemploymen a e; US qua e ly unemploymen a e in 2021:IV
7 pe cen Vacancy a e; US qua e ly acancy a e in 2021:IV
µ4.1 pe cen Qua e ly sepa a ion a e; US da a in 2021:IV
s0.9352 P obabili y o inding a job (implied by he s eady-s a e model equilib ium)
q0.0027 P obabili y o inding a wo ke (implied by he s eady-s a e model equilib ium)
ξ0.6 Elas ici y o ma ches w. . . unemploymen ; see Pe ongolo and Pissa ides (2001)
η0.5 Ba gaining powe o wo ke s; con en ional alue
σm0.0037 E iciency o ma ching; econciles mwi h u=4.2 pe cen and =7 pe cen
ρw0.9 Pe sis ence o he wage-push ac o
¯
eπ0.0228 Wage-push elas ici y w. . . in la ion expec a ions ac oss all esponden s; Tables 2,4
¯
eπ0.114 Wage-push elas ici y w. . . in la ion expec a ions in coun e ac ual analysis; Table 4
γ0.875 Nominal wage s ickiness; pass- h ough ac oss all esponden s in Table 2
γ0.65 Nominal wage s ickiness; uni pass- h ough o coun e ac ual analysis
ζw0.675 Wage indexa ion; pass- h ough ac oss all esponden s in Table 2
ζw0.306 Wage indexa ion; pass- h ough o coun e ac ual analysis
λ0.285 In o ma ion s ickiness; Table 6
In e ms o labo ma ke pa ame e s, as shown in Table 5, we pa ame e ize he model as ol-
lows: he elas ici y o ma ches wi h espec o unemploymen , ζ, is se o 0.6, consis en wi h
Pe ongolo and Pissa ides (2001). Wage ba gaining powe is se o i s con en ional alue in he
li e a u e, i.e., η=0.5. The implied e iciency o ma ching, σm, is se o 0.0037 o be consis en wi h
he s eady-s a e alues o he unemploymen and acancy a es, and ma ching. We assume he
wage-push ac o p ocess is pe sis en wi h an au oco ela ion coe icien o 0.9.
A ew mo e pa ame e s emain o be calib a ed in a way ha is di ec ly ela ed o ou empi -
ical esul s. Fi s , o calib a e λ, we in es iga e how ou esponden s eac o new in o ma ion.27
27As shown by Coibion and Go odnichenko (2015), in a se ing wi h in o ma ion s ickiness simila o ou s, he
equency o upda ing he in o ma ion se (1−λ)is all one needs o pin down he esponse o expec a ions o new
in o ma ion a he ime o o ecas .
30
Speci ically, we ea ange equa ion (11) o ead as:
e
E π +h−e
E −1π +h
| {z }
(pos e io - p io )
= (1−λ)E π +h−e
E −1π +h
| {z }
new in o in pe iod
wi h (1−λ)cap u ing he e ec o new in o ma ion made a ailable in pe iod on in la ion expec-
a ions. To discipline λconsis en ly wi h ou expe imen , we use he es ima es om he ollowing
eg ession:
EihπPos e io
pi−EihπP io
pi=α+βTihIij −EihπP io
pii+εi(14)
whe e Tiis an indica o ha akes alue 1 i indi idual i ecei es ea men s 2, 4, o 5 (and possibly
3, depending on he speci ica ion), and akes a alue o ze o i he indi idual iis in he con ol o
placebo g oup. hIij −EihπP io
piicap u es new in o ma ion due o in o ma ion ea men j.Iij is
he nume ical in o ma ion con ained in ea men s 2, 3, 4, o 5. In his speci ica ion, β= (1−λ).
Table 6p esen s he es ima es o β. As ou benchma k calib a ion, we use he es ima e o β=0.715,
o equi alen ly, λ=0.285, as epo ed in column (4) o Table 6, whe e we accoun o he con ol,
placebo, and wage ea ed g oups.28
Table 6: E ec o New In o ma ion on In la ion Expec a ions
(1) (2) (3) (4)
New in o ma ion 0.742*** 0.711*** 0.742*** 0.715***
(0.014) (0.014) (0.012) (0.012)
Cons an 1.581*** -0.678*** 1.702*** -0.251
(0.163) (0.208) (0.139) (0.181)
Wage T ea men No No Yes Yes
Con ol and Placebo No Yes No Yes
Obse a ions 3,338 5,528 4,430 6,620
R-squa ed 0.730 0.432 0.735 0.483
No es: The able shows es ima es o equa ion (14). Column (1) only con ains
in o ma ion o ea men s 2, 4 and 5. Column (2) includes he placebo and
con ol g oups. Column (3) is (1) plus ea men 3 and column (4) con ains all
ea ed and con ol g oups. We use obus s anda d e o s.
Second, we calib a e nominal wage s ickiness, γ, and wage indexa ion o pas in la ion, ζw, o
28Coibion, Go odnichenko, and Webe (2022) a gue ha he inclusion o he con ol g oup is impo an since he
p io and pos e io ques ions abou in la ion expec a ions a e wo ded di e en ly. Ou esul s emain quali a i ely
simila i we calib a e λ o a lowe alue o abou 0.26.
31
ma ch Fac 1 and Fac 2 quan i a i ely along he IRFs o nominal wage g ow h o a ious shocks.
Sol ing he model unde a ional expec a ions, one can show unde gene al assump ions (see
de ails in Appendix G) ha he esponse o nominal wage g ow h expec a ions o a change in
in la ion expec a ions is gi en by:
∂e
E (ˆ
W +7−ˆ
W +3)
∂e
E ˆ
π +4
=a1−a2
1−λ+1+a3(15)
whe e he elemen s a1,a2, and a3a e con olu ed unc ions o he many s uc u al pa ame e s o
he model.29,30 Howe e , wage indexa ion o pas in la ion, and especially nominal wage s icki-
ness, γ, a e key pa ame e s in hese unc ions, and i is possible o calib a e hem such ha we a e
able o ma ch Fac 1 and Fac 2 quan i a i ely. In pa icula , we can ma ch he in la ion expec a-
ions pass- h ough o nominal wage g ow h ac oss ou esponden s by choosing a wage con ac
du a ion o abou 8 qua e s (γ=0.875) wi h indexa ion o pas in la ion o 0.675.31 To cons uc a
coun e ac ual scena io o uni pass- h ough om in la ion expec a ions o nominal wage g ow h
expec a ions, we se γ=0.65, which implies an a e age wage con ac du a ion o abou 3 qua -
e s. The wage indexa ion o pas in la ion in his case is se o ζw=0.306. No e ha while many
choices o ime ho izons exis o compu ing momen s, we choose he ime ho izons in equa ion
(15) o align wi h hose in he su ey.
Second, o ma ch Fac 3, we se he elas ici y o he wage-push ac o wi h espec o in la ion
expec a ions so ha we ma ch he e idence shown in Tables 2-4. Pa ame e ¯
eπis he elas ici y be-
ween in la ion and nominal wage g ow h expec a ions condi ional on ha ing applied o ano he
job due o highe in la ion expec a ions. Hence, we pa ame e ize ¯
eπas ollows:
¯
eπ=pass- h ough
| {z }
Tables 2,3
×elas ici y o job applica ions w. . . in la ion expec a ions
| {z }
=0.114, Table 4
(16)
6.3 Impulse Response Func ions: Lessons
Nex , we analyze he dynamics o ou model subjec o a posi i e demand shock and a posi i e
(ad e se) cos -push shock, he wo p edominan dis u bances ha we judge we e a ec ing he US
29While he e a e many pa ame e combina ions ha can ma ch he model-implied pass- h ough in (15) wi h he
empi ical one, we in e p e a less han uni pass- h ough as e idence o signi ican nominal wage igidi y and hus
emain ocused on calib a ing his pa ame e oge he wi h he wage indexa ion o pas in la ion.
30Recall ha ou pos e io ques ion abou income g ow h expec a ions in e s e
E (ˆ
W +7−ˆ
W +3).
31Du a ion o a wage con ac is gi en by 1/(1−γ).
32

economy a ound ou su ey pe iod. Two lessons eme ge ha help us unde s and he mechanism
behind households’ associa ion o highe in la ion wi h wo se economic ou comes, consis en wi h
ou empi ical indings and he wo k o Shille (1997) and Candia, Coibion, and Go odnichenko
(2020).
Lesson 1: Nega i e o dampened esponses o eal wages o shocks due o nominal wage igid-
i y ansla e in o g ea e luc ua ions and ola ili y in ou pu and consump ion.
Rega dless o whe he he model is subjec ed o a demand- o supply-side in la iona y dis u -
bance, an economy calib a ed o quan i a i ely ma ch ou empi ical pass- h ough o in la ion ex-
pec a ions o income g ow h expec a ions has la ge ami ica ions o eal wage dynamics ela i e
o a coun e ac ual scena io o a uni pass- h ough. As we subsequen ly explain, se e e nominal
wage igidi y is he d i ing sou ce o consume s’ dislike o in la ion in he model.
Figu e 2: Response o a Posi i e Demand Shock
No es: In black: calib a ion ma ching ou empi ical pass- h ough om in la ion o nominal wage g ow h expec a ions
(γ=0.875,ζw=0.675)acco ding o Equa ion (15). In dashed g ay: calib a ion ma ching coun e ac ual o uni
pass- h ough om in la ion o nominal wage g ow h expec a ions (γ=0.65,ζw=0.306). In ed: x axis.
33
Conside Figu e 2, whe e he economy is subjec o a one s anda d de ia ion posi i e demand
shock.32 Rela i e o he coun e ac ual o uni pass- h ough, eal wages decline, which esul s in
a la ge inc ease in labo hou s ha ampli ies he esponses o ou pu and consump ion.33 The
dynamics o eal wage and in la ion a e such ha he nominal wage g ow h, which is de ined as
he sum o eal wage g ow h and in la ion, inc eases in bo h cases. Consume s’ u ili y is a ec ed
by wo opposing o ces: i declines in esponse o wo king mo e along bo h he ex ensi e and
he in ensi e ma gins, bu i inc eases in esponse o highe consump ion.34 The o me channel
is conside ably la ge in he case o 20 pe cen pass- h ough compa ed wi h ull pass- h ough,
yielding a la ge decline in u ili y e en hough in la ion has isen by less.
32The s anda d de ia ion o he demand shock is se equal o 1.
33On impac , he eal wage is gi en by ˆ
w = (1−γ)ˆ
w∗
−γˆ
π , whe e ˆ
w∗
is he ully lexible eal wage. In con as o
he case o incomple e pass- h ough, unde uni pass- h ough, eal wages a e su icien ly lexible o espond posi i ely
o a posi i e demand shock.
34I is wo h no ing ha hou s in he model luc ua e in esponse o bo h he demand and he supply shocks ha
d i e in la ion up, while he su ey esponden s indica ed ha hey did no expec o change hei hou s in esponse o
highe in la ion, indica ing some ension be ween he heo e ical model and he empi ical da a. We lea e he esolu ion
o his conund um o u u e wo k.
34
Figu e 3: Response o a Posi i e Cos -Push Shock
No es: In black: calib a ion ma ching ou empi ical pass- h ough om in la ion o nominal wage g ow h expec a ions
(γ=0.875,ζw=0.675)acco ding o Equa ion (15). In dashed g ay: calib a ion ma ching coun e ac ual o uni
pass- h ough om in la ion o nominal wage g ow h expec a ions (γ=0.65,ζw=0.306). In ed: x axis.
Figu e 3conside s he case whe e he economy is shocked by a one s anda d de ia ion cos -
push supply dis u bance.35 Rela i e o he coun e ac ual o a uni pass- h ough economy, he
decline in eal wages is smalle , pu ing mo e downwa d p essu e on labo hou s. Since wages
a e mo e lexible in he coun e ac ual scena io o a uni pass- h ough, hey dec ease mo e and
as e compa ed o he incomple e pass- h ough case, esul ing in a decline in he nominal wage
g ow h. The la ge decline in hou s wo ked ansla es in o la ge declines in ou pu and consump-
ion. Unde a supply shock, g ea e nominal wage ic ions cause la ge inc eases in in la ion and
la ge dec eases in consump ion/ou pu , s eng hening consume s’ nega i e associa ion be ween
he wo. As was he case o a posi i e demand shock, a posi i e cos -push supply shock ini ially
causes an inc ease in u ili y, ollowed by a decline a ew pe iods la e , and hen a subsequen
inc ease as consume s ecei e highe u ili y om wo king less and enjoying mo e leisu e.36
35The s anda d de ia ion o he cos -push shock is se equal o 1.
36As wi h he demand shock, we no e ha he luc ua ions along he hou s ma gin un coun e o ou su ey esul s
in which esponden s belie e hey will no adjus hei hou s wo ked in esponse o a change in expec ed in la ion,
35
The compa a i e analysis pe aining o Figu es 2and 3is simila when he model is calib a ed
o ma ch he pass- h ough om in la ion expec a ions o income g ow h expec a ions associa ed
wi h high- e sus low-income esponden s. To a oid epe i ion, we epo hose IRFs in Appendix
I.
We nex show how he co ela ion be ween expec ed pe iod u ili y and in la ion expec a ions
a ies wi h he deg ee o nominal wage s ickiness and wage indexa ion o pas in la ion. A ep e-
sen a i e amily’s pe iod u ili y in de ia ion om i s s eady-s a e alue is gi en by:
U =(c(1−ϱ))1−σ(ˆ
c −ϱˆ
c −1)−κhnh1+φ
1+φˆ
n + (1+φ)ˆ
h (17)
whe e ˆ
c and ˆ
h deno e consump ion and labo hou s, espec i ely, in de ia ion om hei s eady-
s a e alues; ϱis he deg ee o ex e nal habi in consump ion; φis he in e se o labo supply
elas ici y; and κhis a scaling ac o o labo disu ili y.37
We simula e 50 pe iods o expec ed pe iod u ili y and in la ion expec a ions da a when shock-
ing he model wi h demand and cos -push inno a ions, o a gi en pai jo (γ,ζw), and conside
he ollowing eg ession o simula ed da a:38
E Uj, +1=αj+γ +βe
E ˆ
π +1+θγj×e
E ˆ
π +1+ϕζw,j×e
E ˆ
π +1+εj, (18)
whe e αjis an IRF ixed e ec , wi h an IRF being he se ies o expec ed pe iod u ili y and ex-
pec ed in la ion o a gi en combina ion o γand ζw; and γ +1is a ixed e ec o e e y pe iod a e
he shock. In he eg ession we d op he coe icien o each speci ic alue γand ζwas i will be
abso bed by he IRF ixed e ec . Table 7shows he esul s o a demand and a supply shock.
p o iding e ile g ound o explo e al e na i e models ha can cap u e his dimension o he da a.
37See Tables 21 and 22 o hei calib a ion.
38Fo each shock, we conside a o al o 10 ×11 =110 pai s o (γ,ζw), whe e γ∈ {0,0.1,...,0.9}and
ζw∈ {0,0.1,...,0.9,1}
36
Coibion, Oli ie , Yu iy Go odnichenko, Edwa d S. Kno ek II, and Raphael Schoenle. 2020a.
“A e age In la ion Ta ge ing and Household Expec a ions.” Tech. ep., Na ional Bu eau o
Economic Resea ch. DOI h ps://doi.o g/10.3386/w27836.
Coibion, Oli ie , Yu iy Go odnichenko, and Tiziano Ropele. 2020b. “In la ion expec a ions and
i m decisions: New causal e idence.” The Qua e ly Jou nal o Economics 135 (1):165–219. DOI
h ps://doi.o g/10.1093/qje/qjz029.
Coibion, Oli ie , Yu iy Go odnichenko, and Michael Webe . 2022. “Mone a y policy communica-
ions and hei e ec s on household in la ion expec a ions.” Jou nal o Poli ical Economy 130 (6).
DOI h ps://doi.o g/10.1086/718982.
Cu in, Richa d. 2022. “In la iona y Psychology Has Se In. Dislodging I Won’ Be Easy.” Ba on’s
DOI h ps://www.ba ons.com/a icles/in la ion-consume -p ices-sen imen -51649268928.
D’Acun o, F ancesco, Ul ike Malmendie , Juan Ospina, and Michael Webe . 2021. “Exposu e o
g oce y p ices and in la ion expec a ions.” Jou nal o Poli ical Economy 129 (5):1615–1639. DOI
h ps://doi.o g/10.1086/713192.
de Walque, G ego y, Oli ie Pie a d, Hen i Sneessens, and Ra Wou e s. 2009. “Sequen-
ial Ba gaining in a Neo-Keynesian Model wi h F ic ional Unemploymen and S ag-
ge ed Wage Nego ia ions.” Annals o Economics and S a is ics (95/96):223–250. DOI
h ps://doi.o g/10.2307/27917411.
Gali, Jo di, F ank Sme s, and Ra Wou e s. 2012. “Unemploymen in an es ima ed New Keynesian
model.” NBER Mac oeconomics Annual 26:329–360. DOI h ps://doi.o g/10.1086/663994.
Ge le , Ma k and An onella T iga i. 2009. “Unemploymen Fluc ua ions wi h S ag-
ge ed Nash Wage Ba gaining.” Jou nal o Poli ical Economy 117 (1):38–86. DOI
h ps://doi.o g/10.1086/597302.
Gue ei o, Joao, Jona hon Hazell, Chen Lian, and Ch is ina Pa e son. 2024. “Why Do Wo ke s
Dislike In la ion? Wage E osion and Con lic Cos s.” Tech. ep.
Hajdini, Ina, Edwa d S. Kno ek II, Ma hieu Pedemon e, Robe Rich, John Lee , and Raphael
Schoenle. 2022a. “Indi ec Consume In la ion Expec a ions.” Economic Commen a y (2022-03).
DOI h ps://doi.o g/10.26509/ bc-ec-202203.
43

Hajdini, Ina, Edwa d S Kno ek II, John Lee , Ma hieu Pedemon e, Robe Rich, and Raphael
Schoenle. 2024. “Indi ec consume in la ion expec a ions: Theo y and e idence.” Jou nal o
Mone a y Economics :103568.
Jäge , Simon, Ch is ophe Ro h, Nina Roussille, and Benjamin Schoe e . 2023. “Wo ke Belie s
Abou Ou side Op ions.” Wo king Pape 29623, Na ional Bu eau o Economic Resea ch. DOI
h ps://doi.o g/10.3386/w29623.
Jain, Monica, Olena Kos yshyna, and Xu Zhang. 2024. “How do people iew wage and p ice
in la ion?” Jou nal o Mone a y Economics :103552.
Jiang, Jane Hua, Rupal Kamda , Kelin Lu, and Daniela Puzzello. 2024. “How do households
espond o expec ed in la ion? an in es iga ion o ansmission mechanisms.” Tech. ep.,
Wo king pape .
Kamda , Rupal. 2019. “The Ina en i e Consume : Sen imen and Expec a ions.” 2019 Mee ing
Pape 647, Socie y o Economic Dynamics. DOI h ps://ideas. epec.o g/p/ ed/sed019/647.
h ml.
K usell, Pe , Toshihiko Mukoyama, Richa d Roge son, and Ay¸segül ¸Sahin. 2017. “G oss
Wo ke Flows o e he Business Cycle.” Ame ican Economic Re iew 107 (11):3447–3476. DOI
h ps://doi.o g/10.1257/ae .20121662.
Kuchle , The esa and Basi Za a . 2019. “Pe sonal expe iences and expec a ions abou agg ega e
ou comes.” The Jou nal o Finance 74 (5):2491–2542. DOI h ps://doi.o g/10.1111/jo i.12819.
Lo enzoni, Guido and I án We ning. 2023. “Wage P ice Spi als.” Tech. ep. DOI
h ps://bpb-us-w2.wpmucdn.com/ oices.uchicago.edu/dis /c/3483/ iles/2023/02/
WageP iceSpi als.pd .
Mankiw, N. G ego y and Rica do Reis. 2002. “S icky In o ma ion e sus S icky P ices: A P oposal
o Replace he New Keynesian Phillips Cu e.” Qua e ly Jou nal o Economics 117(4):1295–1328.
DOI h ps://doi.o g/10.1162/003355302320935034.
Mo ensen, Dale T. and Ch is ophe A. Pissa ides. 1994. “Job C ea ion and Job Des uc-
ion in he Theo y o Unemploymen .” Re iew o Economic S udies 61 (3):397–415. DOI
h ps://doi.o g/10.2307/2297896.
44
Pe ongolo, Ba ba a and Ch is ophe A. Pissa ides. 2001. “Looking in o he Black Box: A
Su ey o he Ma ching Func ion.” Jou nal o Economic Li e a u e 39 (2):390–431. DOI
h ps://doi.o g/10.1257/jel.39.2.390.
Pilossoph, Lau a and Jane M Ryngae . 2022. “Job Sea ch, wages, and in la ion.”
Sa ignac, F édé ique, E wan Gau ie , Yu iy Go odnichenko, and Oli ie Coibion. 2021. “Fi ms’
In la ion Expec a ions: New E idence om F ance.” Wo king Pape 29376, Na ional Bu eau o
Economic Resea ch. DOI h ps://doi.o g/10.3386/w29376.
Shille , Robe J. 1997. “Why do people dislike in la ion?” In Reducing in la ion: Mo i a ion and
s a egy. Uni e si y o Chicago P ess, 13–70.
Sme s, F ank and Ra Wou e s. 2007. “Shocks and F ic ions in US Business Cy-
cles: A Bayesian DSGE App oach.” Ame ican Economic Re iew 97 (3):586–606. DOI
h ps://doi.o g/10.1257/ae .97.3.586.
S an che a, S e anie. 2024. “Why do we dislike in la ion?” Tech. ep., Na ional Bu eau o
Economic Resea ch.
T iga i, An onella. 2006. “The Role o Sea ch F ic ions and Ba gaining o In la ion Dynamics.”
Wo king Pape 304, IGIER (Innocenzo Gaspa ini Ins i u e o Economic Resea ch), Bocconi
Uni e si y. DOI h ps://ideas. epec.o g/p/igi/igie p/304.h ml.
Webe , Michael, Be na do Candia, Tiziano Ropele, Rod igo Llube as, Se a in F ache, B en H.
Meye , Sa en Kuma , Yu iy Go odnichenko, Dimi is Geo ga akos, Oli ie Coibion e al. 2023.
“Tell me some hing I don’ al eady know: Lea ning in low and high-in la ion se ings.” Tech.
ep., Na ional Bu eau o Economic Resea ch. DOI h ps://doi.o g/10.3386/w31485.
We ning, I án. 2022. “Expec a ions and he Ra e o In la ion.” DOI h ps://economics.mi .edu/
si es/de aul / iles/publica ions/expec a ion-pass h ough.pd .
45
Appendix (Fo Online Publica ion)
A Su ey De ails and Ques ions
The expe imen was pu in o he ield by Mo ning Consul du ing he i s week o Ma ch 2022.
The goal was o sample a o al o 6,600 adul esponden s. The numbe o collec ed esponses was
6,629. The su ey s a s wi h demog aphic ques ions. These a e he ones we include in he pape :
• Wha is you i e-digi ZIP Code?
• Wha is you gende ?
–Male
–Female
• Wha is you age?
–18-34
–35-44
–45-64
–65+
• Which ca ego y ep esen s he o al combined income o all membe s o you HOUSEHOLD
du ing he pas 12 mon hs? This includes money om jobs, ne income om business, a m
o en , pensions, di idends, in e es , Social Secu i y paymen s and any o he money income
ecei ed by membe s o you amily who a e 15 yea s o age o olde .
–Unde 50k
–50k-100k
–100k+
Then, we ha e he p io ques ions o he expe imen :
• Nex we a e asking you o hink abou changes in p ices du ing he nex 12 mon hs in e-
la ion o you income. Gi en you expec a ions abou de elopmen s in p ices o goods and
se ices du ing he nex 12 mon hs, how would you income ha e o change o make you
equally well-o ela i e o you cu en si ua ion, such ha you can buy he same amoun
o goods and se ices as oday? (Fo example, i you conside p ices will all by 2% o e he
nex 12 mon hs, you may s ill be able o buy he same goods and se ices i you income also
dec eases by 2%.) To make me equally well o , my income would ha e o
46
–Inc ease by __%;
–S ay abou he same; and
–Dec ease by __%.
• Do you expec you income o inc ease, dec ease, o s ay abou he same o e he nex 12
mon hs?
–Inc ease by __%;
–S ay abou he same; and
–Dec ease by __%.
A his poin , esponden s we e andomly assigned o ecei e ei he a single ea men o o be
pa o he con ol g oup o esponden s (wi h he numbe o esponden s in pa en heses):
• Con ol (N=1,075)
• The Fede al Rese e a ge s an in la ion a e o 2% pe yea in he long un. (1,155)
• A ecen su ey om he Con e ence Boa d ound ha wages we e expec ed o ise 3.9% in
2022. (1,093)
• Be ween Janua y 2021 and Janua y 2022, he Consume P ice Index (CPI), which measu es
he a e age change in p ices o e ime ha consume s pay o goods and se ices, showed
he in la ion a e in he US was 7.5%. (1,112)
• Acco ding o he Su ey o P o essional Fo ecas e s, he Consume P ice Index (CPI), which
measu es he a e age change in p ices o e ime ha consume s pay o goods and se ices,
showed he in la ion a e will be 3.7% by he end o 2022. (1,074)
• Acco ding o he US Census Bu eau, he Uni ed S a es popula ion was 332,402,978 as o De-
cembe 31, 2021. (1,120)
A e being assigned o he con ol g oup o ecei ing a ea men , we asked e e ybody o hei
pos e io s in he ollowing ques ions:
• In he nex yea , do you hink p ices in gene al will inc ease, dec ease, o s ay abou he same?
–Inc ease by __%;
–S ay abou he same; and
–Dec ease by __%.
• Be ween Decembe 2022 and Decembe 2023, do you expec you income o inc ease, de-
c ease, o s ay abou he same o e he nex 12 mon hs?
47
–Inc ease by __%;
–S ay abou he same; and
–Dec ease by __%.
A e he pos e io s, indi iduals we e asked abou hei likely labo ma ke ac ions o inc ease
hei income o e he nex h ee mon hs.
• How likely a e you o do he ollowing o inc ease you income o e he nex h ee mon hs?
–Apply o a job(s) ha pays mo e
*Ve y likely
*Somewha likely
*Somewha unlikely
*Ve y unlikely
*Don’ know / No opinion
–Wo k longe hou s
*Ve y likely
*Somewha likely
*Somewha unlikely
*Ve y unlikely
*Don’ know / No opinion
–Ask o a aise
*Ve y likely
*Somewha likely
*Somewha unlikely
*Ve y unlikely
*Don’ know / No opinion
–O he (in his case, esponden s a e asked o p o ide a desc ip ion o labo ma ke ac-
ions)
48

B Addi ional Tables
Table 8: Summa y S a is ics and Rela ionship be ween P ice and Wage In la ion
Panel A Panel B
In la ion Exp Nominal Income Real Income Nominal Income
G ow h Exp G ow h Exp G ow h Exp
1s pe cen ile -2 -12 -100 In la ion Exp 0.365***
Fi s qua ile 0 0 -7 (0.012)
Median 0 0 0 Cons an 0.891***
Thi d qua ile 10 2 0 (0.104)
99 h pe cen ile 100 100 50
Mean 12.692 5.523 -7.169
S anda d de ia ion 24.536 18.822 22.735
Obse a ions 20,550 20,550 20,550 20,550
No es: This able shows summa y s a is ics o expec a ions o in la ion and nominal income g ow h. We also epo
a measu e o expec ed eal income g ow h de i ed as he di e ence be ween expec ed nominal income g ow h and
expec ed in la ion a he indi idual le el. The igh pa o he able shows a eg ession o expec ed nominal income
g ow h on expec ed in la ion. Hube - obus s anda d e o s a e in pa en heses. *** deno es s a is ical signi icance a he
1 pe cen le el.
49
Table 9: Robus ness Fi s S age Exe cise wi h T immed and Quan ile Reg essions
(1) (2) (3) (4)
EihπPos e io
piEihπPos e io
piEihπPos e io
yiEihπPos e io
yi
EihπP io
pi0.262*** 0.467***
(0.026) (0.016)
EihπP io
yi0.775*** 1.000
(0.048) -
T2: Ta ge -0.627 0.558 -0.203 -
(0.460) (0.248) (0.104) -
T3: Wages -0.695 1.333** -0.208 -
(0.450) (0.592) (0.230) -
T4: CPI -0.825* 0.533 -0.109 -
(0.456) (0.587) (0.254) -
T5: SPF -0.749 1.556*** -0.100 -
(0.465) (0.596) (0.247) -
T6: Placebo 0.133 1.333** -0.373 -
(0.465) (0.590) (0.248) -
T2 x p io -0.002 -0.079*** -0.127* -
(0.036) (0.022) (0.072) -
T3 x p io -0.003 -0.107*** -0.047 -
(0.035) (0.022) (0.071) -
T4 x p io -0.015 -0.107*** -0.114 -
(0.035) (0.022) (0.074) -
T5 x p io -0.025 -0.189*** -0.039 -
(0.036) (0.023) (0.071) -
T6 x p io 0.047 0.013 -0.078 -
(0.035) (0.022) (0.074) -
Cons an 5.667*** 0.667 0.925*** -
(0.337) (0.419) (0.185) -
Sample OLS Quan ile OLS Quan ile
Obse a ions 6,620 6,620 6,622 6,622
R-squa ed 0.261 0.559
No es: The able shows es ima es o equa ions 3and 4 ha gauge he e ec o ea men s
and hei in e ac ion wi h p io belie s. Columns (1) and (3) show esul s ha exclude
esponses in he ails o he dis ibu ion (less han he 5 h pe cen ile o g ea e han he
95 h pe cen ile) o changes be ween p io s and pos e io s, using obus s anda d e o s.
Columns (2) and (4) use quan ile eg essions a he median.
50
Table 10: E ec o In la ion Expec a ions on Wage Inc ease Ac ions, T immed Sample
Apply o a job(s) Wo k longe hou s Ask o a aise
(1) (2) (3) (4) (5) (6)
EihπPos e io
pi0.005*** 0.018*** 0.004** 0.008** -0.002 0.004
(0.002) (0.004) (0.002) (0.004) (0.002) (0.004)
Cons an 2.212*** 2.103*** 2.263*** 2.225*** 2.110*** 2.063***
(0.023) (0.039) (0.022) (0.039) (0.022) (0.041)
Reg ession OLS IV OLS IV OLS IV
F Tes 423.226 447.834 388.324
dy
dx
¯
x
¯
y0.019 0.067 0.014 0.031 -0.008 0.015
Obse a ions 4,471 4,471 4,406 4,406 4,256 4,256
R-squa ed 0.002 -0.013 0.001 -0.001 0.000 -0.003
No es: This able shows OLS and IV eg essions om equa ion 6.ℓj
iis a alue ha anges om
1 o 4, whe e 1 is “Ve y unlikely, ” 2 is “Somewha unlikely,” 3 is “Somewha likely” and 4 is
“Ve y likely.” Fo columns (1) and (2) ℓj
iis he answe o he ques ion abou “apply o a job(s)
ha pays mo e,” columns (3) and (4) a e he answe s o he ques ion abou “wo k longe hou s,”
and columns (5) and (6) a e he answe s abou “ask o a aise.” We use as an ins umen he
alues gene a ed om column (3) in Table 1Robus s anda d e o s a e in pa en heses.
Table 11: E ec o In la ion Expec a ions on Apply o a Job(s) by Demog aphics
Apply o a Job(s) Tha Pays Mo e
All Male Female <50k 50k-100k 100k+
(1) (2) (3) (4) (5) (6)
EihπPos e io
pi0.029*** 0.021*** 0.042*** 0.019** 0.048*** 0.025***
(0.006) (0.007) (0.010) (0.010) (0.011) (0.007)
Cons an 2.015*** 2.172*** 1.802*** 2.173*** 1.801*** 2.033***
(0.054) (0.060) (0.102) (0.095) (0.096) (0.074)
Reg ession IV IV IV IV IV IV
F-Tes 143.328 82.591 59.017 59.277 36.924 137.812
dy
dx
¯
x
¯
y0.114 0.072 0.184 0.076 0.182 0.094
Obse a ions 4,651 2,371 2,280 1,984 1,662 1,005
No es: This able shows IV eg essions om equa ion 6.ℓj
iis a alue ha anges om 1 o 4,
whe e 1 is “Ve y unlikely, ” 2 is “Somewha unlikely,” 3 is “Somewha likely” and 4 is “Ve y
likely.” ℓj
iis he answe o he ques ion “apply o a job(s) ha pays mo e.” Column (1) is o he
ull sample, column (2) only o male esponden s, column (3) o emale esponden s, column
(4) o esponden s who ha e an income lowe han 50k, column (5) o esponden s wi h income
be ween 50k and 100k, and column (6) o esponden s wi h income highe han 100k. We use
as an ins umen he alues gene a ed om column (3) in Table 1. Robus s anda d e o s a e in
pa en heses.
51
Table 12: E ec o In la ion Expec a ions on Wo k Longe Hou s by Demog aphics
Wo k Longe Hou s
All Male Female <50k 50k-100k 100k+
(1) (2) (3) (4) (5) (6)
EihπPos e io
pi0.009 0.004 0.018** 0.001 0.024** 0.012
(0.005) (0.007) (0.009) (0.009) (0.011) (0.008)
Cons an 2.219*** 2.372*** 2.008*** 2.263*** 2.067*** 2.296***
(0.051) (0.060) (0.091) (0.088) (0.093) (0.078)
Reg ession IV IV IV IV IV IV
F-Tes 149.752 88.642 60.033 61.735 39.939 138.630
dy
dx
¯
x
¯
y0.034 0.014 0.080 0.003 0.088 0.043
Obse a ions 4,573 2,339 2,234 1,942 1,630 1,001
No es: This able shows IV eg essions om equa ion 6.ℓj
iis a alue ha anges om 1 o 4,
whe e 1 is “Ve y unlikely, ” 2 is “Somewha unlikely,” 3 is “Somewha likely” and 4 is “Ve y
likely.” ℓj
iis he answe o he ques ion “wo k longe hou s.” Column (1) is o he ull sam-
ple, column (2) only o male esponden s, column (3) o emale esponden s, column (4) o
esponden s who ha e an income lowe han 50k, column (5) o esponden s wi h income be-
ween 50k and 100k, and column (6) o esponden s wi h income highe han 100k. We use as
an ins umen he alues gene a ed om column (3) in Table 1. Robus s anda d e o s a e in
pa en heses.
52
men s o ins umen o in la ion expec a ions. By con as , columns (5) o (8) show ha he in o -
ma ion ea men s do no seem o a ec consume s’ pos e io income g ow h expec a ions, condi-
ional on he p io , meaning ha he ea ed and con ol g oups a e e ec i ely he same, and p e-
en ing us om doing he same o ins umen o income g ow h expec a ions. As a esul , we un
EihπPos e io
pi=




∑j=2,4,5 γj
p×Tj
i +∑j=2,4,5 θj
p×Tj
i ×EihπP io
pii Ti =Ta ge ,CPI,SPF
0i Ti =Con ol,Placebo
whe e we use he nume ical in o ma ion p o ided wi hin each ea men Tj
i ha a ies o e
ime as abo e. Table 15 shows he esul s o he a e age and by demog aphics
Table 15: Pass- h ough om In la ion Expec a ions o Income G ow h Expec a ions, by Demo-
g aphics Follow-up
EihπPos e io
yi
All Male Female <50k 50k-100k >100k
EihπPos e io
pi0.174*** 0.243*** 0.135** 0.148*** 0.210** 0.253**
(0.043) (0.068) (0.056) (0.056) (0.087) (0.107)
EihπP io
yi0.594*** 0.597*** 0.582*** 0.597*** 0.567*** 0.603***
(0.019) (0.030) (0.026) (0.025) (0.037) (0.062)
Time FE Yes Yes Yes Yes Yes Yes
F- es 314.429 123.973 185.655 185.638 76.927 61.875
Obse a ions 12,882 6,039 6,843 6,029 4,452 2,401
R-squa ed 0.486 0.541 0.441 0.477 0.459 0.559
No es: This able shows esul s om IV eg essions om di e en demog aphics. The eg ession
used is he same as in column (2) in Table 2. Reg essions ha e obus s anda d e o s.
We see a pa e n simila o he one in he baseline exe cise. The es ima ed pass- h ough is a li -
le bi smalle , bu s ill close o 20 pe cen . We ind he same pa e n o he esul s by demog aph-
ics as be o e. Finally, we un he eg essions on he labo ma ke ac ions using he same s a egies,
meaning ha we use he same con ols and ime ixed e ec s. The esul s a e p esen ed in Table 16.
59

Table 16: E ec o In la ion Expec a ions on Wage Inc ease Ac ions, Follow-up
Apply o a job(s) Wo k longe hou s Ask o a aise
ha pays mo e
(1) (2) (3) (4) (5) (6)
EEihπPos e io
pi0.006*** 0.036*** 0.005*** 0.015*** -0.002 0.002
(0.001) (0.004) (0.001) (0.004) (0.001) (0.004)
Time FE Yes Yes Yes Yes Yes Yes
Reg ession OLS IV OLS IV OLS IV
F-Tes 372.1 377.8 359.9
dy
dx
¯
x
¯
y0.020 0.121 0.016 0.049 -0.007 0.007
Obse a ions 4,651 4,651 4,573 4,573 4,409 4,409
No es: This able shows OLS and IV eg essions om equa ion 6.ℓj
iis a alue ha anges
om 1 o 4, whe e 1 is “Ve y unlikely,” 2 is “Somewha unlikely,” 3 is “Somewha likely” and
4 is “Ve y likely.” Fo columns (1) and (2) ℓj
iis he answe o he ques ion abou “apply o a
job ha pays mo e,” columns (3) and (4) a e he answe s o he ques ion abou “wo k longe
hou s,” and columns (5) and (6) a e he answe s abou “ask o a aise.” Reg essions ha e
obus s anda d e o s.
We ind e y simila esul s in e ms o poin es ima es and elas ici ies. O e all, he ollow-up
exe cise con i ms he obus ness o he baseline esul s, sugges ing ha hey a e no d i en solely
by a pa icula ime pe iod in ea ly 2022. In addi ion, i is wo h no ing ha his exe cise om
Sep embe 2022 shows ha ou baseline esul s a e obus o a ying he p ecise ime ame used
in he p io s and pos e io s. In pa icula , in his exe cise we used a ime ame o he pos e io
income g ow h expec a ions ques ion ha had g ea e empo al o e lap wi h he p io han was
he case in ou baseline exe cise conduc ed in Ma ch 2022. Gi en ha ou esul s a e essen ially
unchanged, we a e com o able ha di e en iming assump ions we e no d i ing he esul s
documen ed in he body o he pape .39
In addi ion o his exe cise, we use he a ia ion on he same in o ma ion ea men o lea n
abou he e ec o each ea men on he pass- h ough esul . In o de o do so, we use he “con-
ol” g oups (placebo and con ol) and only one ea men g oup indi idually a a ime. Table 17
desc ibes he esul s o each ea men g oup.
39As a eminde , in he baseline su ey esul s om Ma ch 2022, he in la ion p io asked abou income needed o
o se p ice changes “o e he nex 12 mon hs,” while he in la ion pos e io asked abou he g ow h in p ices “in he
nex yea .” Meanwhile, he income g ow h p io asked abou expec ed income changes “o e he nex 12 mon hs”
while he income g ow h pos e io asked abou expec ed income g ow h “be ween Decembe 2022 and Decembe
2023.” In he su ey esul s om Sep embe 2022, he wo ding o he p io and pos e io ques ions was unchanged,
meaning ha he e was now mo e o e lap in he ime ames o he income p io and pos e io ques ions, whe eas
he e had been li le o e lap in he Ma ch wa e. The ac ha ou esul s a e essen ially he same implies ha he lack
o o e lap in he baseline esul s was no impo an o ou indings.
60
Table 17: IV Resul s o Each Indi idual T ea men
EihπPos e io
yi
(1) (2) (3) (4)
EihπPos e io
pi0.174*** 0.151* 0.148* 0.207**
(0.043) (0.078) (0.079) (0.090)
EihπP io
yi0.594*** 0.598*** 0.602*** 0.606***
(0.019) (0.028) (0.028) (0.030)
Time FE Yes Yes Yes Yes
T ea men All Ta ge CPI SPF
F-Tes 314.429 86.127 96.273 82.905
Obse a ions 12,882 7,792 7,735 7,673
R-squa ed 0.486 0.494 0.478 0.491
No es: This able shows esul s om IV eg essions one ea men a
a ime. The eg ession used is he same as in column (2) in Table 2.
Reg essions ha e obus s anda d e o s.
Table 17 shows ha he e ec changes sligh ly depending on he ea men . The es ima ed
pass- h ough is sligh ly s onge when consume s a e ea ed wi h in o ma ion abou u u e in la-
ion, and sligh ly lowe o he o he ea men s, bu hey a e all compa able. The able shows ha
ou main indings a e highly obus : pass- h ough is on he o de o oughly 20 pe cen . Because
each in la ion ea men is gene a ing a simila pass- h ough es ima e, we do no belie e ha he
imbalance o ha ing h ee in la ion ea men s and one wage ea men is a p ima y d i e o ou
main esul .
E Robus ness o Expe imen o P io on In la ion Expec a ions
He e, we show ha ou no el indi ec measu e o in la ion expec a ions, used o cap u e e-
sponden s’ p io in la ion expec a ions in he expe imen , does no bias he e ec o in la ion ex-
pec a ions on income g ow h expec a ions o labo ma ke ac ions. In Hajdini e al. (2022a), we
desc ibe ou no el measu e o in la ion expec a ions in de ail. In pa icula , we show ha i has
p ope ies simila o o he measu es o in la ion expec a ions such as hose o he Fede al Rese e
Bank o New Yo k’s Su ey o Consume Expec a ions (SCE) o he Su eys o Consume s by
he Uni e si y o Michigan. We use he ICIE as he main a iable on his pape because i s good
p ope ies and because i allows us o ob ain a la ge amoun o obse a ions o he expe imen ,
as i is pa o he main p oduc o he su ey. Rega dless o such e idence, we chose o pe o m
a complemen a y RCT expe imen in June 2023 o explo e whe he elying on ou no el indi ec
61
measu e o in la ion expec a ions biases he e ec o in la ion expec a ions on income g ow h ex-
pec a ions o labo ma ke ac ions. We ind ha he choice o he p io ques ion does no yield any
signi ican di e ences in ou main esul s.
Speci ically, a sample o a ound 4,400 esponden s en e ed ou RCT expe imen in June 2023.
Responden s we e andomly assigned o wo g oups: one g oup was asked ou no el ICIE ques-
ion and he o he g oup was asked he con en ional in la ion expec a ions ques ion om he
Fede al Rese e Bank o New Yo k’s Su ey o Consume Expec a ions. In pa icula , he la e
ques ion asks consume s he ollowing: “In he nex yea , do you hink ha he e will be in la ion o
de la ion? (No e: de la ion is he opposi e o in la ion).” Responden s we e hen p o ided wi h he
ollowing op ions: “1. In la ion (%); 2. De la ion (%); 3. Nei he in la ion no de la ion.” Then, all
esponden s we e asked he same ques ion abou income g ow h expec a ions, as in he egula
exe cise in he main ex : “Do you expec you income o inc ease, dec ease, o s ay abou he
same o e he nex 12 mon hs?” Subsequen ly, hal o each g oup ( andomly assigned) ecei ed a
ea men ela ed o in la ion:
“Acco ding o he Su ey o P o essional Fo ecas e s, he Consume P ice Index (CPI), which measu es
he a e age change in p ices o e ime ha consume s pay o goods and se ices, showed he in la ion a e
will be 3.4% by he end o 2023.”
The es o he esponden s ecei ed no ea men . Finally, all esponden s we e asked abou
hei pos e io in la ion expec a ions and income g ow h expec a ions, espec i ely, elying on he
ollowing wo ques ions:
“In he nex yea , do you hink p ices in gene al will inc ease, dec ease, o s ay abou he same?”
“Be ween Decembe 2023 and Decembe 2024, do you expec you income o inc ease, dec ease, o s ay
abou he same?” Las , we ask esponden s he labo ma ke ac ion ques ions in he same way as in
he main RCT expe imen .
The ul ima e goal o his exe cise is o unde s and whe he he es ima ed pass- h ough om
in la ion expec a ions o income g ow h expec a ions depends on he ques ion used o elici p io
in la ion expec a ions. Ou s a egy is o i s e alua e he e ec o he p io and ea men on
pos e io in la ion expec a ions, unning eg essions simila o (1) and (3). We do so o he wo
dis inc p io s sepa a ely as well as join ly, wi h esul s shown in Table 18.
62
Table 18: E ec s o T ea men s on Expec a ions: Di e en P io s
(1) (2) (3)
EihπPos e io
piEihπPos e io
piEihπPos e io
pi
EihπP io
pi0.491*** 0.218*** 0.399***
(0.003) (0.006) (0.063)
T1: SPF 0.580*** 0.130 0.239***
(0.064) (0.095) (0.005)
T1 x P io -0.446*** -0.057*** -0.192***
(0.005) (0.010) (0.005)
Cons an 0.164*** 0.830*** 0.645***
(0.035) (0.066) (0.041)
Sample ICIE NYFED Pooled
Obse a ions 1,813 1,974 3,846
R-squa ed 0.880 0.576 0.525
No es: The able shows es ima es o equa ion (1) ha ela e p io s and pos-
e io s, as well as es ima es o equa ion (3) ha gauge he e ec o ea -
men s and hei in e ac ion wi h p io belie s. In column (1), EihπP io
pi
e e s o p io in la ion expec a ions elici ed using he ICIE ques ion,
whe eas in column (2), EihπP io
pideno es p io in la ion expec a ions in-
e ed om he NY Fed ques ion. In column (3), bo h p io s a e pooled so
EihπP io
pideno es p io in la ion expec a ions in e ed om bo h he ICIE
and he NY Fed ques ion.
We hen ake ad an age o he exogenous a ia ion in in la ion expec a ions induced by ou in-
o ma ion ea men o cons uc ou ins umen o in la ion expec a ions, simila o he main RCT
expe imen . We cons uc he ins umen al a iable in wo ways: i) using he pooled i s -s age e-
g ession, he eby assuming he same coe icien o bo h p io s, and ii) allowing o p io -speci ic
coe icien s. Speci ically,
Ei
hπPos e io
pi=




γpTi+θpTi×EihπP io
pi i ea ed g oup
0 i con ol g oup
whe e Ti=1 i indi idual iis ea ed wi h he in la ion in o ma ion and 0 o he wise; o he i s
a ian o cons uc ing he ins umen al a iable we ely on es ima es o γpand θp epo ed in
column (3) in Table 18, whe eas o he second a ian we use es ima es o γpand θp epo ed in
column 1 o he esponden s who a e asked he ICIE ques ion and es ima es shown in column (2)
o hose who a e asked he Fede al Rese e Bank o New Yo k’s SCE ques ion.
63
We hen es ima e, analogously o ou p e ious ins umen ed eg ession se up, he ollowing
eg ession
EihπPos e io
yi=α0+α1×NYFed +β0EihπPos e io
pi+β1EihπPos e io
pi]×NYFed+ψEihπP io
yi+εi
(E.1)
whe e NYFed is a dummy a iable aking alue 1 i p io in la ion expec a ions a e elici ed using
he Fede al Rese e Bank o New Yo k’s SCE ques ion and 0 o he wise. We no e ha , di e en ly
om he analysis in he main ex , ou eg ession abo e includes he dummy a iable NYFed as
well as i s in e ac ion wi h he p io in o de o es whe he he e ec s o he choice o p io a e
signi ican ly di e en o no . We ins umen EihπPos e io
piusing Ei
hπPos e io
pi.
Simila ly, we un he ollowing eg ession o he epo ed likelihood o unde aking labo ma -
ke ac ion ℓj
ion expec ed in la ion, o assess he ex en o which in la ion expec a ions d i e labo
ma ke decisions:
ℓj
i=α0+α1×NYFed +β0EihπPos e io
pi+β1EihπPos e io
pi×NYFed+εi(E.2)
whe e ℓj
iis a alue ha anges om 1 o 4, whe e 1 is “Ve y unlikely, ” 2 is “Somewha unlikely,” 3
is “Somewha likely” and 4 is “Ve y likely” o h ee labo ma ke ac ions: i) apply o a job(s) ha
pays mo e, ii) wo k longe hou s, and iii) ask o a aise. As in (E.1), we con ol o he dummy
a iable NYFed and i s in e ac ion wi h he p io o es whe he he choice o p io has signi i-
can ly di e en e ec s on he es ima ed pass- h ough om in la ion expec a ions o labo ma ke
ac ions.
Table 19 shows he pass- h ough esul s and Table 20 shows he indings in e ms o labo
ma ke ac ions.
64

Table 19: Pass- h ough Es ima es o Di e en In la ion Expec a ions P io s
(1) (2) (3) (3)
EihπPos e io
yiEihπPos e io
yiEihπPos e io
yiEihπPos e io
yi
EihπPos e io
pi0.178*** 0.106 0.178*** 0.104
(0.039) (0.131) (0.039) (0.131)
EihπPos e io
pi×NYFed(= 1)-0.060 -0.120 -0.060 -0.119
(0.048) (0.141) (0.048) (0.141)
EihπP io
yi0.531*** 0.558*** 0.531*** 0.558***
(0.029) (0.034) (0.029) (0.034)
NYFed(= 1)-0.311 0.098 -0.311 0.092
(0.233) (0.684) (0.233) (0.681)
Cons an 0.488*** 0.753 0.488*** 0.761
(0.157) (0.574) (0.157) (0.571)
Sample Sepa a ed Sepa a ed Pooled Pooled
Reg ession OLS IV OLS IV
F-Tes 17.489 17.803
Obse a ions 4,405 4,405 4,405 4,405
R-squa ed 0.423 0.409 0.423 0.409
No es: This able shows esul s om OLS and IV eg essions in (E.1). Columns (1) and (2) a e he esul s o
eg essing he pos e io o income g ow h expec a ions on he p io o income g ow h expec a ions and he
pos e io o in la ion expec a ions using he IV cons uc ed sepa a ely o bo h p io s. In column (2) we use
IV, ins umen ing wi h Ei
πPos e io
p. Columns (3) and (4) a e he esul s o eg essing he pos e io o in la ion
expec a ions on he p io o in la ion expec a ions and he pos e io o income g ow h expec a ions using he
pooled es ima ion o he IV. In column (4) we use IV, ins umen ing wi h Ei
πPos e io
p.NYFed(= 1)is a a iable
ha akes a alue o 1 i he p io is he NY Fed ques ion. Robus s anda d e o s a e in pa en heses.
65
Table 20: E ec o In la ion Expec a ions on Labo Ma ke Ac ions
Apply o a job(s) Wo k longe hou s Ask o a aise
(1) (2) (3) (4) (5) (6)
EihπPos e io
pi0.049*** 0.049*** 0.005 0.005 -0.008 -0.008
(0.015) (0.015) (0.014) (0.014) (0.014) (0.014)
EihπPos e io
pix NYFed(= 1)-0.025 -0.025 0.012 0.012 0.003 0.003
(0.018) (0.018) (0.016) (0.016) (0.016) (0.016)
NYFed(= 1)0.049 0.049 -0.146 -0.146 -0.052 -0.053
(0.098) (0.098) (0.093) (0.093) (0.089) (0.089)
Cons an 1.688*** 1.689*** 1.949*** 1.949*** 1.770*** 1.770***
(0.076) (0.076) (0.072) (0.071) (0.072) (0.071)
Sample Sepa a ed Pooled Sepa a ed Pooled Sepa a ed Pooled
F- es 21.521 21.274 21.521 21.274 21.521 21.274
Obse a ions 4,405 4,405 4,405 4,405 4,405 4,405
No es: This able shows IV eg essions om equa ion (E.2). Columns (1) and (2) epo he es ima ed pass- h ough om
in la ion expec a ions o labo ma ke ac ion “apply o a job(s) ha pays mo e,” columns (3) and (4) epo he es ima ed
pass- h ough om in la ion expec a ions o labo ma ke ac ion “wo k longe hou s,” and columns (5) and (6) p o ide
he es ima ed pass- h ough om in la ion expec a ions o labo ma ke ac ion “ask o a aise.” NYFed(= 1)is a a iable
ha akes a alue o 1 i he p io is he NY Fed ques ion. Sample sepa a ed means ha he ins umen is buil sepa a ely
o each p io and pooled means ha i is buil join ly o bo h p io s, as explained in he ex . Robus s anda d e o s a e
in pa en heses.
The ollowing esul s a ise. Fi s , he choice o wo ding o he in la ion expec a ions ques-
ion ha o ms he p io —ICIE o based on he SCE—makes no s a is ically signi ican di e ence
in ou pass- h ough eg essions. The coe icien s on he NY Fed SCE dummy and he in e ac ed
p io wi h he NY Fed SCE dummy a e all s a is ically insigni ican . Second, he le els o he
pass- h ough es ima es a e somewha lowe han in ou main exe cise. This esul indica es ha
consume s may no be s ongly a ec ed by he wo ding o he ques ion, because in his pe iod,
independen ly o he p io , hey expec a low pass- h ough. Thi d, we also ind simila esul s in
e ms o labo ma ke ac ions, which con i ms he esul s o he main exe cise in he pape and e-
in o ces he main esul o he obus ness exe cise— o a di e en ou come a iable— ha esul s
a e independen o he choice o p io .
F Model
The model has been la gely adap ed om Ch is o el and Kues e (2008) and Ch is o el, Kues e ,
and Lize (2009).
66
Households. The e a e a la ge numbe o iden ical amilies wi h uni measu e. Each amily
consis s o a measu e n o employed membe s and u =1−n o unemployed membe s. Each
amily membe has he ollowing u ili y unc ion:
e
E0
∞
∑
=0
β (ci −ϱc −1)1−σ
1−σ−κh
h1+φ
i
1+φ!(F.1)
whe e ci deno es he consump ion o consume i;c −1is he amily’s agg ega e eal consump ion
in pe iod ( −1);hi is he wo king hou s o employed consume i;κh>0 is a pa ame e o wo k
disu ili y; and ϱ∈[0,1)cap u es he deg ee o ex e nal habi in consump ion. Each amily aces
he ollowing cons ain :
c +τ +κ =Z1−u
0wi hi di +u b+ed
d −1
R −1
π
−d +Ψ +n ΦK(F.2)
whe e e
Eis a gene ic expec a ions ope a o ; τ is lump-sum axes pe capi a in eal e ms; κ de-
no es eal cos pe acancy pos ing ;wi is he eal wage o employed consume i;d deno es he
isk- ee one-pe iod eal bond holdings wi h e u n ed
R and ed
being a shock o he isk p emium;
and bis eal unemploymen bene i s. Va iable Ψ deno es he eal di idends o he amily om
i ms in he economy, such ha Ψ =ΨC
+R1−u
0Ψh
i di, whe e ΨC
and Ψh
i a e di idends a ising
om he di e en ia ed goods and labo goods i ms, espec i ely, o be desc ibed in wha ollows.
The model does no accoun o capi al income, so we assume ha he amily ecei es a ixed
sha e n ΦK,ΦK≥0, ou o cu en e enue o labo i ms as “capi al income.” The amily makes
op imal decisions on behal o i s membe s by maximizing he agg ega e u ili y unc ion in (F.1)
wi h espec o consump ion and eal bond holdings, subjec o he budge cons ain in (F.2).
Fi ms. The e a e h ee ypes o i ms: i) i ms ha p oduce a homogeneous in e media e good,
“labo good”; ii) wholesale i ms ha pu chase labo goods in a pe ec ly compe i i e ma ke ,
and use hem as inpu s o p oduce di e en ia ed goods; and iii) e ail i ms ha pu chase di e -
en ia ed goods om he wholesale s and bundle hose goods in o a homogeneous consump ion
baske sold o consume s and he go e nmen .
Re aile s’ demand o di e en ia ed good jis gi en by:
yj =Pj
P −ε
y (F.3)
whe e Pj is he j h good p ice; ε>1 is he own-p ice elas ici y o demand; P is he agg ega e p ice
67
le el; and y deno es he inal good/economy’s agg ega e ou pu .
The wholesale sec o has a uni mass wi h i ms indexed by j∈[0,1]. Each i m p oduces a-
ie y jacco ding o yj =ld
j , whe e ld
j deno es i m j’s demand o he in e media e labo good,
which i can acqui e in a pe ec ly compe i i e ma ke a eal p ice xh
. Wholesale s ace Cal o-
ype p ice s ickiness such ha in e e y pe iod, a ac ion ω∈(0,1)o hem canno ese he p ice.
Simila o Ch is iano, Eichenbaum, and E ans (2005), we assume ha he i ms ha canno e-
op imize can adjus p ices by he index ac o πζp
−1¯
π1−ζp, whe e ζp∈[0,1]deno es he deg ee o
in la ion indexa ion. The p oblem o wholesale s is hen exp essed as ollows:
max
Pj e
E
∞
∑
h=0
ωhΓ , +h
Pj πζp
−1, −1+h(¯
π1−ζp)h
P +h
−mc +h
yj, +h
(F.4)
whe e Γ , +h=βhλ +h
λ , wi h λ being households’ ma ginal u ili y o consump ion; π −1, −1+h=
P −1+h/P −1; and mc =xh
eC
is he ma ginal cos , wi h eC
being a cos -push shock. To al p o i s o
he wholesale sec o in pe iod a e gi en by
ΨC
=Z1
j=0Pj
P
−mc yj dj (F.5)
Finally, he labo good i ms a e homogeneous and hey need exac ly one wo ke o ope a e.
So, he e is a mass o n = (1−u )o such i ms a any gi en ime. Ma ch ican p oduce li labo
good uni s ia li =z hα
i , whe e z is a p oduc i i y shock and α∈(0,1).
Labo ma ke s. The ma ching p ocess be ween wo ke s and labo i ms is go e ned by a
Cobb-Douglas unc ion,
m =σmuξ
1−ξ
(F.6)
whe e m is ma ches o med in pe iod ;u is unemploymen ; is acancies; ξ∈[0,1]is he elas-
ici y o ma ching wi h espec o unemploymen ; and σm>0 is a scaling ac o . Labo ma ke
igh ness is de ined as:
θ =
u
(F.7)
Then, he p obabili ies ha a acancy is illed and ha an unemployed wo ke ma ches wi h a
i m a e, espec i ely,
q =m
,s =m
u
(F.8)
68
Figu e 12: Response o a Posi i e Demand Shock
No es: In do ed ed: calib a ion ma ching ou empi ical pass- h ough om in la ion o nominal wage g ow h ex-
pec a ions o high-income consume s (γ=0.8515,ζw=0.35). In dashed blue: calib a ion ma ching ou empi ical
pass- h ough om in la ion o nominal wage g ow h expec a ions o low-income consume s (γ=0.895,ζw=0.6). In
black: x axis.
75

Figu e 13: Response o a Posi i e Cos -Push Shock
No es: In do ed ed: calib a ion ma ching ou empi ical pass- h ough om in la ion o nominal wage g ow h ex-
pec a ions o high-income consume s (γ=0.8515,ζw=0.35). In dashed blue: calib a ion ma ching ou empi ical
pass- h ough om in la ion o nominal wage g ow h expec a ions o low-income consume s (γ=0.895,ζw=0.6). In
black: x axis.
76