Supply chain networks and the macroeconomic expectations of firms

Hajdini, Ina; Kuma , Sa en; Malik, Sam een; No is, Jo dan J.; Pedemon e, Ma hieu
Wo king Pape
Supply chain ne wo ks and he mac oeconomic
expec a ions o i ms
IDB Wo king Pape Se ies, No. IDB-WP-1721
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; Kuma , Sa en; Malik, Sam een; No is, Jo dan J.; Pedemon e,
Ma hieu (2025) : Supply chain ne wo ks and he mac oeconomic expec a ions o i ms, IDB Wo king
Pape Se ies, No. IDB-WP-1721, In e -Ame ican De elopmen Bank (IDB), Washing on, DC,
h ps://doi.o g/10.18235/0013598
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/324838
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/
Supply Chain Ne wo ks and he
Mac oeconomic Expec a ions o Fi ms
Ina Hajdini
Sa en Kuma
Sam een Malik
Jo dan J. No is
Ma hieu Pedemon e
WORKING PAPER No IDB-WP-1721
In e -
A
me ican De elopmen Bank
Depa men o Resea ch and Chie Economis
July 2025
* Fede al Rese e Bank o Cle eland
** Auckland Uni e si y o Technology
*** NYU Abu Dhabi
**** In e -Ame ican De elopmen Bank
Supply Chain Ne wo ks and he
Mac oeconomic Expec a ions o Fi ms
Ina Hajdini*
Sa en Kuma **
Sam een Malik***
Jo dan J. No is***
Ma hieu Pedemon e****
In e -
A
me ican De elopmen Bank
Depa men o Resea ch and Chie Economis
July 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
Supply chain ne wo ks and he mac oeconomic expec a ions o i ms / Ina
Hajdini, Sa en Kuma , Sam een Malik, Jo dan J. No is, Ma hieu Pedemon e.
p. cm. — (IDB Wo king Pape Se ies; 1721)
Includes bibliog aphical e e ences.
1. Business logis ics-New Zealand. 2. Business en e p ises-New Zealand. 3.
Business communica ion-New Zealand. 4. Mac oeconomics. I. Hajdini, Ina. II.
Kuma , Sa en. III. Malik, Sam een. IV. No is, Jo dan J. V. Pedemon e,
Ma hieu. VI. In e -Ame ican De elopmen Bank. Depa men o Resea ch and
Chie Economis . VII. Se ies.
IDB-WP-1721
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 andomized con ol ial o app oxima ely 1,000 i m pai s in New Zealand
ha ha e a cus ome -supplie ela ionship, we p o ide an in o ma ion ea men o
analyze bo h he di ec e ec s on expec a ions and ac ions o i ms ecei ing his in o -
ma ion and he spillo e e ec s on connec ed i ms ha did no di ec ly ecei e in o ma-
ion. In a ollow-up h ee mon hs la e , we ind di ec and spillo e e ec s on expec a-
ions and ac ions ha a e bo h signi ican and o compa able magni ude. An inc ease
in expec ed GDP g ow h inc eases p ices and employmen ; an inc ease in expec ed
GDP unce ain y educes p ices, in es men , and employmen . We p o ide e idence
ha i is communica ion be ween he i ms, as opposed o obse able ac ions, d i ing
he spillo e e ec on he expec a ions o connec ed i ms. This is consequen ial as we
ind communica ion o be symme ic ups eam s downs eam, while p opaga ion ia
ac ions is asymme ic. We embed i m- o- i m communica ion along he supply chain
in a New Keynesian p icing p oblem and discuss i s implica ions o he ansmission
o agg ega e unce ain y o i ms’ p icing decisions and agg ega e in la ion.
JEL classi ica ions: D8, E3, E4, E5, L14
Keywo ds: Communica ion, Fi ms, Mac oeconomic expec a ions, Ne wo ks, Spillo e s
*This s udy has been app o ed by he Ao ea oa Resea ch E hics Commi ee, New Zealand (AREC 24 20),
Auckland Uni e si y o Technology E hics Commi ee (24/306) and New Yo k Uni e si y Abu Dhabi’s
E hics Boa d (HRPP-2024-110). We hank Oli ie Coibion, Yu iy Go odnichenko, Juan He e˜
no, Zhen Huo,
Hi oshi Toma, Michael Webe , semina pa icipan s a Uni e si ´
e La al, he Fede al Rese e Bank o Cle e-
land, and he IADB o hei help ul commen s and sugges ions. 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 IADB, he Fede al Rese e Bank o Cle eland
o he Fede al Rese e Sys em. Sam een Malik and Jo dan No is acknowledge he suppo by Tamkeen un-
de he NYU Abu Dhabi Resea ch Ins i u e Awa d o he Cen e o Beha io al Ins i u ional Design CG005.
(ADHPG-CG005). The usual disclaime applies. Emails: Ina.Hajdini@cle . b.o g, [email p o ec ed],
sam [email p o ec ed], [email p o ec ed], [email p o ec ed].

1 In oduc ion
Mac oeconomic expec a ions a e a key d i e o agg ega e dynamics (Beaud y and Po ie ,
2006,2007;Jaimo ich and Rebelo,2009). Empi ical esea ch o e he pas decade has
in ensi ied he analysis o unco e ing how hese expec a ions a e o med (Coibion and
Go odnichenko,2015;Coibion e al.,2018a), and wha hei e ec is on indi idual decision-
making (Coibion e al.,2020b;Geo ga akos e al.,2024). Special in e es has been placed
on i ms’ decision-making gi en hei p ice-se ing powe (Coibion e al.,2018b,2020a),
and he in luence o agg ega e unce ain y in shaping hese decisions (Bloom e al.,2007;
Bloom,2009;Coibion e al.,2024;Kuma e al.,2023).
The in e dependence o belie s be ween i ms is hough o play a key ole, o ex-
ample, being a he ounda ion o sen imen -d i en business cycles (Angele os and La’O,
2013;Gaballo,2018). Mo i a ed by supply chain in e ac ions being undamen al o shock
p opaga ion and ampli ica ion (Acemoglu e al.,2012,2016;Ca alho e al.,2021;Ozdagli
and Webe ,2023;Pas en e al.,2020), we in es iga e how a i m’s supply chain shapes i s
mac oeconomic expec a ions, and he consequences o i s decision-making. We p o ide expe -
imen al e idence on he p esence and ele ance o in o ma ion di usion be ween i ms,
and e eal di ec communica ion o be a cen al, ye unexplo ed, mechanism. Unlike s an-
da d shock p opaga ion ia p ices o ou pu being asymme ic ups eam s downs eam
(Acemoglu e al.,2016;Ca alho and Tahbaz-Salehi,2019), we ind communica ion o be
symme ic. This po en ially econ igu es ou unde s anding o how shocks p opaga e
h ough supply chains, and we embed communica ion in o a mac oeconomic model o
explo e i s agg ega e consequences.
To p oceed, we su eyed app oxima ely 1,000 i m- i m pai s in New Zealand, wi h
one i m being he p ima y supplie o he o he . We inco po a ed a andomized con-
olled ial (RCT) using an in o ma ion-based ea men o o icial o ecas s o GDP g ow h.
We had wo ea men g oups, one ecei ing he mean o he o ecas s and he o he ecei -
ing he ange ac oss o ecas s, in o de o assess he impac o unce ain y. Impo an ly,
in each ea ed pai , only one i m ecei es he in o ma ion (ei he he supplie o he cus-
ome in he ela ionship, chosen a andom). This design he e o e allows us o iden i y
1
bo h he di ec e ec o ecei ing mac oeconomic in o ma ion on he “main” i m ha he
ea men was applied o, and he spillo e e ec on he “connec ed” i m in he same pai
ha did no di ec ly ecei e he ea men . The su ey consis s o wo wa es (baseline
and ollow-up), and in o ma ion is p o ided a he end o he baseline. The ollow-up
akes place h ee mon hs la e , allowing us o iden i y he di usion o mac oeconomic in-
o ma ion be ween a i m and i s supplie o cus ome , and o measu e he e ec s on eal
decisions.
The p o ision o in o ma ion caused bo h he main and he connec ed i ms o upda e
hei expec a ions. The di ec e ec on he main i ms co obo a es indings in he li e -
a u e (Coibion e al.,2018b;Kuma e al.,2023).1The spillo e e ec s on he connec ed
i ms, howe e , a e new. As expec ed, he connec ed i ms’ expec a ions show a change
only in he ollow-up pe iod, no in he baseline pe iod, consis en wi h i aking ime o
he in o ma ion o di use. In e es ingly, he spillo e e ec s a e la ge, wi h hei magni-
ude being compa able o ha o he di ec e ec s, implying ha he in o ma ion di usion
is s ong.
Analyzing he impac o ou in o ma ion ea men on i ms’ decisions (ac ions)—
p ices, in es men , employmen , and wages—we ind signi ican e ec s in he ollow-up
o he di ec ly ea ed and connec ed i ms, bo h wi h simila magni udes. Using a ia ion
induced by he ea men o ins umen i ms’ GDP g ow h expec a ions, we ind ha a
1 pe cen age poin inc ease in expec ed GDP g ow h inc eases i ms’ p ices by 0.28 pe -
cen age poin s and employmen by 0.92 pe cen age poin s, compa ed o hei plan h ee
mon hs ago. We ind ha a 1 pe cen age poin inc ease in unce ain y, measu ed as he
dis ance be ween he mos and leas likely GDP g ow h scena io, dec eases p ices by 0.31
pe cen age poin s, in es men by 0.63 pe cen age poin s, and employmen by 0.79 pe -
cen age poin s, compa ed o hei plans h ee mon hs ago. We ind no e ec on wages
om ei he he mean o he unce ain y ea men s.
1Despi e he o icial o ecas s being public in o ma ion, he ina en ion o i ms wi h espec o his in o -
ma ion is well-es ablished, bo h in ou se ing (see Coibion e al.,2018b) and mo e gene ally (Candia e al.,
2024;Song and S e n,2024). Fo heo e ical mechanisms a ionalizing his ina en ion, see, o ins ance,
A ouzi and Yang (2018); Gabaix (2020); Sims (2003).
2
The signi ican shi in expec a ions and decisions o connec ed i ms sugges s ha
he supply chain ne wo k is a highly ele an sou ce o in o ma ion o mac oeconomic
expec a ions and decisions, wi h meaning ul in e ac ions and in o ma ion spillo e s be-
ween connec ed i ms. Nex , we assess he po en ial mechanisms unde lying his in o -
ma ion di usion. Speci ically, we wan o disen angle whe he communica ion be ween
i ms, o in e ence om changes in obse able ac ions o he o he i m, can explain he
lea ning we ind. We o e a numbe o pieces o e idence sugges ing ha communica ion
is impo an .
Fi s , we decompose he spillo e e ec on a connec ed i m’s pos e io belie s in o
he impac coming om he main i m’s pos e io belie s as opposed o he main i m’s
ac ions. We ind only he o me o be signi ican , sugges ing ha connec ed i ms o m
hei belie s by di ec ly lea ning he main i m’s belie s, concei ably h ough communica-
ion. Second, in he su ey, we asked he i ms di ec ly abou hei GDP communica ion
wi hin he pai . We ind a la ge and s a is ically signi ican e ec : 85% (73%) o mean (un-
ce ain y) ea ed i ms epo ed communica ing, compa ed o only 35% o con ol i ms.
Thi d, we show ha he spillo e e ec s a e symme ic whe he lowing ups eam o
downs eam ( he main i m is a cus ome o supplie in he pai , espec i ely). I lea ning
we e media ed exclusi ely h ough obse ing he ac ions o he main i ms, one would
expec asymme ic ea men e ec s, as he li e a u e has documen ed ha shocks end o
p opaga e ia p ices o ou pu mo e in one di ec ion (Acemoglu e al.,2016;Ca alho and
Tahbaz-Salehi,2019). Con e sely, using ou su ey ques ion on GDP communica ion, we
ind ha he e ec on communica ion om ea men is he same whe he he main i m
is he cus ome o supplie in he pai , a inding ha is consis en wi h he ea men e ec
on expec a ions and ac ions being symme ic.
We end he pape by in es iga ing he implica ions o ou indings in a New Keynesian
p icing p oblem, whe e agg ega e ou pu (GDP) g ow h is exogenously gi en. Building
on he p icing block o he p oduc ion ne wo k model in Rubbo (2023), we inco po a e
a communica ion ne wo k ha i ms u ilize when o ming hei expec a ions abou ou -
pu g ow h, he la e being impe ec ly obse ed by i ms. Mo i a ed by ou empi ical
3
indings, we assume ha he communica ion ne wo k is symme ic (equal communica-
ion ups eam and downs eam) and ha i ms a e ambigui y-a e se, as a ac able way
o unce ain y o be ele an o decisions despi e he ac ha we log-linea ize he model
(Ilu and Schneide ,2014).2Ou se up allows us o show ha , in equilib ium, i ms’ p ic-
ing decisions a e in luenced by bo h he p oduc ion and he communica ion ne wo ks.
Consis en wi h ou empi ical e idence, he model implies a nega i e esponse o i ms’
p ices o a ea men o highe unce ain y abou u u e ou pu g ow h.
We cha ac e ize he implica ions o communica ion using he model bo h heo e ically
and quan i a i ely. Fo he la e , we pa ame e ize he model o closely ma ch key compo-
nen s o i ms in ou su ey da a and simula e ou comes when a subse o i ms is gi en
in o ma ion abou highe unce ain y abou u u e ou pu g ow h. Ou analysis yields
h ee key insigh s. Fi s , whe he he ea men was p o ided o he supplie o he cus-
ome i m does no ma e o i s impac on all i ms’ p ices when i ms communica e,
highligh ing ha communica ion gene a es symme y in ups eam s downs eam ans-
mission o a ea men .
Second, communica ion educes he dispe sion o p ice esponse o a ea men ac oss
i ms when compa ed o no communica ion. We empi ically alida e his esul o he
model by compu ing he connec ed i m’s p ice change associa ed wi h a 1 pe cen age
poin exogenous inc ease in he ea ed i m’s p ice in ou model simula ions and su -
ey da a. We show ha when he e is communica ion, he es ima ed p ice ela ionship
be ween he ea ed and connec ed i m app oaches uni y bo h empi ically and in he
model, ega dless o whe he he ea men was p o ided o a cus ome o supplie i m.
Thi d, communica ion esul s in a s onge and sho e -li ed esponse o in la ion
o u u e ou pu g ow h unce ain y, consis en wi h communica ion bo h p opaga ing
and homogenizing he ini ial p ice esponse o mos i ms, ela i e o no communica ion.
Mo eo e , he esponse o he agg ega e p ice le el o u u e ou pu g ow h unce ain y
is, on a e age, highe when i ms communica e wi h one ano he compa ed o when hey
2Ambigui y-a e sion e e s o Knigh ian unce ain y whe eby i ms canno assess he p obabili y dis i-
bu ion o ou comes accu a ely. See Eps ein and Wang (1994) o an ea ly applica ion o such unce ain y o
asse p icing and Ilu and Schneide (2023) o a ecen e iew.
4
age. Addi ionally, while all 1,074 pai s o i ms we e con ac ed o he ollow-up, only
539 pai s o i ms pa icipa ed, because ei he one o bo h i ms in he pai s did no an-
swe he su ey. Appendix Table A-3 shows ha nei he he ea men assignmen no he
obse able cha ac e is ics can p edic pa icipa ion in he ollow-up su ey.
3 T ea men E ec s
3.1 T ea men E ec s on Expec a ions
We s a by e alua ing whe he ou ea men s a ec ed i ms’ GDP expec a ions. To do
so, we compa e how ea ed and con ol i ms changed hei pos e io GDP expec a ions
ela i e o hei p io GDP expec a ions. Speci ically, we un he ollowing eg ession,
which is widely used in his ype o se ing (see, o example, Coibion e al.,2018b;Kuma
e al.,2023)7:
Pos e io mean
i=α+βP io mean
i+
2
X
n=1
γnTn,i +
2
X
n=1
θnP io mean
i×Tn,i +εi,(1)
whe e P io mean
iis he belie o i m i’s manage abou he mean o GDP g ow h in he
baseline pe iod be o e he ea men in e en ion. Pos e io mean
iis he belie a e he
ea men in e en ion. We use wo measu es o pos e io belie s. The i s is an ins an-
aneous measu e, asked immedia ely a e he ea men in e en ion in he baseline, and
he second is a pe sis en measu e, asked in he ollow-up h ee mon hs a e he baseline.
Tn,i is a dummy ha akes a alue o 1 i i m iis in a pai ha ecei ed ea men n(n= 1
is mean, n= 2 is unce ain y) and 0 o he wise.
We un he eg ession sepa a ely o he main i ms in o de o es ima e he di ec
e ec , and sepa a ely o he connec ed i ms in o de o es ima e he spillo e e ec s
7Some o hese pape s also employ Hube - obus eg essions, which inc ease powe by down-weigh ing
obse a ions wi h la ge esiduals, ypically hose wi h subs an ial p io -pos e io e isions. We epo ou
esul s using his app oach in Appendix C-3. The indings emain quali a i ely unchanged and, i any hing,
become quan i a i ely s onge .
11

(co esponding o Figu e 1). No e ha he ea men s a us Tn,i is he same o bo h main
and connec ed i ms wi hin he same pai ; ha is, a connec ed i m is ea ed i he main
i m i is pai ed wi h is also ea ed. We also e un he eg ession o he pos e io and
p io on unce ain y, a he han he mean.
The coe icien βcap u es he co ela ion be ween p io and pos e io o he con ol
g oup. As he con ol ecei ed no in o ma ion, we expec ha βis close o one. β+θn
cap u es he co ela ion be ween p io and pos e io o ea ed g oup n. I ea men n
is e ec i e, we will see changes in expec a ions such ha ea ed i ms place some posi-
i e weigh on he new in o ma ion. Consequen ly, θnwill be nega i e, as he co ela ion
be ween he p io and pos e io would be lowe han in he con ol g oup. Because he
ea men is andomized, we can in e p e θnas he causal e ec o in o ma ion on he
p io -pos e io co ela ion. γnis he he causal e ec when p io expec a ions equal ze o
( he y-in e cep ), which we expec o be posi i e i he co ela ion ( he slope) dec eases. Vi-
sually, he ela ionship be ween he pos e io and p io o a es clockwise due o ea men
(see Figu e 2).
We p esen he esul s on he belie s o mean GDP g ow h in Table 1. Columns (1) and
(2) show he ea men e ec s in he baseline pe iod o he main and connec ed i ms,
espec i ely. Columns (3) and (4) show he ea men e ec s in he ollow-up pe iod o
he main and connec ed i ms, espec i ely. Figu e 2p esen s he co esponding dis i-
bu ions o pos e io agains p io belie s in he ou cases. As expec ed, he es ima ed
co ela ion be ween he p io and pos e io o he con ol g oup, β, is close o one ac oss
all ou speci ica ions in Table 1, and he dis ibu ion along he 45-deg ee line in Figu e 2,
indica ing no sys ema ic change in he con ol i ms’ belie s be o e o a e ea men .
The ea men e ec on he main i m — he di ec e ec — in he baseline pe iod
(Column 1) om ea men one (p o ision o he mean o icial o ecas o GDP g ow h) is
θ1=−0.723, a educ ion in co ela ion o app oxima ely h ee-qua e s ela i e o he con-
ol i ms. Fi ms di ec ly ecei ing he in o ma ion, he e o e, immedia ely upda e hei
p io s. T ea men wo (p o ision o he ange o o icial GDP g ow h o ecas s) also leads
o a signi ican educ ion in co ela ion, hough o smalle magni ude. This is expec ed
12
Table 1: T ea men E ec on GDP Expec a ions in Baseline and
Follow-up
(1) (2) (3) (4)
P io mean 0.972*** 0.964*** 0.945*** 0.938***
(0.023) (0.016) (0.020) (0.013)
T11.799*** -0.063 1.787*** 1.772***
(0.068) (0.044) (0.070) (0.112)
T21.567*** -0.040 1.773*** 1.433***
(0.068) (0.045) (0.095) (0.147)
T1×P io mean -0.723*** 0.017 -0.603*** -0.586***
(0.032) (0.019) (0.032) (0.046)
T2×P io mean -0.492*** 0.006 -0.503*** -0.502***
(0.032) (0.018) (0.046) (0.061)
Cons an 0.025 0.062 0.080 0.120**
(0.048) (0.043) (0.047) (0.036)
Pe iod Pos e io Baseline Baseline Follow-Up Follow-Up
Type o i m Main Connec ed Main Connec ed
Obse a ions 999 1020 510 505
R-squa ed 0.739 0.955 0.760 0.743
No e: The able epo s esul s o eg ession 1, whe e he ou come a iable
Pos e io mean is he a e age GDP o ecas o i m ia e he ea men . P io mean
is he a e age GDP o ecas be o e he ea men . T1is an indica o ha is equal
o one i i m i ecei ed he in o ma ion ea men abou he a e age GDP o ecas
and T2is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea -
men abou he GDP unce ain y. Columns (1) and (2) show esul s o he baseline
su ey, and columns (3) and (4) show esul s o he ollow-up su ey. Columns
(1) and (3) show esul s o he i ms ha ecei ed he in o ma ion ea men in he
baseline pe iod, and columns (2) and (4) show esul s o he i ms ha a e con-
nec ed o he ea ed i ms. Robus s anda d e o s a e shown in pa en heses.
gi en ha in o ma ion abou unce ain y in GDP g ow h is no di ec ly in o ma i e abou
he mean g ow h. In Panel A o Figu e 2, we see a co esponding clockwise o a ion o he
ela ionship be ween pos e io and p io , e lec ing he educ ion in he co ela ion due o
he ea men .
The ea men e ec on he connec ed i m — he spillo e e ec — in he baseline
pe iod (Column 2) om ei he ea men is insigni ican ly di e en om ze o. Co e-
spondingly, we see no o a ion o he dis ibu ion in Panel B o Figu e 2. As expec ed, his
sugges s no in o ma ion has ye been di used om he main o he connec ed i m in he
13
Figu e 2: Co ela ion be ween P io and Pos e io o Main and Connec ed Fi ms in he
Baseline and Follow-up
Baseline
A: Main Fi m B: Connec ed Fi m
Follow-up
C: Main Fi m D: Connec ed Fi m
No e: This igu e shows a sca e plo o he expec a ions abou GDP asked be o e he ea men in he
baseline pe iod (p io , x-axis) wi h ei he he pos e io in he baseline pe iod o he pos e io in he ollow-
up pe iod (y-axis). Panels A and B plo he p io and he pos e io in he baseline pe iod. Panel A p esen s
esul s o ea ed i ms, while Panel B shows he same o connec ed i ms. Panels C and D plo p io
expec a ions in he baseline pe iod agains pos e io expec a ions in he ollow-up—Panel C o ea ed
i ms, Panel D o connec ed i ms. Each do ep esen s a i m’s esponse; he lines a e linea i s by g oup.
Black indica es con ol i ms, g ay co esponds o hose ecei ing T ea men 1 (a e age GDP o ecas ), and
blue o hose ecei ing T ea men 2 (unce ain y in o ma ion).
pai . This is because bo h i ms wi hin a pai we e su eyed e y close in ime, while i
concei ably akes ime o in o ma ion o be di used be ween i ms.
Now, u ning o he ea men e ec in he ollow-up pe iod, he di ec e ec in he
ollow-up (Column 3) is e y simila o he e ec in he baseline (Column 1). This sug-
14
Table 2: T ea men E ec on Expec ed GDP Unce ain y in Baseline and Follow-up
(1) (2) (3) (4)
P io Unce ain y 0.960*** 0.993*** 0.978*** 0.974***
(0.019) (0.010) (0.019) (0.018)
T11.395*** 0.025 1.310*** 2.044***
(0.198) (0.084) (0.302) (0.328)
T21.145*** -0.015 1.142*** 1.139***
(0.163) (0.083) (0.264) (0.267)
T1×P io Unce ain y -0.766*** -0.008 -0.717*** -0.761***
(0.033) (0.013) (0.042) (0.046)
T2×P io Unce ain y -0.720*** -0.008 -0.689*** -0.610***
(0.031) (0.014) (0.042) (0.045)
Cons an 0.220* 0.067 0.187* 0.276*
(0.095) (0.070) (0.090) (0.122)
Pos e io Pe iod Baseline Baseline Follow-Up Follow-Up
Fi m Type Main Connec ed Main Connec ed
Obse a ions 1012 1022 514 513
R-squa ed 0.835 0.973 0.809 0.700
No e: The able epo s esul s o eg ession 1, whe e he ou come a iables
Pos e io unce ain y is he unce ain y in he GDP o ecas o i m ia e he ea -
men , measu ed as he absolu e alue on he dis ance be ween he mos and less
likely scena io. P io unce ain y is he unce ain y o ecas be o e he ea men . T1
is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou
he a e age GDP o ecas and T2is an indica o ha is equal o one i i m i ecei ed
he in o ma ion ea men abou he GDP unce ain y. Columns (1) and (2) show e-
sul s o he baseline su ey, and columns (3) and (4) show esul s o he ollow-up
su ey. Columns (1) and (3) show esul s o he i ms ha ecei ed he in o ma ion
ea men in he baseline pe iod, and columns (2) and (4) show esul s o he i ms
ha a e connec ed o he ea ed i ms. Robus s anda d e o s a e shown in pa en-
heses.
ges s ha he ea men e ec is highly pe sis en , wi h he belie s o he main i m in he
ollow-up con inuing o be highly in luenced by he ea men h ee mon hs ea lie . Mo e
impo an ly, he spillo e e ec in he ollow-up (Column 4) is now highly signi ican .
Tha is, e en hough he connec ed i ms did no di ec ly ecei e he in o ma ion in he
baseline, hei belie s h ee mon hs la e had been upda ed as i hey had ecei ed he
in o ma ion. Mo eo e , he magni udes o spillo e e ec s a e ema kably simila o he
di ec e ec s (Column 4 s Column 3). Panels C and D o Figu e 2p esen he co espond-
ing dis ibu ions, showcasing he simila i y in hei e ec .
15
These indings on he spillo e e ec s a e he mos in e es ing and no el pa o his
pape . This implies ha he in o ma ion abou GDP expec a ions has been di used om
he main i m o he connec ed i m — i.e., along he supply chain ne wo k. In Sec ion 4,
we explo e he mechanism o his di usion, speci ically whe he he i ms a e engaging in
di ec communica ion abou GDP expec a ions, o whe he hey a e in e ing hem om
obse able ac ions.
We p esen he analogous esul s o p io s and pos e io s on he unce ain y o GDP
g ow h, a he han he mean, in Table 2. We ind e y simila esul s, bo h quali a i ely
and quan i a i ely. This shows ha no only is in o ma ion abou he mean ansmi ed
h ough he inpu -ou pu ne wo k bu also in o ma ion abou unce ain y.
We examine he he e ogenei y o he ea men e ec s wi h espec o i m cha ac e is-
ics (size, ma ke sha e, age, and sec o ) in Table A-4 (a). We de ec no sys ema ic a ia ion
ac oss hese dimensions. This sugges s ha he in o ma ion di uses b oadly, a he han
being limi ed o a speci ic ype o i m.
3.2 T ea men E ec s on Ac ions
In his sec ion, we e alua e whe he i ms changed hei decisions/ac ions due o he
in o ma ion ea men , sugges ing ha he in o ma ion con en is economically ele an
and meaning ul. We examine ou measu es o decisions: p ice, in es men , employmen ,
and wages. These a e measu ed bo h as planned changes epo ed in he baseline su ey
(ex-an e plans o he nex h ee mon hs) and as ac ual ac ions eco ded in he ollow-up
su ey (ex-pos decisions a he endline). Fi s , we es ima e he educed- o m e ec o he
ea men on he ac ions, e ealing whe he he in o ma ion caused i ms’ ac ions o be
less co ela ed wi h hei ini ial plans. Second, we use he ea men as an ins umen o
he i ms’ GDP expec a ions o es ima e he elas ici y o a change in ac ions wi h espec
o a change in GDP expec a ions.
The educed- o m eg ession is he ollowing:
16

Ac ioni=α+βPlani+
2
X
n=1
γnTn,i +
2
X
n=1
θnPlani×Tn,i +εi,(2)
whe e Ac ioniis he ac ion he manage o i m i epo ed in he ollow-up pe iod. Plani
is he i m’s plan epo ed in he baseline pe iod. As in eg ession 1,Tn,i is a dummy ha
akes a alue o one i i m i ecei ed he ea men nand ze o o he wise. θn e lec s he
co ela ion o he ac ion and he plan. I he ea men has an e ec on he i m’s ac ion,
hen he plan-ac ion co ela ion will be educed, co esponding o a nega i e θn. As be o e,
he speci ica ion is un sepa a ely o main and connec ed i ms o es ima e he di ec and
spillo e e ec s, espec i ely. We p esen he esul s in Table 3.
We ind signi ican ea men e ec s on p ices, employmen , and in es men (Columns
1 o 6), hough no on wages (Columns 7 o 8).8Mos in e es ingly, his is he case no
only o he di ec e ec s (odd numbe ed columns) bu also o he spillo e e ec s (e en
numbe ed columns). Mo eo e , he magni udes o he di ec and spillo e e ec s o an
ac ion a e simila , especially o p ices and in es men s. This sugges s ha he di usion
o in o ma ion be ween i ms is highly ele an and meaning ul o i m ope a ions.
We examine he e ogenei y o he ea men e ec s wi h espec o i m cha ac e is ics
in Table A-4 (b). We de ec no sys ema ic a ia ion ac oss hese dimensions.
Nex , we es ima e he elas ici ies o ac ions wi h espec o expec a ions. This is use ul
o gi e a clea e sense o he magni ude o he changes in ac ions due o he ea men , as
i accoun s o he changes in expec a ions a ibu able o he in o ma ion ea men . We
ex end he ins umen al a iable s a egy o di ec e ec s om Coibion e al. (2022,2023)
and Kuma e al. (2023) o spillo e e ec s. Fo he di ec e ec s, he second s age is gi en
by:
Ac ioni=α+βPlani+γPos e io mean
i+θPos e io unce ain y
i+X′
iδ+εi,(3)
The eg ession is un on he main i ms i.Xiincludes p io s o mean and unce ain y
om he baseline pe iod. The es o he a iables a e de ined as in speci ica ions 1and
2. We ins umen Pos e io mean
iand Pos e io unce ain y
iby he ea men in e ac ed wi h
8No di ec e ec on wages is consis en wi h esul s in he li e a u e in ou se ing (Kuma e al.,2023).
17
Table 3: T ea men E ec on Wage, Employmen , and In es men Plans
P ice In es men Employmen Wage
(1) (2) (3) (4) (5) (6) (7) (8)
Plan 1.006*** 1.012*** 0.975*** 0.979*** 1.014*** 1.017*** 0.995*** 0.998***
(0.009) (0.011) (0.018) (0.019) (0.020) (0.012) (0.015) (0.019)
T11.583*** 1.841*** 3.448*** 3.128*** 2.837*** 2.291*** -0.024 0.011
(0.136) (0.136) (0.199) (0.205) (0.540) (0.498) (0.019) (0.041)
T21.722*** 1.815*** 2.819*** 2.552*** 3.388*** 2.883*** -0.016 -0.028
(0.125) (0.125) (0.190) (0.167) (0.568) (0.472) (0.016) (0.028)
T1×Plan -0.323*** -0.401*** -0.679*** -0.625*** -0.741*** -0.491*** 0.005 -0.040
(0.089) (0.080) (0.092) (0.096) (0.178) (0.145) (0.017) (0.033)
T2×Plan -0.381*** -0.533*** -0.483*** -0.366*** -1.017*** -0.845*** -0.001 -0.005
(0.068) (0.081) (0.081) (0.069) (0.196) (0.181) (0.021) (0.023)
Cons an -0.013 -0.041 -0.002 -0.012 -0.050 0.009 0.012 0.030
(0.022) (0.026) (0.030) (0.029) (0.074) (0.047) (0.011) (0.028)
Fi m Type Main Connec ed Main Connec ed Main Connec ed Main Connec ed
Obse a ions 512 506 505 512 508 511 505 511
R-squa ed 0.715 0.629 0.577 0.586 0.324 0.438 0.980 0.981
No e. The able epo s esul s o eg ession 2, whe e he ou come a iables a e ac ions he i m ook in
he h ee mon hs leading up o he ollow-up su ey. Those ac ions a e he change in p ices (columns (1)
and (2)), change in in es men (columns (3) and (4)), change in employmen (columns (5) and (6)), and
change in wages (columns (7) and (8)). Plan is he plan ha he i m had in he baseline su ey o he
nex h ee mon hs. T1is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou
he a e age GDP o ecas , and T2is an indica o ha is equal o one i i m i ecei ed he in o ma ion
ea men abou GDP unce ain y. Columns (1), (3), (5), and (7) show esul s o he i ms ha ecei ed he
in o ma ion ea men in he baseline pe iod, and columns (2), (4), (6), and (8) show esul s o he i ms
ha a e connec ed o he ea ed i ms. Robus s anda d e o s a e shown in pa en heses.
he p io s. As we con ol o he p io s, he ins umen uses he a ia ion only om he
change in expec a ions induced by he ea men . The e o e we can in e p e he es ima es
β, γ as causal e ec s.
To es ima e he spillo e e ec s, he speci ica ion ocuses on connec ed i ms i, while
con olling o he co esponding ac ion (and plan) o he main i m, which is ins u-
men ed by he ea men in e ac ed wi h he main i m’s plan. The eason o he ad-
di ional con ol is he exclusion es ic ion. The pos e io ins umen is alid i he only
way he ea men a ec s he connec ed ac ion is h ough he connec ed pos e io s. How-
e e , i is also possible ha he ea men a ec s he connec ed i m’s ac ion h ough he
main i m’s ac ion, wi hou e e ha ing changed he connec ed pos e io s (e.g., he con-
nec ed i m simply changes i s p ice in esponse o he main i m changing i s p ice). This
would iola e he exclusion es ic ion. We con ol o he main i m’s ac ion o mili a e
agains his possibili y.
18
Table 4 epo s he esul s. A 1 pe cen age poin inc ease in i ms’ mean GDP g ow h
expec a ions leads o a s a is ically insigni ican 0.16 pe cen age poin inc ease in he main
i m’s p ices and a signi ican 0.42 pe cen age poin inc ease o connec ed i ms. Employ-
men ises signi ican ly by 0.91 pe cen age poin s o he main i m and 0.64 pe cen age
poin s o connec ed i ms, espec i ely, ela i e o hei ini ial plans.9We ind no signi i-
can e ec on in es men o wages. Rega ding expec a ions o unce ain y, a 1 pe cen age
poin inc ease in unce ain y leads o a 0.34 (0.33) pe cen age poin dec ease in he main
(connec ed) i m’s p ices, a 0.82 (0.52) pe cen age poin decline in in es men , and a 0.81
(0.78) pe cen age poin d op in employmen , all o which a e signi ican . We ind no sig-
ni ican e ec on wages.
Table 4: Causal E ec o Expec a ions on Ac ions
P ice In es men Employmen Wage
(1) (2) (3) (4) (5) (6) (7) (8)
Pos e io mean 0.163 0.419∗∗∗ 0.008 0.065 0.912∗∗ 0.644∗0.024 -0.019
(0.114) (0.125) (0.224) (0.170) (0.419) (0.386) (0.026) (0.015)
Pos e io unce ain y -0.335∗∗∗ -0.331∗∗∗ -0.824∗∗∗ -0.515∗∗∗ -0.810∗∗∗ -0.779∗∗∗ 0.005 0.007
(0.042) (0.071) (0.083) (0.103) (0.173) (0.217) (0.010) (0.011)
Ac ionmain 0.236∗0.317∗∗∗ 0.091 0.348
(0.139) (0.083) (0.083) (0.323)
Obse a ions 485 453 478 452 479 454 479 452
Fi m Type Main Connec ed Main Connec ed Main Connec ed Main Connec ed
F(mean) 110.8 50.7 151.8 48.1 118.9 60.1 109.3 44.0
F(unce ain y) 365.3 187.8 777.3 158.7 402.0 191.0 386.8 191.9
F(ac ion) 45.5 64.6 16.1 0.9
No e. The able epo s esul s o eg ession 3, whe e he ou come a iables a e ac ions he i m ook in he
h ee mon hs leading up o he ollow-up su ey. Those ac ions a e he change in p ices (columns (1) and
(2)), change in in es men (columns (3) and (4)), change in employmen (columns (5) and (6)) and change
in wages (columns (7) and (8)). P os e io mean is he GDP o ecas o he i m in he ollow-up pe iod.
Pos e io unce ain y is he unce ain y abou he GDP o ecas o i m iin he ollow-up pe iod, measu ed
as he absolu e alue o he dis ance be ween he mos and leas likely scena io. Ac ionmain is he ac ion o
he main i m in he ollow-up pe iod. Va iables no shown bu included in he speci ica ion: P lan is he
plan ha he i m had in he baseline su ey o he nex h ee mon hs; P io mean is he GDP o ecas o he
i m in he baseline pe iod be o e ecei ing he ea men , and P io unce ain y is he unce ain y o ecas
be o e he ea men . We ins umen he pos e io a iables wi h he p io s in e ac ed by he ea men
dummy. Fo connec ed i ms, we also ins umen he co esponding ac ion o he main i m wi h he plan
in e ac ion by he ea men dummy. Columns (1), (3), (5), and (7) show esul s o he i ms ha ecei ed
he in o ma ion ea men in he baseline pe iod, and columns (2), (4), (6), and (8) show esul s o he i ms
ha a e connec ed o he ea ed i ms. The i s s age F-s a is ics a e shown a he end o he able. Robus
s anda d e o s a e shown in pa en heses.
9When we pool main and connec ed i ms, he a e age e ec s o mean expec a ions on p ices and em-
ploymen a e signi ican ac oss all i ms; see Table A-6 in he Online Appendix.
19
The opposing e ec s o he pos e io mean and unce ain y on i ms’ ac ions align wi h
economic in ui ion. When i ms an icipa e economic g ow h, hey inc ease hei p ices
and employmen , as i hey expec highe demand o hei goods. Con e sely, highe
unce ain y educes hei p ices, in es men , and employmen decisions, ela ed o he
con ac iona y e ec o highe unce ain y (Bake e al.,2024). The es ima ed impac o
unce ain y is pa icula ly obus . Mo eo e , he magni udes a e e y simila be ween
main and connec ed i ms.
Summa izing, we ind ha changes in expec a ions (bo h i s and second momen s)
signi ican ly a ec i ms’ decisions; his esul con i ms he indings in Kuma e al. (2023).
Mos impo an ly, we p esen a no el inding: changes in expec a ion a ec he connec ed
i ms’ ac ions, and wi h a magni ude simila o ha o he main i ms. These indings sug-
ges ha in o ma ion om ea ed i ms is eaching hei connec ed i ms, ei he h ough
di ec communica ion o by in e ing expec a ions om obse ed changes in ac ions. We
in es iga e hese channels in Sec ion 4. Addi ionally, in Sec ion 5, we discuss he implica-
ions o hese indings o he s eng h o communica ion.
Rega dless o he ansmission channel, ou indings ha e impo an implica ions o
he con agion o expec a ions wi hin he inpu -ou pu ne wo k. F om a policy pe spec-
i e, cen al banks could le e age his mechanism o s a egically dissemina e in o ma ion
h oughou he economy. A he same ime, i also aises conce ns abou he po en ial o
pessimis ic expec a ions o p opaga e, ampli ying down u ns h ough ne wo k e ec s.
4 The Role o Communica ion
In his sec ion, we p o ide empi ical e idence suppo ing he hypo hesis ha communica-
ion be ween a i m and i s connec ed i m is an impo an d i e o in o ma ion di usion.
We especially ule ou he al e na i e explana ion ha connec ed i ms a e solely upda ing
hei belie s because hey obse e changes in he ac ions o he main i m. Fo ins ance,
Table 4shows ha an inc ease in GDP unce ain y causes he (main) i m ha is di ec ly
ecei ing his in o ma ion o educe i s in es men (Column 3). A connec ed i m, obse -
20
5.1 Se up
We conside he sec o -le el Phillips cu e de i ed in Rubbo (2023) and assume each o
he Nsec o s is ep esen ed by a i m. The p ice ec o p is gi en by13
p = ∆ κy +βΩe
E p +1+ Ωp −1,(5)
whe e y is a measu e o slack in he economy ha we assume o be ou pu g ow h in
de ia ion om he s eady s a e; e
E is a gene ic expec a ions ope a o , possibly di e en
om he ull-in o ma ion a ional expec a ions one; βdeno es he discoun ac o ; Ωis
an in e ible ma ix whose elemen s a e con olu ed exp essions o he in ensi y o inpu -
ou pu linkages (IO ≡[ιij]ma ix) among i ms as well as hei labo sha es (α≡[αi]
ec o ), p ice lexibili y (Φ≡[ϕi]diagonal ma ix), and consump ion sha es (ψ≡[ψi]
ec o ); κis he ec o o Phillips cu e slopes; and ∆ = (I+βΩ)−1.14
Ou pu g ow h is assumed o be exogenously gi en and is d i en by an iid shock, ε∗
+1,
wi h mean ze o and a iance σ2
ε, and a de e minis ic sequence µ∗
:
y +1 =µ∗
+ε∗
+1.(6)
The long- un beha io o µ∗
is assumed o con e ge o ha o an iid no mal s ochas ic
p ocess wi h mean 0 and s anda d de ia ion σµ∗, and ha is independen o he p ocess
o ε∗
. Simila o Ilu and Schneide (2014), we assume ha i ms canno dis inguish he
de e minis ic sequence om he iid shocks e en i hey obse e an in ini ely la ge amoun
o da a. As a esul , equa ion (6) desc ibes a la ge amily o possible p ocesses ha can
ha e a he di e en implica ions in he sho un, o example, because hey di e in he
condi ional mean µ∗
. This is consequen ial because, by i e a ing equa ion (5) o wa d, a
13The ec o o i m-le el in la ion a es is desc ibed by π =p −p −1=βΩe
E [π +1]+κy −(I−Ω)p −1.
We no e ha ou se ing abs ac s om p oduc i i y shocks, implying ha all o ou esul s go h ough
e en i he slack in he economy is measu ed by he ou pu gap—which equals ou pu in he absence o
p oduc i i y shocks—as in Rubbo (2023).
14As p o ed by Rubbo (2023), Ωis in e ible as long as no i m has ully lexible p ices. We desc ibe he
s uc u e o he ma ices and ec o s in equa ion (5) in Appendix B.1 in mo e de ail.
27

i m’s op imal p icing decision depends on i s expec ed u u e ou pu g ow h, ha is, i s
pe cei ed µ∗
.
5.2 Ou pu G ow h Expec a ions
To sol e he model we ha e o discipline i ms’ expec a ions abou u u e g ow h. In do-
ing so, we conside h ee componen s ha we desc ibe in de ail below.
Unce ain y. We assume ha i ms ace Knigh ian unce ain y and a e a e se o ambigu-
i y a ising om no being able o dis inguish he de e minis ic componen om he iid
componen o g ow h. Fi ms hen base hei ac ions on he mos pessimis ic possible ou -
come. To discipline he i ms’ belie se o ou pu g ow h, we ollow a s a egy simila o
ha in Ilu and Schneide (2014). Speci ically, i m j’s pe cei ed law o mo ion abou ou pu
g ow h in de ia ion om he s eady s a e is gi en by
y +1 =µj +εj, +1, µj ∈[−aj ,−aj + 2|aj + ¯a|](7)
whe e i m jpe cei es he de e minis ic componen o g ow h o ange be ween −aj and
−aj + 2|aj + ¯a|wi h aj being a mean 0 iid shock a ound he s eady-s a e ambigui y le el
¯a > 0, and he ealized aj is assumed small enough so ha ¯a+aj >0.15 A wide ange
o µj , ha is, a la ge |aj + ¯a|o equi alen ly highe aj , also implies a lowe wo s case
scena io o ou pu g ow h. Since i ms base hei ac ions on he mos pessimis ic possible
ou come, hei p io expec a ions abou u u e ou pu g ow h a e gi en by
e
Ep io
j y +1 = min
µj ∈[−aj ,−aj +2|aj +¯a|]µj =−aj .(8)
15Ilu and Schneide (2014) lay ou he equa ion ha is analogous o ou equa ion (7) be o e exp essing
a iables in de ia ion om hei s eady s a e. The co esponding ange a ound he de e minis ic compo-
nen in ha case would be −aj −¯a, −aj −¯a+ 2|aj + ¯a|. Mo eo e , ou pu g ow h and i s de e minis ic
componen a e pe cei ed o con e ge o −¯a; hence sub ac ing by −¯ain he ange abo e yields he ange
o µj in equa ion (7). The e o e, aj desc ibes ambigui y a ound a o ecas o ou pu g ow h in de ia ion
om he s eady s a e ha equals ¯a, o ambigui y a ound a o ecas o no ou pu g ow h.
28
In o ma ion ea men . As in he expe imen , ea ed i m j ecei es in o ma ion abou
p o essional o ecas e s’ ange a ound hei ou pu g ow h o ecas , which i m jin e -
p e s o equal 2|a∗
+ ¯a|so ha i is cen e ed a ¯ajus like µj .16 Upon ecei ing his in o -
ma ion, he i m upda es i s expec a ions abou u u e ou pu g ow h o
e
Epos
j y +1 =−aj (1 −gj) + gjsj ,(9)
whe e sj =−a∗
i i m jis ea ed and sj = 0 o he wise; gj∈[0,1] deno es he gain
om he in o ma ion ea men .17 We he ea e in e changeably e e o sj as a signal o
ea men .
Communica ion. Consis en wi h ou empi ical e idence, we assume i ms communica e
hei expec a ions abou u u e ou pu g ow h wi h each o he acco ding o an exogenous
communica ion ma ix C, desc ibed in De ini ion 1, ha i ms ake as gi en.18
DEFINITION 1. The communica ion ne wo k is desc ibed by ma ix C= [cij], whe e cij ∈[0,1]
quan i ies he in ensi y wi h which i m jcommunica es i s expec a ions abou u u e ou pu
g ow h o i m i, so ha PN
j=1 cij = 1. No communica ion co esponds o C=I, and we de-
ine e en communica ion as C=1N×N/N.
E en communica ion is he case whe e all i ms communica e equally wi h one-ano he ,
which is mo i a ed by he empi ical e idence o symme ic communica ion (Figu e 4).19
16Ou analysis goes h ough simila ly in he case o a ea men abou µ∗
, co esponding o ea men 1 in
ou empi ical analysis.
17The ea ed i m upda es he lowe bound o he de e minis ic componen ange o −(1 −gj)aj −gja∗
.
18This ules ou i ms s a egically choosing o communica e pa s o in o ma ion wi h o he i ms (i.e.,
endogenous C). This is easonable in ou se ing as we do no ind e idence o any s a egic beha io (see
Sec ion 4).
19Ou empi ical e idence highligh s symme y in communica ion be ween any supplie -cus ome pai s
o i ms, which we ex end o apply o he es o he ne wo k o which ou pai is connec ed. I is plausible
ha he in ensi y o communica ion be ween he supplie o cus ome i ms and he es o he ne wo k is
asymme ic. Howe e , since we do no obse e he in ensi ies wi h which ou supplie and cus ome i ms
communica e wi h he es o hei ne wo k, e en communica ion is a use ul and in o ma i e benchma k
ha also enables us o de i e some analy ical esul s.
29
The inal expec a ions o any i m iabou ou pu g ow h a e gi en by
e
Ei y +1 = 1−
N
X
j=i
cij!
| {z }
=cii e
Epos
i y +1 +
N
X
j=i
cij e
Epos
j y +1,(10)
whe e PN
j=icij = (1 −cii)is he o al exposu e o in o ma ion om communica ion. The
ec o o all i ms’ expec a ions abou u u e g ow h can be w i en as
e
E y +1 =C[−(I−G)a +Gs ](11)
whe e y =1y ,a is he ec o o i m ambigui y, Gis a diagonal ma ix whose diagonal
equals i m gain om in o ma ion ea men , and s is he ec o o signals.20
5.3 Solu ion and Implica ions
We now u n o he solu ion o he model and he implica ions o communica ion o i ms’
p ices, he agg ega e p ice le el, and in la ion.
P ice ec o . P oposi ion 1p o ides he equilib ium p ice ec o and shows ha an in-
o ma ion ea men abou highe unce ain y (lowe sj ) leads o lowe p ices, as docu-
men ed by ou empi ical esul s in Table 4.
PROPOSITION 1. The equilib ium p ice ec o is desc ibed by
p =βMC(Gs −(I−G)a ) + Myy +Mpp −1,
whe e M=Mp×Myand all i s elemen s a e posi i e.
20We no e ha i he e is pe ec in o ma ion abou µ∗
, and, as a esul , he e is no ambigui y abou he
de e minis ic componen o g ow h, ha is, ai = 0 o any i m i, hen he model eco e s he one in
Rubbo (2023), which abs ac s om impe ec in o ma ion, ambigui y a e sion, and communica ion ne -
wo ks. Fi ms being ully in o med abou µ∗
and a ional implies ha i ms’ communica ion abou hei
expec a ions o ou pu g ow h is i ele an since all i ms sha e he same expec a ions, e
Ei y +1 =µ∗
o any
i.
30
The impac o a ea men o i m jon he p ice o i m ican be decomposed in o
∂pi
∂sj
=βMijcjjgj
| {z }
ea ed i m ac ion channel
+βX
k=j
Mikckjgj
| {z }
ea ed i m communica ion channel
≥0.(12)
The i s componen desc ibes he e ec o ea men s o he ex en ha he ac ions o he
ea ed i m a ec i m i, as cap u ed by Mij. The second componen desc ibes he commu-
nica ion e ec ha esul s om he ea ed i m sha ing in o ma ion wi h i s p oduc ion
ne wo k (including i) and hose i ms eac ing o he new in o ma ion. Absen communi-
ca ion (C=I), i m i’s p ice will only espond o he ea men sj ia he ac ions o he
ea ed i m j( he i s componen ). The communica ion channel ( he second componen )
becomes mo e impo an o he e ec on i m i’s p ice as communica ion inc eases.21
Co olla y 1 o malizes ou symme y esul : unde e en communica ion (C=1N×N/N),
he eac ion o all i ms’ expec a ions and p ices is independen o whe e he ea men
o igina es. Hence, e en communica ion implies a symme ic ups eam s downs eam
p opaga ion o shocks o ou pu g ow h expec a ions.
COROLLARY 1. Suppose all ea ed i ms place he same weigh on he ea men (gj=g,∀j).
The i m whe e he ea men o igina es is i ele an o he esponse o ou pu g ow h expec a ions
and p ices unde e en communica ion, ha is, ∂
e
Ei y +1
∂sj =∂
e
Ei y +1
∂sk and ∂pi
∂sj =∂pi
∂sk ,∀iand ∀j=k.
In con as o Co olla y 1 ha cha ac e izes he esponse o a gi en i m’s p ice when
a ying he ea ed i m, Co olla y 2cha ac e izes he dispe sion in p ice esponses ac oss
i ms o a gi en ea ed i m. Speci ically, Co olla y 2p o es ha o a gi en in o ma ion
ea men sj , he ini ial p ice esponse a ies ac oss i ms when communica ion is absen .
While, in he p esence o e en communica ion, he e is p ice dispe sion only o he ex en
ha i ms’ Phillips cu e slopes a e he e ogeneous.
COROLLARY 2. Suppose all ea ed i ms place he same weigh on he ea men (gj=g,∀j).
Following a ea men o i m j, he e will always be dispe sion in he ini ial i m p ice esponses
21We no e ha i all i ms ecei e he same signal abou u u e g ow h and upda e hei expec a ions
simila ly (G=gI), communica ion has no e ec on ou pu g ow h expec a ions.
31
absen communica ion; unde e en communica ion, he e will be dispe sion in he ini ial i m p ice
esponses only o he ex en ha he Phillips cu e slopes, κ, a e he e ogeneous.
Agg ega e p ice le el and in la ion. A one- ime ea men abou u u e ou pu g ow h
unce ain y causes a pe manen shi in he p ice ec o and in he agg ega e p ice le el,
de ined as P =ψ′p , whe e ψis he ec o o consump ion sha es. I is s aigh o wa d
ha he symme y esul in Co olla y 1ca ies o e simila ly o he agg ega e p ice. Fi ms
adjus hei p ices un il hey con e ge o he new long- un agg ega e p ice le el, desc ibed
by P oposi ion 2in he Appendix.22 Co olla y 3 o malizes he condi ion o which com-
munica ion ampli ies he esponse o he long- un agg ega e p ice le el o an in o ma ion
ea men ecei ed by i m j.
COROLLARY 3. Rela i e o no communica ion, e en communica ion ampli ies he esponse o he
long- un agg ega e p ice o a ea men ecei ed by i m ji he ollowing inequali y is sa is ied:
λjκj(1 −ϕj)
ϕj
<
N
X
i=j
λiκi(1 −ϕi)
ϕi(N−1) ,(13)
whe e λ=ψ′(I−IO)−1is he ec o o Doma weigh s ha measu e i m size. A mo e posi i e
di e ence be ween he igh -hand side and le -hand side in (13) implies a la ge ampli ying e ec
o communica ion on he long- un agg ega e p ice compa ed o no communica ion.
When mul iplied by gj, he le -hand side o he inequali y in Equa ion (13) coincides
wi h he esponse o he long- un agg ega e p ice o a ea men p o ided o i m j, in
he absence o communica ion. An insigh o Co olla y 3is hen ha communica ion has
s onge p opaga i e e ec s— ela i e o no communica ion—on he long- un p ice le el
when ea ing a i m ha , absen communica ion, igge s a weake e ec on he long- un
agg ega e p ice.
In he special case when i ms ha e simila p ice lexibili y and Phillips cu e slopes,
communica ion has s onge p opaga i e e ec s on he long- un agg ega e p ice he smalle
22The esponse o he long- un agg ega e p ice le el o a one- ime ea men abou u u e ou pu g ow h
unce ain y, depends on he in e ac ion o i ms’ Doma weigh s, hei p ice lexibili y, hei Phillips cu e
slopes, and he in ensi y o communica ion.
32

he ea ed i m’s Doma weigh is ela i e o he es o he i ms in he ne wo k. I i ms
also ha e he same consump ion sha es (ψ=1/N), a smalle Doma weigh co esponds
o being mo e downs eam, implying highe p opaga i e e ec s o communica ion when
he ea men o igina es om he cus ome i m (downs eam) as opposed o he supplie
i m (ups eam) in any supplie -cus ome pai .23 In ui i ely, a cus ome i m’s p oduc
p ice is less ele an o i s supplie , han he supplie ’s p ice is o he cus ome . Thus,
absen communica ion, shock p opaga ion is weake when o igina ing downs eam.
Agg ega e in la ion is equal o ¯π =ψ′(p −p −1). Gi en ha p −1is p e-de e mined,
he symme y esul o Co olla y 1 o p ca ies o e wi h he ea men o igin being
i ele an o agg ega e in la ion unde e en communica ion. Using P oposi ion 1 o sol e
o agg ega e in la ion,
¯π =ψ′βMC(Gs −(I−G)a ) + Myy + (Mp−I)p −1.(14)
No ing ha PN
j=1(Mp−I)ij = 0 ∀i, hen communica ion lowe s he pe sis ence o he
agg ega e in la ion esponse o he ex en ha i homogenizes he ini ial p ice esponses
ac oss i ms. In ui i ely, he ea men will a ec u u e in la ion o he ex en ha cu en
p ices ha e no eached he new long- e m agg ega e p ice le el. Homogenei y in he
ini ial p ice esponses implies immedia e adjus men o all p ices o he new agg ega e
p ice le el.24
5.4 Quan i a i e Analysis
We now explo e he ole o communica ion quan i a i ely. To be close o he empi ical
se up, we conside a h ee- i m model, whe e one i m is he cus ome i m ( i m 3), one
he supplie i m ( i m 2), and he o he one cap u es he es o he ne wo k ( i m 1). The
23The ups eamness is measu ed by 1′((I−IO)−1−I); see Ca alho and Tahbaz-Salehi (2019).
24F om P oposi ion 1and Pj(Mp)ij = 1 ∀i, i ollows ha i he ini ial p ice esponse is homogeneous,
p ∝1, hen p +h=p and P +h=P ,∀h > 0. The dynamics in his case mi o s ha o he ep esen a i e
i m model, wi h s anda d Phillips cu e π =βe
E π +1 +κy . A one- ime ea men abou u u e ou pu
g ow h unce ain y in pe iod will lead o a one- ime change in in la ion in pe iod , a e which in la ion
e e s o i s s eady s a e.
33
cus ome i m pu chases i s inpu s om i ms 1 and 2, while he supplie pu chases i s
inpu s only om i m 1.25 We simula e he p ice ec o esponse o a one- ime uni in o -
ma ion ea men o di e en alues o he supplie ’s and cus ome ’s inpu sha es in he
case o no communica ion and e en communica ion.26 We calib a e he model o ma ch
key cha ac e is ics o i ms in ou sample, such as p ice lexibili y and labo cos sha e,
while se ing o he pa ame e s o s anda d alues. We desc ibe he calib a ion s a egy
and de ails o he simula ion exe cise in Appendix B.1 and ocus nex on ou h ee esul s
om he simula ion exe cise.
Resul 1: Communica ion gene a es symme y in ups eam s downs eam ansmis-
sion o a ea men . Figu e 5plo s he dis ibu ion o he ini ial esponse o i m p ices o
an in o ma ion ea men abou highe g ow h unce ain y p o ided o he supplie (le
panel) and o he cus ome ( igh panel). The ed dis ibu ions in Figu e 5show ha when
i ms communica e, whe he he ea men was p o ided o he supplie o he cus ome
i m does no ma e o i s impac on p ices, as s a ed in Co olla y 1. By con as , absen
communica ion, he o igin o he ea men ma e s o he p ice esponses, as isualized
by he as ly di e en blue dis ibu ions in he le and igh panels. This esul is consis-
en wi h ou empi ical e idence in Table 6o he symme ic spillo e e ec s o ea men s
on connec ed i ms.
Resul 2: Communica ion educes he dispe sion o p ice esponses o a ea men . Fig-
u e 5 u he shows ha communica ion hea ily educes he dispe sion o p ice changes
on impac ollowing a ea men o he supplie o cus ome : as Co olla y 2highligh s, any
emaining p ice dispe sion when i ms communica e is explained by hei he e ogeneous
25The assumed s uc u e o he ne wo k implies he ollowing inpu -ou pu ma ix and labo sha e ec o :
IO =

0 0 0
ι21 0 0
ι31 ι32 0
and α=1α2α3′, so ha IO +α=1.
26In Online Appendix C-4, we conside ano he communica ion s a egy wi h C= Ω, so ha he in ensi y
o communica ion equals he sensi i i y o he i ms’ p ice changes o he ec o o u u e expec ed p ice
changes.
34
Figu e 5: Dis ibu ion o he Impac on P ices a e T ea men s o Highe Unce ain y
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Pe cen
0
1
2
3
4
5
6
7#10-3 T ea men o supplie - m
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Pe cen
0
1
2
3
4
5
6
7#10-3 T ea men o cus ome - m
No communica ion
E en communica ion
No e: Dis ibu ion o ini ial p ice esponses ac oss all h ee i ms when he ea ed i m is he supplie (le
panel) and when he ea ed i m is he cus ome ( igh panel). In ed: e en communica ion; in blue: no
communica ion.
Phillips cu e slopes.27 The low p ice dispe sion unde communica ion is consis en wi h
ou empi ical esul s: mos ea ed i ms epo communica ing (Figu e 3) and he a e -
age ea men e ec on p ices is almos iden ical o he main and connec ed i ms (Table
3, Columns 1 and 2).
A mo e di ec empi ical alida ion would be con as ing he ea men e ec s on p ices
o i ms ha do and do no communica e, expec ing g ea e di e gence in he la e . Al-
hough exogenous a ia ion in communica ion would be needed o de ini i e e idence
— which we do no ha e — we none heless show he esul s o his exe cise in Table
C-7 in he Appendix. In bo h he da a and he simula ions, we compu e he connec ed
i m’s p ice change associa ed wi h a 1 pe cen age poin exogenous ise in he ea ed
i m’s p ice. We ind ha when he e is communica ion, he es ima ed p ice ela ion-
ship be ween he ea ed and connec ed i m app oaches uni y bo h empi ically and in
he model, ega dless o whe he he ea men was p o ided o a cus ome o a supplie
i m.28 Absen communica ion, he poin es ima e o he p ice ela ionship is always less
27Figu e C-1 in he Online Appendix shows ha he educ ion in p ice dispe sion holds ue ac oss i ms
wi hin each h ee- i m ne wo k.
28This esul can explain why o he wo ks, such as Ca alho e al. (2021), ind a simila p opaga ion o
shocks o downs eam and ups eam i ms.
35
hen uni y, bu highe when he ea men is p o ided o he supplie compa ed o when
he ea men is gi en o he cus ome —bo h in he su ey da a and in he model simula-
ions.
Figu e 6: E olu ion o P ices and Agg ega e In la ion a e T ea men s o Highe Unce -
ain y
02468
Time
-0.8
-0.6
-0.4
-0.2
0
P ice le el
A: P ices, ea men o supplie
02468
Time
-0.8
-0.6
-0.4
-0.2
0
P ice le el
B: P ices, ea men o cus ome
Fi m 1, no comm
Supplie , no comm
Cus ome , no comm
Fi m 1, comm
Supplie , comm
Cus ome , comm
02468
Time
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
Pe cen
C: In.a ion, ea men o supplie
02468
Time
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
Pe cen
D: In.a ion, ea men o cus ome
No communica ion
E en communica ion
No e: Panels A and B plo he e olu ion o p ice le els, a e aged ac oss simula ions, when a ea men o
highe unce ain y abou u u e ou pu g ow h is p o ided o he supplie and cus ome . Panels C and D
plo he e olu ion o agg ega e in la ion, implied by he p ice dynamics in panels A and B. In ed: e en
communica ion; in blue: no communica ion.
Resul 3: Communica ion leads o a s onge and sho e -li ed esponse o in la ion o
u u e ou pu g ow h unce ain y. Panels A and B in Figu e 6plo he e olu ion o i m
p ices o e ime, a e aged ac oss simula ions, when he supplie is ea ed (le panel)
and when he cus ome is ea ed ( igh panel). Communica ion ampli ies he a e age
esponse o he long- un agg ega e p ice le el, wi h la ge ampli ica ion compa ed o he
no communica ion case when he ea ed i m is he cus ome compa ed o he supplie .
This is consis en wi h Co olla y 3, no ing ha he cus ome i m is he mos downs eam
36
Geo ga akos, D., Go odnichenko, Y., Coibion, O., and Kenny, G. (2024). The causal e ec s
o in la ion unce ain y on households’ belie s and ac ions. Technical epo , Na ional
Bu eau o Economic Resea ch.
Ghassibe, M. (2021). Mone a y policy and p oduc ion ne wo ks: an empi ical in es iga-
ion. Jou nal o Mone a y Economics, 119:21–39.
Ilu , C. and Schneide , M. (2023). Chap e 24 - ambigui y. In Bachmann, R., Topa, G., and
an de Klaauw, W., edi o s, Handbook o Economic Expec a ions, pages 749–777. Aca-
demic P ess.
Ilu , C. L. and Schneide , M. (2014). Ambiguous business cycles. Ame ican Economic Re-
iew, 104(8):2368–99.
Jackson, M. O. (2008). Social and economic ne wo ks, olume 3. P ince on uni e si y p ess
P ince on.
Jaimo ich, N. and Rebelo, S. (2009). Can News abou he Fu u e D i e he Business Cycle?
Ame ican Economic Re iew, 99(4):1097–1118.
Jamilo , R., Kohlhas, A., Tala e a, O., and Zhang, M. (2024). G anula sen imen s.
Kie en, P., K¨
onig-Ke s ing, C., Schmid , R., T au mann, S., and Heinicke, F. (2025). Fi s -
o de and highe -o de in la ion expec a ions: E idence abou Households and Fi ms.
Jou nal o Economic Beha io & O ganiza ion, 233:106988.
Kuma , S., Go odnichenko, Y., and Coibion, O. (2023). The e ec o mac oeconomic un-
ce ain y on i m decisions. Econome ica, 91(4):1297–1332.
La’O, J. and Tahbaz-Salehi, A. (2022). Op imal mone a y policy in p oduc ion ne wo ks.
Econome ica, 90(3):1295–1336. Publishe : Wiley Online Lib a y.
Liu, L. and Hudgens, M. G. (2014). La ge Sample Randomiza ion In e ence
o Causal E ec s in he P esence o In e e ence. Jou nal o he Ame ican
S a is ical Associa ion, 109(505):288–301. Publishe : Taylo & F ancis ep in :
h ps://doi.o g/10.1080/01621459.2013.844698.
43

Lucas J , R. E. (1972). Expec a ions and he Neu ali y o Money. Jou nal o economic heo y,
4(2):103–124. Publishe : Academic P ess.
Nikolakoudis, G. (2024). Incomple e in o ma ion in p oduc ion ne wo ks.
Ozdagli, A. and Webe , M. (2023). Mone a y policy h ough p oduc ion ne wo ks: E i-
dence om he s ock ma ke . Technical epo , Na ional Bu eau o Economic Resea ch.
Pas en, E., Schoenle, R., and Webe , M. (2020). The p opaga ion o mone a y policy shocks
in a he e ogeneous p oduc ion economy. Jou nal o Mone a y Economics, 116:1–22.
Pelle , T. and Tahbaz-Salehi, A. (2023). Rigid p oduc ion ne wo ks. Jou nal o Mone a y
Economics, 137:86–102.
Pollmann, M. (2023). Causal In e ence o Spa ial T ea men s. Numbe : a Xi :2011.00373
a Xi :2011.00373 [econ, s a ].
Rubbo, E. (2023). Ne wo ks, phillips cu es, and mone a y policy. Econome ica,
91(4):1417–1455.
Sebbesen, A. and Obe ho e , H. (2024). The p opaga ion o business expec a ions
wi hin he Eu opean Union. Jou nal o Applied Econome ics, 39(6):1082–1103. ep in :
h ps://onlinelib a y.wiley.com/doi/pd /10.1002/jae.3075.
Sims, C. A. (2003). Implica ions o a ional ina en ion. Jou nal o mone a y Economics,
50(3):665–690.
Song, W. and S e n, S. (2024). Fi m Ina en ion and he E icacy o Mone a y Policy: A
Tex -Based App oach. The Re iew o Economic S udies, page dae102.
Uhlig, H. (2001). A oolki o analysing nonlinea dynamic s ochas ic models easily. In
Compu a ional Me hods o he S udy o Dynamic Economies. Ox o d Uni e si y P ess.
an Rooij, M., Coibion, O., Geo ga akos, D., Candia, B., and Go odnichenko, Y. (2024).
Keeping Up wi h he Jansens: Causal Pee E ec s on Household Spending, Belie s and
Happiness.
44
Webe , M., Candia, B., A ouzi, H., Ropele, T., Llube as, R., F ache, S., Meye , B., Kuma ,
S., Go odnichenko, Y., Geo ga akos, D., Coibion, O., Kenny, G., and Ponce, J. (2025).
Tell me some hing i don’ al eady know: Lea ning in low- and high-in la ion se ings.
Econome ica, 93(1):229–264.
45
Appendix
A Addi ional Figu es and Tables
A.1 Figu es
Figu e A-1: Communica ion abou GDP: Ups eam s Downs eam (Connec ed Fi ms)
No e: This igu e shows he ea men e ec on he numbe o imes he connec ed i m epo ed commu-
nica ing wi h he main i m abou GDP, du ing he h ee mon hs p io o he ollow-up. This is shown
sepa a ely o each ea men , Tn, and sepa a ely whe he he main i m is he supplie o cus ome in he
pai . 95% con idence in e als a e displayed.
Figu e A-2: Reasons o Sha ing In o ma ion abou GDP
No e: The numbe o i ms (main and connec ed) lis ing he labeled esponse as a eason o sha ing in o -
ma ion abou GDP.
1
A.2 Tables
Table A-1: Fi m Coun s and Pe cen ages by Sec o and Size Ca ego y
5 o less Wo ke s 6–19 Wo ke s 20–49 Wo ke s 50+ Wo ke s To als
Numbe % Numbe % Numbe % Numbe % Numbe %
Panel A: S a s NZ Reco ds
Manu ac u ing 5286 48 3663 33 1239 11 771 7 10959 100
Wholesale T ade 4107 54 2328 31 705 9 396 5 7536 100
Re ail T ade 7317 58 3945 31 735 6 618 5 12615 100
To als 16710 54 9936 32 2679 9 1785 6 31110 100
Panel B: Fi ms App oached
Manu ac u ing 2610 46 1934 34 729 13 347 6 5620 51
Wholesale T ade 2451 51 1622 34 433 9 307 6 4813 64
Re ail T ade 3122 54 1996 35 295 5 364 6 5777 46
To als 8183 50 5552 34 1457 9 1018 6 16210 52
Panel C: Main Wa e Fi ms Sample
Manu ac u ing 70 3 444 23 362 50 251 72 1127 20
Wholesale T ade 45 2 212 13 157 36 99 32 513 11
Re ail T ade 95 3 195 10 175 59 43 12 508 9
To als 210 3 851 15 694 48 393 39 2148 13
Panel D: Follow-up Fi ms Sample
Manu ac u ing 31 44 230 52 198 55 130 52 589 52
Wholesale T ade 18 40 111 52 72 46 47 47 248 48
Re ail T ade 33 35 109 56 73 42 26 60 241 47
To als 82 39 450 53 343 49 203 52 1078 50
No e: This able summa izes he numbe o i ms and hei pe cen age sha es by sec o and i m size ca ego y
ac oss di e en su ey s ages. Panels A and B a e popula ion and app oached samples, espec i ely, while
Panels C and D ep esen he main and ollow-up su ey samples. Pe cen ages in A and B a e sha es o he
popula ion and sum o 100 wi hin a ow (excluding he las column). Pe cen ages in C and D a e esponse
a es: hey a e he sha e o obse a ions ela i e o he same cell in he panel abo e (e.g. in ou main
wa e sample, panel C, we ha e 70 manu ac u ing i ms o 5 o less wo ke s, which is 3% o he 2610 i ms
app oached, panel B).
2

Table A-2: T ea men P edic ion o Fi ms’ Cha ac e is ics
(1) (2) (3) (4) (5)
Employmen (log) Indus y N o Rela ionships Fi m Age GDP P io
T ea men 1 -0.017 0.121 0.374 -2.891** 0.113
(0.053) (0.130) (0.450) (1.331) (0.097)
T ea men 2 -0.077 -0.013 0.226 -6.752*** 0.031
(0.051) (0.128) (0.439) (1.279) (0.097)
Cons an 3.016*** 5.168*** 9.800*** 32.653*** 1.418***
(0.036) (0.093) (0.352) (0.928) (0.070)
Obse a ions 2,010 2,010 1,799 1,722 2,010
R-squa ed 0.001 0.001 0.000 0.016 0.001
No e. The able epo s esul s o eg ession whe e he dependen a iable is ei he Employmen
in logs (Column 1), whe he he i m is in he manu ac u ing o ade indus y (Column 2),
numbe o i ms’ cus ome s and supplie (Column 3), i ms’ age (Column 4), o he p io GDP
expec a ion (Column 5). The independen a iables ake a alue o one i he i m ecei ed
ea men 1 o 2 and ze o o he wise, espec i ely. Robus s anda d e o s in pa en hesis.
3
Table A-3: P edic abili y o Pa icipa ion in he Follow-up Wa e
(1) (2) (3) (4) (5) (6)
T ea men 1 0.036 0.036 0.039 0.038 0.023 0.034
(0.027) (0.027) (0.027) (0.027) (0.029) (0.032)
T ea men 2 -0.023 -0.022 -0.021 -0.023 -0.023 -0.002
(0.027) (0.027) (0.027) (0.027) (0.029) (0.032)
Employmen (log) 0.019 0.016 0.014 0.004 0.004
(0.011) (0.012) (0.012) (0.013) (0.015)
Indus y: T ade -0.042*
(0.022)
Subsec o : Equipmen and Machine y 0.035 0.015 0.075
(N=203) (0.052) (0.056) (0.060)
Subsec o : Food and Be e age 0.087* 0.062 0.063
(N=268) (0.049) (0.053) (0.056)
Subsec o : Pape , wood, p in ing and u ni u e 0.092* 0.066 0.066
(N=262) (0.049) (0.053) (0.056)
Subsec o : Re ail T ade 0.012 -0.005 0.001
(N=480) (0.045) (0.047) (0.051)
Subsec o : Tex ile and clo hing 0.069 0.041 0.007
(N=155) (0.056) (0.059) (0.063)
Subsec o : Wholesale ade 0.025 -0.007 0.006
(N=485) (0.045) (0.048) (0.051)
Numbe o Rela ionships 0.002 0.001
(0.002) (0.002)
Fi m age 0.000
(0.001)
Cons an 0.497*** 0.440*** 0.469*** 0.413*** 0.441*** 0.418***
(0.019) (0.040) (0.043) (0.055) (0.059) (0.063)
Obse a ions 2,024 2,024 2,024 2,024 1,790 1,534
R-squa ed 0.004 0.004 0.005 0.008 0.006 0.006
No e. The able epo s esul s o eg ession whe e he dependen a iable is a a iable ha akes a alue o one
i he i m pa icipa ed in he ollow-up, and ze o o he wise. Numbe o Rela ionships is he numbe o supplie
and cus ome i ms ha he i m has. Robus s anda d e o s in pa en hesis. The i m in he subsec o “O he
S o e Re ailing” is included in “Re ail T ade” as he e is only one i m belonging o ha g oup.
4
Table A-4: He e ogeneous T ea men E ec s: Fi m Cha ac e is ics
(a) Expec a ions
Pos e io mean P os e io unce ain y
(1) (2) (3) (4) (5) (6) (7) (8)
T1×P io ×H0.070 0.020 0.003 -0.076 0.103∗∗ 0.007 0.060 -0.012
(0.047) (0.018) (0.066) (0.101) (0.052) (0.017) (0.066) (0.094)
T2×P io ×H0.009 0.019∗∗ -0.050 0.141 -0.040 -0.010 -0.062 -0.060
(0.056) (0.008) (0.094) (0.124) (0.054) (0.009) (0.063) (0.107)
He e ogenei y, HEmploymen Ma ke Sha e Age Manu ac u ing Employmen Ma ke Sha e Age Manu ac u ing
N505 275 419 505 513 280 427 513
(b) Ac ions
P ice In es men
(1) (2) (3) (4) (5) (6) (7) (8)
T1×Plan ×H-0.080 -0.039 -0.164 0.105 0.037 0.002 -0.041 0.163
(0.080) (0.037) (0.123) (0.158) (0.074) (0.020) (0.119) (0.173)
T2×Plan ×H0.167∗∗ -0.013 -0.099 -0.321∗-0.054 0.041 -0.177∗0.270∗∗
(0.071) (0.020) (0.115) (0.169) (0.067) (0.031) (0.103) (0.132)
He e ogenei y, HEmploymen Ma ke Sha e Age Manu ac u ing Employmen Ma ke Sha e Age Manu ac u ing
N506 288 435 506 512 288 438 512
Employmen Wages
(1) (2) (3) (4) (5) (6) (7) (8)
T1×Plan ×H0.197∗∗ 0.021 -0.009 -0.217 0.014 -0.008 -0.035 -0.107∗
(0.084) (0.103) (0.185) (0.296) (0.036) (0.007) (0.036) (0.060)
T2×Plan ×H0.133 0.023 0.380 0.207 -0.045 -0.016 0.002 -0.015
(0.184) (0.068) (0.264) (0.400) (0.033) (0.012) (0.015) (0.048)
He e ogenei y, HEmploymen Ma ke Sha e Age Manu ac u ing Employmen Ma ke Sha e Age Manu ac u ing
N511 291 435 511 511 288 435 511
No e. Panels (a) and (b) ex end he eg essions o equa ions (1) and (2), espec i ely, o include an in e ac ion
o each e m wi h a iable H, o he sample o connec ed i ms. Each column uses a di e en cha ac e is ic
o he main i m o H, as labeled (log employmen , ma ke sha e, log i m age, and a dummy equal o one
i in manu ac u ing). Only he iple in e ac ion e ms a e displayed, which iden i ies he e ec o Hon he
co ela ion o he pos e io wi h he p io , in he case o Panel (a), and he co ela ion o he ac ion wi h he
plan, in he case o Panel (b). S anda d e o s a e displayed in pa en heses.
5
Table A-5: He e ogeneous Spillo e E ec s on Ac ions: Ne wo k Cha ac e is ics
P ice In es men
(1) (2) (3) (4) (5) (6)
T1×Plan ×H-0.194 -0.014∗∗ -0.030 0.077 -0.006 -0.110
(0.158) (0.007) (0.158) (0.186) (0.008) (0.124)
T2×Plan ×H-0.083 0.014∗0.090 0.189 0.004 -0.163
(0.163) (0.008) (0.121) (0.132) (0.009) (0.125)
He e ogenei y, HUps eam Exp. Sha e N connec ions Ups eam Exp. Sha e N connec ions
N506 357 377 512 360 383
Employmen Wages
(1) (2) (3) (4) (5) (6)
T1×Plan ×H-0.687∗∗∗ 0.024∗∗∗ 0.029 -0.019 -0.004 -0.031
(0.191) (0.007) (0.283) (0.064) (0.003) (0.031)
T2×Plan ×H-0.485 0.009 0.217 0.080∗-0.001 -0.073
(0.321) (0.026) (0.277) (0.047) (0.003) (0.048)
He e ogenei y, HUps eam Exp. Sha e N connec ions Ups eam Exp. Sha e N connec ions
N511 356 385 511 358 385
No e. The speci ica ions ex end he eg ession o equa ion (2) o include an in e ac ion o each e m wi h
a iable H, o he sample o connec ed i ms. Each column uses a di e en cha ac e is ic o he main i m
o H, as labeled (a dummy equal o one i he cus ome , sha e o sales o o expendi u e on he connec ed
i m, and numbe o cus ome s o supplie s — i hey a e a supplie o cus ome , espec i ely, in he la e
wo). Only he iple in e ac ion e ms a e displayed, which iden i ies he e ec o Hon he co ela ion o
he ac ion wi h he plan. S anda d e o s a e displayed in pa en heses.
6
B.2.1 P oo o P oposi ion 1
The op imal p ice ec o depends on he cu en ou pu g ow h, ec o o signals, ec o
o ambigui y, and he ec o o pas p ices. Hence, we guess he ollowing solu ion:
p =Mss +Myy +Maa +Mpp −1,
implying ha expec a ions abou he p ice ec o in + 1 a e
e
E p +1 =Mye
E y +1 +Mpp =−MyC(I−G)a +MyCGs +Mpp .
Plugging expec a ions in o he op imal p ice equa ion and le ing K=diag(κ), we ha e
p = ∆Ky + ∆Ωp −1+β∆Ω [−MyC(I−G)a +MyCGs +Mpp ]
= (I−β∆ΩMp)−1∆Ky + ∆Ωp −1+β∆Ω (−MyC(I−G)a +MyCGs ).
F om he e, i ollows ha
Mp−β∆ΩM2
p= ∆Ω
My= (I−β∆ΩMp)−1∆K
Ma=−β(I−β∆ΩMp)−1∆ΩMyC(I−G)
Ms=β(I−β∆ΩMp)−1∆ΩMyCG
(17)
I is s aigh o wa d o see om he i s equa ion abo e ha Ms=βMpMyCGand Ma=
−βMpMyC(I−G). To ease no a ion, we se M=Mp×My. To sol e o Mp, we ely on
Theo em 3.5 in Uhlig (2001); o ensu e ha p ice dynamics a e s able, we only conside
he solu ion o Mpwhose eigen alues a e wi hin he uni ci cle. To p o e ha all he
elemen s o ma ix Ma e posi i e, we i s p o e he ollowing lemma:
LEMMA 1. All he elemen s o ma ix Mpa e posi i e and less han uni y, and he sum o elemen s
in each ow o Mpequals 1.
To p o e he lemma abo e, we show ha Mpand Ωsha e he same eigen ec o s. Recall
13

ha Mpis he solu ion o he quad a ic ma ix equa ion: M2
p−(Ω−1/β +I)Mp+I/β =0N.
Le
Ξ = 
Ξ11 Ξ12
Ξ21 Ξ22
=
Ω−1/β +I−I/β
I0N

Le λbe an eigen alue o Ξ, hen he eigen ec o associa ed wi h i is he ec o X=
hX1X2i′, ha is,
(Ξ −λI)X= 0 ⇒(Ξ11 −λI)X1=X2/β, X1=λX2
Hence, he eigen ec o associa ed wi h λis X=hλX2X2i′. The e o e,
(Ξ11 −λI)X1−X2/β = 0 ⇐⇒ (Ω−1−(βλ −β+ 1/λ)
| {z }
e- alue o Ω−1
I)X2= 0
Uhlig (2001) shows ha he eigen ec o o Mpis gi en by X2.Ω−1and Ωsha e he same
eigen ec o s and, as a esul , i ollows ha Mpand Ωalso sha e he same eigen ec o s.
The la ges eigen alue o Ωis 1; he eigen ec o associa ed wi h i is e=1N. I ol-
lows ha eis also an eigen ec o o Mp, hence Mpe=e, implying ha he sum o each
ow o Mequals 1 and ha 1 is an eigen alue o Mp. To gua an ee a s able solu ion, i
has o be ha he emaining eigen alues o Mpa e wi hin he uni ci cle. By he Ge -
shgo in ci cle heo em, each eigen alue λio Mphas o be wi hin he ollowing ange
h1−PN
j=1 mp
ij −PN
j=1 |mp
ij|,1−PN
j=1 mp
ij +PN
j=1 |mp
ij|i. The bounds canno exceed 1 o
-1, implying ha PN
j=1 mp
ij =PN
j=1 |mp
ij|, and ha each elemen o Mpis posi i e.
I is easy o see ha MpMy=M2
pΩ−1K, whe e all he diagonal elemen s in Ka e pos-
i i e. Since Mpand Ω−1a e s ochas ic ma ices, i ollows ha M2
pΩ−1is also a s ochas ic
ma ix and ha all he elemen s in Ma e posi i e.
B.2.2 P oo o Co olla y 2
The esponse o he p ice ec o is gi en by ∂p
∂sj =βgM2
pΩ−1diag(κ)C:j. Absen com-
munica ion, he esponse is βgκj(M2
pΩ−1):,j; hence, p ice dispe sion depends on he dis-
pe sion o he elemen s in he j h column o M2
pΩ−1. F om he p oo o P oposi ion 1,
14
Mp−β∆ΩM2
p= ∆Ω; p e-mul iplying his exp ession by ∆−1, eigendecomposing Mpand
Ω, and using he ac ha Mpand Ωsha e he same eigen ec o s, we ha e ha
(I+βQΛΩQ−1)QΛpQ−1−βQΛΩΛ2
pQ−1=QΛΩQ−1⇒βΛΩΛ2
p−(I+βΛΩ)Λp−ΛΩ=0N,N ,
whe e Qis he ma ix con aining he eigen ec o s; ΛΩand ΛΩa e he diagonal ma ices
con aining he eigen alues o Mpand Ω, espec i ely. Pinning down he eigen alues o
Mpis equi alen o sol ing o he oo s o Nquad a ic polynomials. I is easy o see ha
0 is an eigen alue o Mponly i 0is also an eigen alue o Ω, which canno happen since
Ωis in e ible. As a esul , 0is no an eigen alue o M2
pΩ−1, hus he elemen s in any j h
column o M2
pΩ−1a e di e en om one ano he . The e o e, absen communica ion he e
is always dispe sion in he ini ial esponse o p ices o he in o ma ion ea men .
The esponse o he p ice ec o unde e en communica ion is gi en by
∂p
∂sj
=βg hPj(M2
pΩ−1)1jκj... Pj(M2
pΩ−1)Njκji′,
whe e Pj(M2
pΩ−1)ij = 1,∀i. F om abo e, we know ha he elemen s in any column o
M2
pΩ−1a e dis inc om one ano he . Hence, any dispe sion in he ini ial esponse o he
p ice ec o has o be due o he e ogeneous Phillips cu e slopes. I all i ms sha e he
same Phillips cu e slope κ, hen he ini ial esponse o any p ice would be βgκ.
B.2.3 P oo o P oposi ion 2
F om P oposi ion 1in he pape , he ec o o p ices con e ges o
lim
h→∞ p +h=βgjlim
h→∞ Mh+1
pMyC:,j =βgjlim
h→∞ Mh+2
pΩ−1diag(κ)C:,j,(18)
whe e he second equali y ollows om he ac ha MpMy=M2
pΩ−1diag(κ). As shown
in he p oo o P oposi ion 1,Mpand Ω−1sha e he same eigen ec o s, and he absolu e
alue o all he eigen alues o Mp, o he han he uni one, lie wi hin he uni ci cle. Hence,
15
he eigendecomposi ion o Mh+2
pΩ−1as happ oaches ∞is gi en by
lim
h→∞ Mh+2
pΩ−1=Q
10
0 0
Q−1=hQ11Q−1
1: Q12Q−1
1: ... Q1NQ−1
1: i′
The eigen ec o o Mpassocia ed wi h he uni eigen alue is Q1: =1, whe eas Q−1
1: is
he eigen ec o o Ω′associa ed wi h i s uni eigen alue – since Mp,Ω, and Ω−1sha e he
same eigen ec o s. Rubbo (2023) p o es in he Appendix ha Q1: =λ(I−Φ)Φ−1, so ha
Ω′Q1: =Q1:.
16
Online Appendix
July 10, 2025
C-1 Design De ails
C-1 Sample
Ou popula ion in he su ey is qui e ep esen a i e o he i ms in New Zealand. Panel
A in Table A-1 p esen s he o al numbe and pe cen age o i ms in manu ac u ing and
ade (wholesale and e ail). As pe S a is ics New Zealand, he e a e sligh ly o e 31,000
i ms in manu ac u ing, e ail, and wholesale ade. Mo e han hal he i ms in hese in-
dus ies a e small i ms employing ewe han six employees. The popula ion o i ms in
his su ey is d awn om hese indus ies, and we use employmen size dis ibu ion as
he benchma k o con ol o sample ep esen a ion. Ou su ey main ains ai ly simila
p opo ions o i ms a each i m size dis ibu ion, ha is, i ms wi h ewe han 5 em-
ployees, 6-19 employees, 20-49 employees, and a leas 50 employees. Fo example, 33
pe cen o manu ac u ing i ms in New Zealand employ be ween 6-19 wo ke s. In ou
su ey, he p opo ion o manu ac u e s in his employmen size g oup is a ound 34 pe -
cen . Compa ing Panel A and Panel B shows he p opo ions o i ms in each employmen
size dis ibu ion in ou su ey popula ion. O e all, he su ey popula ion ame includes
a ound 52 pe cen o i ms om he gene al popula ion o i ms in hese indus ies.
I is no uncommon in su eys o a ain a ying esponse a es om i ms ac oss di -
e en indus ies. One o he objec i es o he su ey was o achie e highe esponse a es
om i ms ha employ a leas six employees. Fi ms ha a e oo small in size a e qui e
ulne able and hei business con inui y is always ques ionable. Fu he mo e, du ing he
p ocess o de eloping he popula ion da a, i was e iden ha e y small i ms end o
change hei inpu supplie s qui e equen ly; his is no ideal o ou RCT exe cise. In his
su ey, he e o e, a lo o ocus was gi en o i ms ha employ a leas six employees. The
esponse a es o di e en employmen size g oups a e also epo ed in Panel C and D.
The o e all esponse a e is a ound 13 pe cen o he main wa e and a ound 50 pe cen
o he ollow-up wa e. Wi h he assis ance o su ey ec ui men specialis s, he su ey
e ained nea ly hal he i ms o pa icipa e in he ollow-up wa e.
The pa icipan s in he su ey a e manage s o di ec o s o he i m. One o he c i e ia
1

o pa icipan ec ui men was ha he manage o di ec o mus play an in eg al ole in
he i m in se ing p oduc p ices and wages, and also be an in luen ial igu e in in es -
men and employmen decisions. This c i e ion was applied o ec ui pa icipan s o he
popula ion da abase compiled by New Zealand Ma ke Resea ch and Su eys Limi ed.
Pa icipan de ails and hei i m’s supply chain ela ionships a e egula ly upda ed o
keep he eco ds ac i e. The ime gap in pa icipa ion be ween he main wa e and he
ollow-up wa e was app oxima ely h ee mon hs. Pa icipan s om all RCT g oups we e
su eyed h oughou he su ey pe iod; ha is, he e was no case whe e a pa icula RCT
g oup was p io i ized in ime.
C-2 Powe Calcula ions
Wi h he sample size o 150 (N1=75 ea ed and N2=75 con ol pai s), signi icance (α) equal
o 5% and powe (1−κ) equal o 80%, he minimum de ec able e ec (MDE) is 0.46SD.
When we a y he sample size o 200, MDE = 0.398SD.
Based on a pilo we collec ed in o ma ion o 20 pai s o i ms: 10 ea ed and 10 con-
ol. We a e in e es ed in ne wo k e ec s so we p o ide he powe analysis o un a ge ed
ea ed i ms. The es ima ed e ec size o ea men on he un a ge ed i m’s mean GDP
expec a ions in he ollow-up was 1.39, signi ican a he 5% le el.
We epea he same es ima ion o he e ec s on economic decisions o he un a ge ed
i ms. The e ec size o p ices, in es men , and employmen a e 3.39, 1.45, and 3.24,
espec i ely. Fo wages, we de ec ze o e ec in he pilo . We summa ize his in o ma ion
in he able below:
2
Table B-1: Powe Calcula ion
Pilo Es ima ed E ec Size Minimum De ec able E ec
N=150 N=200
GDP mean o ecas 1.39
0.460 0.398
P ices 3.39
Employmen 1.45
In es men 3.24
Wages -
No es: The a iables all co espond o he ollow-up wa e o he linked i m. The e ec sizes a e in uni s o
s anda d de ia ion.
C-2 Su ey
C-1 P e-Su ey In o ma ion
The su ey company (New Zealand Ma ke Resea ch and Su eys Limi ed) p o ided
some i m cha ac e is ics ha hey collec ed independen ly o , and mon hs be o e, ou
su ey. These include employmen , in en o y sha e om main supplie , numbe o cus-
ome s, and numbe o supplie s.
A ew days be o e he baseline o ou su ey, he su ey company e i ied supplie
and cus ome iden i ica ion. Speci ically:
Ask his ques ion o cus ome /main supplie i m: You i m is lis ed in he da abase a
New Zealand Ma ke Resea ch and Su eys Limi ed. The da abase indica es ha XXX
[ i m name] is you cus ome /main supplie o he main p oduc line. Is his in o ma-
ion co ec ?
1. Yes
2. No
3
C-2 Baseline Su ey
Sec ion A. Fi m Cha ac e is ics
1How many yea s old is he i m?
Answe : yea s
2How many wo ke s a e employed in his i m?
Answe : wo ke s
3Ou o he o al e enue o he i m, wha ac ion is used o compensa ion o all
employees and wha ac ion is used o he cos s o ma e ials and in e media e
inpu s ( aw ma e ials, ene gy inpu s, e c. .. )?
Sha e o e enues:
Labo cos % , Cos o ma e ials %
4Fo i s main p oduc line, wha is he i m’s cu en ma ke sha e?
Answe : %
5How many weeks ago did you i m change he p ice o he main p oduc ?
Answe : Weeks ago.
6Using he ollowing equencies, please iden i y how o en his i m ( o mally)
changes he p ice o i s main p oduc :
(a) Daily
(b) Weekly
(c) Mon hly
(d) Qua e ly
4
(e) Hal annually
( ) Annually
(g) Less equen ly han annually
Sec ion B. Manage Cha ac e is ics
7How many yea s o wo k expe ience do you ha e a his i m: Answe :
yea s.
8Wha is you highes educa ional quali ica ion?
(a) Less han high school
(b) High school diploma
(c) Some college o Associa e deg ee
(d) College Diploma
(e) G adua e S udies (Mas e s o PhD)
Sec ion C. Mac oeconomic Expec a ions
9Wha do you hink will be he annual g ow h a e o eal GDP in New Zealand in
wel e mon hs?
Answe : % pe yea .
10 Could you p o ide us wi h an app oxima e ange o wha you hink annualized
eal GDP g ow h in New Zealand will be o e he nex 12 mon hs?
Be ween % pe yea (lowes o ecas ) and % pe yea (highes
o ecas ).
5
C-3 Resul s wi h Hube Weigh s
Table C-1: T ea men E ec on Expec ed GDP Unce ain y in Baseline and Follow-up wi h
Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P io mean
i0.972*** 0.977*** 0.964*** 0.964*** 0.945*** 0.957*** 0.938*** 0.934***
(0.008) (0.005) (0.016) (0.016) (0.020) (0.016) (0.013) (0.010)
T11.799*** 1.825*** -0.063 -0.062 1.787*** 1.849*** 1.772*** 1.948***
(0.045) (0.030) (0.044) (0.044) (0.070) (0.050) (0.112) (0.063)
T21.567*** 1.362*** -0.040 -0.039 1.773*** 1.821*** 1.433*** 1.681***
(0.074) (0.039) (0.045) (0.045) (0.095) (0.065) (0.147) (0.071)
T1×P io mean
i-0.723*** -0.749*** 0.017 0.017 -0.603*** -0.633*** -0.586*** -0.657***
(0.022) (0.015) (0.019) (0.019) (0.032) (0.024) (0.046) (0.027)
T2×P io mean
i-0.492*** -0.378*** 0.006 0.006 -0.503*** -0.508*** -0.502*** -0.604***
(0.039) (0.020) (0.018) (0.018) (0.046) (0.033) (0.061) (0.032)
Cons an 0.025 0.013 0.062 0.061 0.080* 0.038 0.120*** 0.101***
(0.024) (0.008) (0.043) (0.043) (0.047) (0.029) (0.036) (0.028)
Reg ession OLS Hube OLS Hube OLS Hube OLS Hube
Pe iod Baseline Baseline Baseline Baseline Follow Up Follow Up Follow Up Follow Up
Type T ea ed T ea ed Connec ed Connec ed T ea ed T ea ed Connec ed Connec ed
Obse a ions 999 956 1,020 1,020 510 507 505 494
R-squa ed 0.739 0.920 0.955 0.956 0.760 0.851 0.743 0.871
No e. The able epo s esul s o eg ession 1, whe e he ou come a iables Pos e io mean
iis he a e age
GDP o ecas o i m ia e he ea men . P io mean
iis he a e age GDP o ecas be o e he ea men .
T1is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou he a e age GDP
o ecas and T2is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou GDP
unce ain y. Columns (1), (2), (3) and (4) show esul s o he baseline su ey, and columns (5), (6), (7)
and (8) show esul s o he ollow-up su ey. Columns (1), (2), (5) and (6) show esul s o he i ms ha
ecei ed he in o ma ion ea men in he baseline pe iod, and columns (3), (4), (7) and (8) show esul s
o he i ms ha a e connec ed o he ea ed i ms. Columns (2), (4), (6) and (8) show esul s wi h Hube
weigh s. Robus s anda d e o s a e shown in pa en heses.
12

Table C-2: T ea men E ec on Expec ed GDP Unce ain y in Baseline and Follow-up wi h
Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P io unce ain y
i0.960*** 0.997*** 0.993*** 0.993*** 0.978*** 0.989*** 0.974*** 0.994***
(0.019) (0.002) (0.010) (0.010) (0.019) (0.009) (0.018) (0.012)
T11 1.395*** 1.157*** 0.025 0.025 1.310*** 0.958*** 2.044*** 1.953***
(0.198) (0.072) (0.084) (0.084) (0.302) (0.184) (0.328) (0.216)
T21.145*** 1.414*** -0.015 -0.015 1.142*** 1.368*** 1.139*** 1.440***
(0.163) (0.065) (0.083) (0.083) (0.264) (0.132) (0.267) (0.154)
T1×P io unce ain y
i-0.766*** -0.753*** -0.008 -0.008 -0.717*** -0.684*** -0.761*** -0.767***
(0.033) (0.012) (0.013) (0.013) (0.042) (0.025) (0.046) (0.031)
T2×P io unce ain y
i-0.720*** -0.801*** -0.008 -0.008 -0.689*** -0.736*** -0.610*** -0.663***
(0.031) (0.012) (0.014) (0.014) (0.042) (0.020) (0.045) (0.025)
Cons an 0.220** 0.020 0.067 0.067 0.187** 0.117** 0.276** 0.100
(0.095) (0.017) (0.070) (0.070) (0.090) (0.050) (0.122) (0.063)
Pe iod Pos e io Baseline Baseline Baseline Baseline Follow Up Follow Up Follow Up Follow Up
Type o i m T ea ed T ea ed Connec ed Connec ed T ea ed T ea ed Connec ed Connec ed
Obse a ions 1,012 961 1,022 1,022 514 506 513 504
R-squa ed 0.835 0.965 0.973 0.973 0.809 0.910 0.700 0.856
No e. The able epo s esul s o eg ession 1, whe e he ou come a iablePos e io unce ain y
iis he un-
ce ain y on he GDP o ecas o i m ia e he ea men , measu ed as he absolu e alue o he dis ance
be ween he mos and leas likely scena io. P io unce ain y
iis he unce ain y o ecas be o e he ea men .
T1is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou he a e age GDP
o ecas and T2is an indica o ha is equal o one i i m i ecei ed he in o ma ion ea men abou GDP
unce ain y. Columns (1), (2), (3) and (4) show esul s o he baseline su ey, and columns (5), (6), (7) and
(8) show esul s o he ollow-up su ey. Columns (1), (2), (5) and (6) show esul s o he i ms ha e-
cei ed he in o ma ion ea men in he baseline pe iod, and columns (3), (4), (7) and (8) show esul s o he
i ms ha a e connec ed o he ea ed i ms. Columns (2), (4), (6) and (8) show esul s wi h Hube weigh s.
Robus s anda d e o s a e shown in pa en heses.
13
Table C-3: Causal E ec o GDP Fo ecas and Unce ain y on Ac ions wi h Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P ice P ice In In Emp Emp Wage Wage
Pos e io mean
i0.292*** 0.232*** 0.138 0.158 0.868*** 0.848*** 0.003 -0.002
(0.082) (0.081) (0.141) (0.132) (0.295) (0.278) (0.013) (0.010)
Pos e io unce ain y
i-0.369*** -0.388*** -0.805*** -0.808*** -0.834*** -0.882*** 0.005 0.004
(0.031) (0.031) (0.058) (0.053) (0.121) (0.114) (0.007) (0.006)
Plans 0.741*** 0.754*** 0.534*** 0.549*** 0.519*** 0.574*** 0.990*** 0.991***
(0.027) (0.027) (0.038) (0.038) (0.066) (0.061) (0.007) (0.007)
P io mean
i-0.144** -0.104 -0.026 -0.042 -0.603*** -0.629*** -0.003 0.002
(0.067) (0.067) (0.107) (0.104) (0.229) (0.223) (0.012) (0.008)
P io unce ain y
i0.268*** 0.291*** 0.594*** 0.612*** 0.685*** 0.710*** -0.004 -0.003
(0.029) (0.030) (0.053) (0.050) (0.120) (0.115) (0.004) (0.004)
Cons an 0.634*** 0.615*** 1.452*** 1.292*** 0.194 0.451 0.013 0.012
(0.130) (0.123) (0.227) (0.217) (0.471) (0.427) (0.020) (0.021)
Type All All All All All All All All
Reg ession OLS Hube OLS Hube OLS Hube OLS Hube
F (mean) 143 379.7 169.2 369.4 140.8 362.8 138.3 363.6
F (unce ) 592.4 1406 740.8 1433 622.5 1483 599.4 1418
Obse a ions 960 940 959 939 960 938 958 937
R-squa ed 0.639 0.665 0.480 0.507 0.272 0.323 0.981 0.982
No e. The able epo s esul s o eg ession 3, whe e he ou come a iables a e ac ions ha he i m ook in
he h ee mon hs be o e he ollow-up su ey. Those ac ions a e he change in p ices (columns (1) and (2)),
change in in es men (columns (3) and (4)), change in employmen (columns (5) and (6)) and change in wages
(columns (7) and (8)). Columns (2), (4), (6) and (8) use Hube weigh s. P lan a e he plans ha he i m had in he
baseline su ey o he nex h ee mon hs. P os e io mean
iis he GDP o ecas o he i m in he ollow-up pe iod.
Pos e io unce ain y
iis he unce ain y abou he GDP o ecas o i m iin he ollow-up pe iod, measu ed as he
absolu e alue on he dis ance be ween he mos and leas likely scena io. P os e io mean
iis he GDP o ecas
o he i m in he baseline pe iod be o e ecei ing he ea men and P io unce ain y
iis he unce ain y o ecas
be o e he ea men . We ins umen he pos e io a iables wi h he p io s in e ac ed by he ea men dummy
and a ea men dummy. Robus s anda d e o s a e shown in pa en heses.
14
Table C-4: Causal E ec o GDP Fo ecas and Unce ain y on Ac ions, o T ea ed Fi ms wi h
Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P ice P ice In In Emp Emp Wage Wage
Pos e io mean
i0.163 0.099 0.008 0.039 0.912** 0.732* 0.024 0.006
(0.114) (0.128) (0.224) (0.218) (0.419) (0.431) (0.026) (0.014)
Pos e io unce ain y
i-0.335*** -0.357*** -0.824*** -0.822*** -0.810*** -0.906*** 0.005 -0.000
(0.042) (0.047) (0.083) (0.081) (0.173) (0.173) (0.010) (0.008)
Plans 0.766*** 0.759*** 0.486*** 0.521*** 0.379*** 0.443*** 0.995*** 0.999***
(0.037) (0.038) (0.057) (0.055) (0.104) (0.100) (0.008) (0.007)
P io mean
i-0.061 -0.022 0.015 -0.025 -0.746** -0.642* -0.016 0.002
(0.093) (0.105) (0.157) (0.156) (0.325) (0.341) (0.022) (0.010)
P io unce ain y
i0.258*** 0.282*** 0.571*** 0.586*** 0.631*** 0.725*** -0.009 -0.005
(0.039) (0.043) (0.073) (0.068) (0.161) (0.157) (0.007) (0.005)
Cons an 0.541*** 0.561*** 1.742*** 1.600*** 0.444 0.591 0.008 0.019
(0.168) (0.172) (0.355) (0.344) (0.766) (0.717) (0.031) (0.030)
Type T ea ed T ea ed T ea ed T ea ed T ea ed T ea ed T ea ed T ea ed
Reg ession OLS Hube OLS Hube OLS Hube OLS Hube
F (mean) 110.8 233.2 151.8 228.8 118.9 235.6 109.3 233.7
F (unce ) 365.3 1132 777.3 1191 402 1229 386.8 1142
Obse a ions 485 476 478 471 479 470 479 470
R-squa ed 0.688 0.689 0.448 0.480 0.174 0.229 0.978 0.983
No e. The able epo s esul s o eg ession 3only o ea ed i ms, whe e he ou come a iables a e ac ions
ha he i m ook in he h ee mon hs be o e he ollow-up su ey. Those ac ions a e he change in p ices
(columns (1) and (2)), change in in es men (columns (3) and (4)), change in employmen (columns (5) and (6))
and change in wages (columns (7) and (8)). Columns (2), (4), (6) and (8) use Hube weigh s. Plan a e he plans
ha he i m had in he baseline su ey o he nex h ee mon hs. P os e io mean
iis he GDP o ecas o he i m
in he ollow-up pe iod. Pos e io unce ain y
iis he unce ain y in he GDP o ecas o i m iin he ollow-up pe-
iod, measu ed as he absolu e alue o he dis ance be ween he mos and leas likely scena io. P os e io mean
i
is he GDP o ecas o he i m in he baseline pe iod be o e ecei ing he ea men and P io unce ain y
iis he
unce ain y o ecas be o e he ea men . We ins umen he pos e io a iables wi h he p io s in e ac ed by
he ea men dummy and a ea men dummy. Robus s anda d e o s a e shown in pa en heses.
15
Table C-5: Causal E ec o GDP Fo ecas and Unce ain y on Ac ions, o Connec ed Fi ms
wi h Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P ice P ice In In Emp Emp Wage Wage
Pos e io mean
i0.412*** 0.368*** 0.161 0.188 0.558 0.722** -0.015 -0.013
(0.116) (0.106) (0.176) (0.165) (0.365) (0.328) (0.012) (0.011)
Pos e io unce ain y
i-0.425*** -0.443*** -0.842*** -0.844*** -0.902*** -0.912*** 0.007 0.008
(0.046) (0.041) (0.075) (0.071) (0.176) (0.165) (0.010) (0.010)
Plans 0.712*** 0.742*** 0.583*** 0.585*** 0.644*** 0.693*** 0.986*** 0.984***
(0.041) (0.039) (0.048) (0.048) (0.083) (0.083) (0.012) (0.013)
P io imean -0.226** -0.191** -0.011 -0.019 -0.326 -0.476* 0.008 0.008
(0.094) (0.087) (0.140) (0.136) (0.294) (0.275) (0.010) (0.010)
P io unce ain y
i0.302*** 0.321*** 0.663*** 0.679*** 0.787*** 0.768*** 0.001 0.000
(0.042) (0.041) (0.070) (0.071) (0.176) (0.175) (0.005) (0.005)
Cons an 0.717*** 0.691*** 1.257*** 1.083*** 0.148 0.354 0.009 0.006
(0.200) (0.187) (0.307) (0.295) (0.533) (0.490) (0.028) (0.029)
Type Conn Conn Conn Conn Conn Conn Conn Conn
Reg ession OLS Hube OLS Hube OLS Hube OLS Hube
F (mean) 69.73 164.1 70.15 171.1 70.32 157.6 66.85 158.8
F (unce ) 247.7 487.3 233.1 492.7 260.5 517.4 249.2 495.4
Obse a ions 475 459 481 463 481 463 479 462
R-squa ed 0.601 0.641 0.523 0.548 0.403 0.452 0.983 0.982
No e. The able epo s esul s o eg ession 3only o connec ed i ms, whe e he ou come a iables a e ac-
ions ha he i m ook in he h ee mon hs be o e he ollow-up su ey. Those ac ions a e he change in p ices
(columns (1) and (2)), change in in es men (columns (3) and (4)), change in employmen (columns (5) and (6))
and change in wages (columns (7) and (8)). Columns (2), (4), (6) and (8) use Hube weigh s. Plan a e he plans
ha he i m had in he baseline su ey o he nex h ee mon hs. P os e io mean
iis he GDP o ecas o he i m
in he ollow-up pe iod. Pos e io unce ain y
iis he unce ain y in he GDP o ecas o i m iin he ollow-up pe-
iod, measu ed as he absolu e alue o he dis ance be ween he mos and leas likely scena io. P os e io mean
i
is he GDP o ecas o he i m in he baseline pe iod be o e ecei ing he ea men and P io unce ain y
iis he
unce ain y o ecas be o e he ea men . We ins umen he pos e io a iables wi h he p io s in e ac ed by
he ea men dummy and a ea men dummy. Robus s anda d e o s a e shown in pa en heses.
16
Table C-6: Causal E ec s o GDP Fo ecas , Unce ain y and O he s’ Ac ions on Connec ed
Fi ms wi h Hube Weigh s
(1) (2) (3) (4) (5) (6) (7) (8)
P ice P ice In In Emp Emp Wage Wage
Pos e io mean
i0.419*** 0.389*** 0.065 0.087 0.644* 0.826** -0.019 -0.016
(0.125) (0.110) (0.170) (0.164) (0.386) (0.354) (0.015) (0.016)
Pos e io unce ain y
i-0.331*** -0.360*** -0.515*** -0.516*** -0.779*** -0.747*** 0.007 0.009
(0.071) (0.072) (0.103) (0.106) (0.217) (0.209) (0.011) (0.012)
Ac ion O he j−i0.236* 0.199 0.317*** 0.337*** 0.091 0.111 0.348 0.583
(0.139) (0.144) (0.083) (0.088) (0.083) (0.082) (0.323) (0.521)
Plani0.708*** 0.736*** 0.564*** 0.569*** 0.643*** 0.694*** 0.981*** 0.978***
(0.042) (0.040) (0.051) (0.050) (0.086) (0.086) (0.014) (0.016)
Plan Ac ion O he j−i-0.103 -0.079 -0.192*** -0.195** -0.047 -0.048 -0.346 -0.581
(0.139) (0.145) (0.072) (0.080) (0.071) (0.074) (0.323) (0.519)
P io mean
i-0.230** -0.202** 0.023 0.018 -0.427 -0.594** 0.013 0.010
(0.099) (0.088) (0.138) (0.136) (0.314) (0.302) (0.012) (0.014)
P io unce ain y
i0.230*** 0.257*** 0.492*** 0.490*** 0.673*** 0.617*** 0.002 0.001
(0.063) (0.067) (0.086) (0.095) (0.209) (0.211) (0.006) (0.007)
Cons an 0.486** 0.494*** 0.331 0.278 0.079 0.316 0.007 0.003
(0.194) (0.176) (0.314) (0.283) (0.554) (0.521) (0.028) (0.029)
Reg ession OLS Hube OLS Hube OLS Hube OLS Hube
F (mean) 50.68 114.8 48.08 119.7 60.13 116.1 43.98 110
F (unce ) 187.8 340.8 158.7 341.8 191 364.7 191.9 358.1
F (Ac ion T) 45.47 41.95 64.57 61.14 16.08 17.74 0.851 0.629
Obse a ions 453 438 452 435 454 437 452 436
R-squa ed 0.610 0.654 0.490 0.502 0.388 0.435 0.979 0.975
No e: The able epo s esul s o eg ession 3, whe e he ou come a iables a e ac ions ha he i m ook in he
h ee mon hs be o e he ollow-up su ey. Those ac ions a e he change in p ices (columns (1) ans (2)), change in
in es men (columns (3) and (4)), change in employmen (columns (5) and (6)) and change in wages (columns (7)
and (8)). Plania e he plans ha he i m had in he baseline su ey o he nex h ee mon hs. Ac ion O he j−i
a e he ac ions o he di ec ly ea ed i m ha is connec ed o a i m in his sample. Plan Ac ion O he j−ia e
hei plans. Pos e io mean
iis he GDP o ecas o he i m in he ollow-up pe iod. Pos e io unce ain y
iis he
unce ain y in he GDP o ecas o i m iin he ollow-up pe iod, measu ed as he absolu e alue o he dis ance
be ween he mos and leas likely scena io. P os e io mean
iis he GDP o ecas o he i m in he baseline pe iod
be o e ecei ing he ea men and P io unce ain y
iis he unce ain y o ecas be o e he ea men . This eg ession
is un o he connec ed i m. We ins umen he pos e io a iables and Ac ion O he j−iwi h he ea men
dummy, p io s in e ac ed by he ea men dummy and he plan o he di ec ly ea ed i m ha is connec ed o a
i m in his sample. Robus s anda d e o s a e shown in pa en heses.
C-4 Addi ional Resul s om he Model
Wi hin-ne wo k ini ial p ice esponse a ia ion. Figu e C-1 shows he his og am o he
wi hin-ne wo k a iance o he ini ial p ice esponses o a ea men o highe unce ain y.
17

Figu e C-1: Dis ibu ion o Wi hin-Ne wo k P ice Va iance a e T ea men o Highe Un-
ce ain y
T ea men o supplie - m
0 0.02 0.04 0.06 0.08 0.1
Va iance o p ice esponses
0
5
10
15
20
25 T ea men o cus ome - m
0 0.01 0.02 0.03 0.04
Va iance o p ice esponses
0
5
10
15
20
25
No communica ion
E en communica ion
No e: Dis ibu ion o he a iance o ini ial p ice esponses ac oss ne wo ks when he ea ed i m is he
supplie (le panel) and when he ea ed i m is he cus ome ( igh panel). In ed: e en communica ion;
in blue: no communica ion.
Es ima es o he ela ionship be ween he p ice o he ea ed and ha o he con-
nec ed i m. We es ima e he change in he p ice o he connec ed i m ha is associa ed
wi h a 1 pe cen age poin change in he p ice o he main ( ea ed) i m in he model sim-
ula ions and su ey da a. In he su ey da a, we ins umen he p ice change o he main
i m wi h in e ac ions be ween he ea men dummy and he p ice plan ha he i m had.
In pa icula , we es ima e β1and β2in he ollowing eg ession
P iceconnec ed
j−i=β1(P ice ea ed
i|Talk GDPj−i) + β2(P ice ea ed
i|No Talk GDPj−i) + Xi δ′+εi,
while ins umen ing bo h eg esso s by i m i’s ea men indica o , he in e ac ion be-
ween i m i’s ea men indica o and i s planned p ice change, and he in e ac ion be-
ween i m i’s planned p ice change and an indica o o whe he he pai o i ms commu-
nica ed abou GDP. Vec o Xi embeds i m iand j’s planned p ice changes.
Impo an ly, we sepa a e he e ec be ween he i ms ha communica ed a leas once
abou GDP and he i ms ha did no . While communica ion is also a ec ed by he ea -
18
Table C-7: P ice Pass- h ough om T ea ed o Connec ed wi h Di e en Le els o Com-
munica ion
(1) (2) (3) (4)
P ice ea ed|Talk GDP 0.816*** 0.810***
(0.158) (0.224)
P ice ea ed|No Talk GDP 0.261 0.631***
(0.179) (0.205)
P ice ea ed|E en Comm 1.020*** 0.980***
(0.004) (0.004)
P ice ea ed|No Comm 0.348*** 0.508***
(0.005) (0.009)
Type Supplie Supplie Cus ome Cus ome
Shock o Cus ome Cus ome Supplie Supplie
Da a Su ey Model Su ey Model
F (Talk) 17.89 15.34
F (No Talk) 32.93 112
Obse a ions 181 180 164 180
R-squa ed 0.513 0.820 0.441 0.662
No e: The able epo s esul s o a eg ession ha es ima es he empi ical p ice pass- h ough om he
ea ed i m o he connec ed i m (columns (1) and (3)), and he equi alen eg ession in he model (columns
(2) and (4)). We ins umen bo h P ice ea ed|Talk GDP and P ice ea ed|No Talk GDP wi h he p ice
change plans o he ea ed i m, in e ac ed by he ea men , he ea men indica o and he in e ac ion
wi h whe he hey alked o no . We con ol o he ea ed and connec ed i ms’ plans. Robus s anda d
e o s in all eg essions.
men , po en ially biasing hese es ima es, we see hese esul s as sugges i e e idence o
he model mechanisms.
Al e na i e communica ion ma ix. Figu e C-2 shows ha in he case o C= Ω, he
p ice esponses a e ampli ied ela i e o he no communica ion scena io, bu no by as
much as hey a e unde e en communica ion. Simila o he case o no communica ion, Ω
communica ion implies ha he p ice esponses a e bigge when he ea men o igina es
om he supplie i m han when hey come om he cus ome i m.
19
Figu e C-2: Dis ibu ion o he Impac on P ices a e T ea men s o Highe Unce ain y
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Pe cen
0
1
2
3
4
5
6
7T ea men o supplie - m
-0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1
Pe cen
0
1
2
3
4
5
6
7T ea men o cus ome - m
No communica ion
E en communica ion
+communica ion
No e: Dis ibu ion o p ice changes ac oss all h ee i ms when he ea ed i m is he supplie (le panel)
and when he ea ed i m is he cus ome ( igh panel). In ed: e en communica ion; in blue: no communi-
ca ion; in black: Ωcommunica ion.
20