Boeck, Maximilian; Mo i, Lo enzo
Wo king Pape
Has globaliza ion changed he in e na ional ansmission
o U.S. mone a y policy?
Wo king Pape , No. 15/2023
P o ided in Coope a ion wi h:
No ges Bank, Oslo
Sugges ed Ci a ion: Boeck, Maximilian; Mo i, Lo enzo (2024) : Has globaliza ion changed
he in e na ional ansmission o U.S. mone a y policy?, Wo king Pape , No. 15/2023, ISBN
978-82-8379-305-5, No ges Bank, Oslo,
h ps://hdl.handle.ne /11250/3166793
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/310389
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
Wo king Pape
Has Globaliza ion Changed he In e na ional T ansmission
o
U.S. Mone a y Policy?
No ges Bank Resea ch
Au ho s
:
Maximilian Boeck
Lo enzo Mo
i
K
eywo ds:
Mone a y
policy, in e na ional
Spillo e s, TVP-VARs.
15 | 2023
No ges Bank Wo king Pape
Wo king pape s a No ges Bank, a 1992/1 il 2009/2 kan bes illes på e-pos : se icesen e @no ges-
bank.no
F a 1999 og sene e e publikasjonene ilgjengelige på www.no ges-bank.no
Wo king pape s inneholde o skningsa beide og u edninge som anlig is ikke ha å sin endelige o m.
Hensik en e blan anne a o a e en kan mo a kommen a e a kollege og and e in e esse e. Synspunk e
og konklusjone i a beidene s å o o a e nes egning.
Wo king pape s om No ges Bank, om 1992/1 o 2009/2 can be o de ed by e-mail:
se icesen e @no ges-bank.no
Wo king pape s om 1999 onwa ds a e a ailable on www.no ges-bank.no
No ges Bank’s wo king pape s p esen esea ch p ojec s and epo s (no usually in hei inal o m) and a e
in ended in e alia o enable he au ho o bene i om he commen s o colleagues and o he in e es ed
pa ies. Views and conclusions exp essed in wo king pape s a e he esponsibili y o he au ho s alone.
ISSN 1502-8143 (online)
ISBN 978-82-8379-305-5 (online)
Has Globaliza ion Changed he In e na ional T ansmission o
U.S. Mone a y Policy?∗
Maximilian Boeck†Lo enzo Mo i‡
Decembe 2023
Abs ac
We es ima e a ime- a ying pa ame e ec o au o eg ession o examine he e olu ion o in-
e na ional spillo e s o U.S. mone a y policy in ligh o inc easing globaliza ion in eal and
inancial ma ke s. We ind ha he ad e se in e na ional e ec s o a U.S. igh ening ha e sub-
s an ially inc eased o e he pas h ee decades, peaking du ing he G ea Recession. Based
on a c oss-coun y analysis and coun e ac ual simula ions, we a gue ha such ampli ica ion
can p ima ily be a ibu ed o he su ge in ade in eg a ion, while he ole o ising inancial
in eg a ion in explaining he ime- a ia ion is limi ed.
Keywo ds: Mone a y Policy; In e na ional Spillo e s; TVP-VARs.
JEL Codes: C32, E32, E52, F42, F62.
∗This wo king pape should no be epo ed as ep esen ing he iews o No ges Bank. The iews exp essed a e
hose o he au ho s and do no necessa ily e lec hose o No ges Bank. A p e ious e sion o his pape ci cula ed
unde he i le “U.S. Mone a y Policy and Globaliza ion: A Time-Va ying Pe spec i e”. We hank, wi hou implying
endo semen , Knu A e Aas ei , Juan An olin-Diaz, D ago Be ghol , Alessand o Bucciol, Gio anni Caggiano, Fabio
Cano a, Jin Cao, E em Cas elnuo o, Luca De Angelis, Timo Dimi iadis, Alessand o Do is, Zeno Ende s, F ancesco
Fu lane o, F ancesco Fusa i, Luca Gambe i, Geo gios Geo giadis, Ch is ian Glocke , Eleono a G anzie a, Pia Heckl,
Ma ie Hoe o a, Flo ian Hube , Ragna Juels ud, Thomas A. Lubik, Massimiliano Ma cellino, Ral R. Meisenzahl,
Ge no Mülle , Pascal Paul, Ge Pee sman, Gio anni Pelleg ino, Gio anni Ricco, Thomas Zö ne , and pa icipan s a
a ious esea ch semina s and con e ences o hei aluable commen s. Global indus ial p oduc ion da a o he p ojec
we e kindly p o ided by Vale ie G ossman, En ique Ma ínez-Ga cía, and B aden S ackman. Pa o his esea ch was
conduc ed du ing Boeck’s isi ing pe iod a he Uni e si y o Tübingen and Mo i’s isi ing pe iod a Ghen Uni e si y,
whose kind hospi ali y is g a e ully acknowledged. Maximilian Boeck g a e ully acknowledges inancial suppo by he
Eu opean Union - Nex Gene a ionEU, in he amewo k o he GRINS - G owing Resilien , INclusi e and Sus ainable
p ojec (GRINS PE00000018 – CUP B43C22000760006). The iews and opinions exp essed a e solely hose o he
au ho s and do no necessa ily e lec hose o he Eu opean Union, no can he Eu opean Union be held esponsible o
hem.
†Uni e si à Bocconi. E-mail: [email p o ec ed].
‡Uni e si y o Pado a and No ges Bank. E-mail: [email p o ec ed].
1
1. In oduc ion
Two well-known s ylized ac s in in e na ional economics – po ayed in Figu e 1 – ead as ollows.
Fi s , he global mac o- inancial sys em is cen e ed a ound he U.S. dolla , which ep esen s he
dominan cu ency in ade in oicing and issuance o inancial asse s (Gopina h, 2015; Rey, 2016).
This implies ha U.S. hikes a ec global ou comes by dep essing global ade and inancial con-
di ions.1Second, globaliza ion led o a massi e inc ease in global ade and inancial in eg a ion,
which could ha e subs an ially modi ied he global ansmission o U.S. shocks. Mo i a ed by he
in e ac ion o hese wo ac s, his pape examines he implica ions o globaliza ion o he in e -
na ional ansmission o U.S. mone a y policy. We es ima e a p oxy-SVAR model ha allows o
ime- a ying pa ame e s o cap u e he possible change in he impac o policy dis u bances on he
global inancial and eal cycles. We do so because – as shown once again in Figu e 1 – globaliza ion
has ma e ialized a a changing pace o e ime, wi h a change in di ec ion a e he G ea Recession,
a phenomenon known as “slowbaliza ion”.
We documen ha globaliza ion has led o subs an ial ime- a ia ion in he in e na ional ami-
ica ions o policy shocks: U.S. policy hikes gene a e s onge ecessiona y e ec s o e ime, wi h
a la ening ou o e ec s a e he G ea Recession. Whe eas he e is as e idence showing ha
a mone a y hike enginee ed by he Fede al Rese e (Fed) gene a es global ecessiona y e ec s
(e.g., Dedola e al., 2017; Iaco iello and Na a o, 2019; Degaspe i e al., 2020; Geo giadis and
Schumann, 2021; B ei enlechne e al., 2022; Kalemli-Özcan and Unsal, 2023; and B äuning and
She emi o , 2023), ou esul s on he ele ance o he ime- a ia ion is no el.
To s udy he e ec s o exogenous U.S. mone a y policy shocks and hei in e na ional spillo e s
o e ime, we use a ime- a ying pa ame e ec o au o eg ession (TVP-VAR). We analyze he
ansmission o U.S. mone a y policy media ed by he eac ion o U.S. indus ial p oduc ion, U.S.
p ices, and global economic, ade, and inancial indica o s. To gauge he e ec s on global ou comes,
we ely on he global (excluding U.S.) p oduc ion and ade indices and in e na ional asse p ices.
We es ima e ou ime- a ying model using Bayesian echniques (P imice i, 2005; Paul, 2020) and
achie e shocks’ iden i ica ion using he high- equency ins umen o Mi anda-Ag ippino and Ricco
(2021), which di ec ly con ols o he in o ma ion channel o mone a y policy.
The e ec s o U.S. mone a y policy a e ansmi ed globally h ough ade and inancial channels.
The o me e e s o he adi ional Mundell-Fleming e ec s, in which he cu en accoun adjus s
because mone a y policy a ec s bo h agg ega e demand (o o eign goods) and he cu ency. A
domes ic policy igh ening gene a es ecessiona y e ec s a home, which dep esses demand o
impo s ( es o he wo ld - RoW - expo ). This demand e ec should be pa ly coun e ac ed
1The pi o al ole o he dolla as an in e na ional cu ency b ough many obse e s o iew he Fede al Rese e as a
wo ld banke , see o ins ance Gou inchas and Rey (2007), Gopina h (2015), Rey (2016), Gopina h and S ein (2018),
Gou inchas e al. (2019) and Ilze zki e al. (2019).
2
Figu e 1: Mo i a ing E idence: Global T ade and Financial In eg a ion as Sha e o GDP.
No es: Global (dolla ) ade in eg a ion is de ined as he sum o global expo s and impo s (in oiced in dolla ) as a pe cen age o GDP (Wo ld
Bank na ional accoun s da a and Boz e al. (2022)). Global (dolla ) inancial in eg a ion is de ined as he sum o global ex e nal asse s and liabili ies
(denomina ed in dolla ) as a pe cen age o GDP (da a om he Ex e nal Weal h o Na ions da abase (Lane and Milesi-Fe e i (2018)) and Béné ix
e al. (2019)).
h ough a e alua ion o he cu ency, which leads o expendi u e-swi ching om home goods o
o eign goods. Howe e , since U.S. impo s a e mainly in oiced in dolla , a e alua ion o he
cu ency nei he boos s no dampens he compe i i eness o o eign goods; hus, he expendi u e-
swi ching channel is o mino impo ance o he U.S. S ill, he U.S. dolla is no only he dominan
in oicing cu ency o U.S. ade bu o global ade in gene al (Goldbe g and Tille, 2008; Gopina h,
2015; Gopina h e al., 2020; Boz e al., 2022). Hence, expendi u e-swi ching plays a subs an ial ole
in ade be ween non-U.S. coun ies ha use he dolla as hei in oicing cu ency. The p e alence
o dolla in oicing in global ade leads o bo h in la ion spillo e s ia a widesp ead su ge in impo
p ices as well as nega i e ou pu spillo e s in esponse o he app ecia ion o he dolla (Gopina h
and Neiman, 2014; Gopina h e al., 2020; Geo giadis and Schumann, 2021; Cook and Pa el, 2023).2
Mone a y policy also ope a es globally h ough a inancial channel by a ec ing asse p ices, which
ei he elaxes o binds balance shee o le e age cons ain s and hus shapes in es o s’ isk a e sion
(see, in e alia, Fa hi and We ning, 2014; B uno and Shin, 2015; Rey, 2016). This, in u n,
2Geo giadis and Schumann (2021) a gue ha asymme ies in dolla in oicing sha es be ween coun ies and im-
po s/expo s lead o highe ou pu spillo e s. In he case o ull dominan cu ency pa adigm (all ade is in oiced
in dolla ), hi d-coun y expendi u e-swi ching e ec s would nulli y (in e ms o ou pu ). Asymme ies hen cause
expendi u e-swi ching be ween and wi hin coun ies. Cook and Pa el (2023) a gue ha global alue chains lead o
asymme ic adjus men s in ade o he dominan -cu ency economy compa ed o egional economies.
3
has signi ican e ec s on c oss-bo de capi al lows, unding cos s o agen s, and e en ually eeds
back in o asse p ices. This causes a s ong case o an in e na ional isk channel o mone a y
policy (Mi anda-Ag ippino and Rey, 2020). Pa icula ly, he exis ence o he global inancial cycle
(see Rey, 2015, 2016) ac s as a ansmi e o he inancial channel. Again, he cu ency plays
an impo an ole, because he dolla domina es he global inancial sys em and is an impo an
ansmi e o global isk shocks (Geo giadis e al., 2021, 2023). Agains he e idence p esen ed,
his pape is pa icula ly in e es ed in unde s anding whe he and how he ele ance o he ade
and inancial channel has changed o e ime due o globaliza ion. As indica ed by he mo i a ing
e idence in Figu e 1, i is plausible ha bo h channels ha e gained s eng h du ing globaliza ion
and disen angling hem is ul ima ely an empi ical ma e .
We ind s ong e idence o g owing global spillo e s o U.S. mone a y policy. The nega i e
esponse o RoW indus ial p oduc ion has inc eased signi ican ly o e he las decades o ising
globaliza ion. The spillo e sizes only s abilized wi h he onse o he G ea Recession, which is
consis en wi h he slowdown in ade and inancial in eg a ion ha we see in he da a. The obse ed
ime- a ia ion is subs an ial and s a is ically signi ican . While a one pe cen age poin (pp) hike in
mone a y policy leads o abou a −0.6% con ac ion in RoW indus ial p oduc ion in 1993, a same-
sized shock in 2008 esul s in a down u n o abou −3.2%. When we look a he ansmission ia he
ade and inancial channel, we ind ha mone a y policy shocks gene a e global ade con ac ions
and inancial ic ions. Howe e , we ind a disconnec be ween he wo channels when looking a
he ime- a ia ion. While he e ec s on RoW ade a e signi ican ly ime- a ying and ack well he
pa e n we ind o RoW indus ial p oduc ion, he ones on global inancial condi ions a e ela i ely
cons an . Finally, ising spillo e s in eal ac i i y a e somehow mi o ed by a s eng hening o
domes ic e ec s on U.S. p ices and ou pu . G ea e in e na ional spillo e s imply a g ea e po en ial
o hese e ec s o spillback o he U.S. economy. Based on p e ious es ima es by B ei enlechne
e al. (2022), we a gue ha spillo e -spillback loop e ec s a e he likely d i e o he ising e ec s
o U.S. policy shocks on he U.S. economy ha we ind in he da a. This calls o inco po a ing
spillback e ec s in he calib a ion o U.S. policy decisions oday mo e han in he pas .
Gi en he disconnec in he e olu ion o e ec s on global ade (highly ime-dependen ) and
inancial condi ions ( ela i ely cons an o e ime), ou es ima ions sugges ha ade in eg a ion
is he p ima y ac o d i ing he a iabili y in spillo e e ec s. Howe e , in he p esence o a
la ge and ime- a ying global inancial mul iplie , e en mino shi s in he impac on inancial
condi ions could explain a signi ican p opo ion o he a ia ions in he eac ions o RoW indus ial
p oduc ion. We dig deepe in o his aspec by pe o ming a he e ogenei y analysis and coun e ac ual
simula ions. Rega ding he he e ogenei y analysis, we gene ally ind s onge spillo e e ec s o
eme ging ma ke s (EMEs) han ad anced (AE) economies. Gi en ha EMEs a e mo e impac ed by
U.S. dis u bances and hei sha e o wo ld economic ac i i y ha e inc eased om app oxima ely 20%
4
o 40% o e he pas h ee decades (Lane, 2019), pa o he obse ed ime- a ia ion is explained
mechanically by composi ion e ec s. We hen cons uc a (balanced) panel o 22 coun ies and
es ima e he esponse o coun y-speci ic indus ial p oduc ion o U.S. mone a y policy shocks o e
ime. We analyze he ou comes o his analysis in conjunc ion wi h coun y-speci ic inancial and
ade in eg a ion da a. This enables us o explo e he co ela ion be ween he g ow h o spillo e s and
he inc easing economic in eg a ion and unde s and he na u e o his ela ionship. O e all, we ind
ha coun ies exhibi ing g ea e his o ical le els o ade and inancial in eg a ion end o unde go
mo e p onounced economic down u ns in he a e ma h o con ac iona y U.S. policy shocks.
Bo h channels a e ac i e on a e age. Howe e , when i comes o explaining he ime- a ia ion in
e ec s, ou es ima ions again sugges ha he ade dimension holds g ea e signi icance. While
coun ies ha ha e subs an ially enhanced hei ade in eg a ion end o expe ience exace ba ed
ecessiona y impac s as ime p og esses, no such associa ion eme ges in ela ion o heigh ened
inancial in eg a ion. We also ind e idence ha ising inancial in eg a ion is no esponsible
o ime- a ia ion in eal spillo e s by unning coun e ac ual simula ions. To do so, we i s
iden i y a global inancial shock ollowing he app oach o Gilch is and Zak ajšek (2012), and
hen simula e coun e ac ual scena ios in which U.S. mone a y dis u bances do no impac global
inancial condi ions (Sims and Zha, 2006; An olín-Díaz e al., 2021; McKay and Wol , 2023). By
shu ing o he inancial channel in he ansmission o U.S. mone a y policy shocks, we ob ain
wo esul s. On he one hand, he inancial channel is impo an on a e age, accoun ing o abou
30 −40% o he o e all esponse o RoW indus ial p oduc ion. When i comes o explaining he
ime- a ia ion in he RoW indus ial p oduc ion esponse, on he o he hand, i s ole is again ound
o be limi ed.
The pape con ibu es o he as li e a u e ha uses linea models o documen he nega i e
e ec s o U.S. mone a y policy hikes on he global and eal inancial cycles.3We a e closely
ela ed o he con ibu ions o Mi anda-Ag ippino and Rey (2020), Dedola e al. (2017), Iaco iello
and Na a o (2019), and Degaspe i e al. (2020). While he i s pape ex ensi ely in es iga es he
impac o U.S. mone a y policy on he global inancial cycle, he o he s ocus on examining he
he e ogenei y o spillo e s in eal economic ac i i y. Di e en ly, we use a ime- a ying model and
documen he inc easing in e na ional spillo e s o U.S. policy shocks.4
3Cano a (2005), Maćkowiak (2007), Geo giadis (2016), Feldki che and Hube (2016), Dedola e al. (2017), Iaco iello
and Na a o (2019), Degaspe i e al. (2020), Geo giadis and Schumann (2021), and Ca’Zo zi e al. (2023) analyze he
o eign esponses o eal ac i i y o U.S. policy decisions. See Rey (2016), Ge ko and Rey (2017), Jo dà e al. (2019),
Habib and Vendi i (2019), Dées and Galesi (2021), and Mi anda-Ag ippino and Rey (2020) o he ansmission o
in e na ional inancial condi ions. Obs eld (2020) p o ides a comp ehensi e o e iew on he global dimension o
U.S. mone a y policy.
4The ime- a ying na u e o mone a y policy shocks on domes ic ou comes has been documen ed in many s udies, see
o ins ance Cogley and Sa gen (2005), P imice i (2005), Boi in and Giannoni (2006), Cano a and Gambe i (2009),
Galí and Gambe i (2015) and Aas ei e al. (2023).
5
Addi ionally, we con ibu e o he li e a u e examining he ime- a ying dimension o in e na-
ional mone a y policy. Ou pape connec s closely o Liu e al. (2022), who es ima e a ime- a ying
pa ame e model o join ly model mone a y policy decisions in he U.S., U.K., and Eu o a ea. While
hei s udy highligh s ime- a ying ne wo k s uc u es in cen al banks’ decisions, we in es iga e he
global spillo e s o U.S. shocks and hei unde lying d i e s. Ilze zki and Jin (2021) compa e he in-
e na ional ansmission o U.S. mone a y policy shocks p io o and a e he 1990s. They ind ha ,
be o e he 1990s, wo ld indus ial p oduc ion declines in esponse o a U.S. mone a y igh ening,
while du ing he pe iod 1990-2007, U.S. con ac ions a e expansiona y ab oad. In con as o hei
pape , we model he changes in in e na ional ansmission channels using a ime- a ying model ( s.
sample-spli ing s a egy) and ocus on he dynamics wi hin he pos -1990s pe iod. Fu he mo e
and con a y o hei indings, we ind e idence in a o o a g owing (nega i e) ole o U.S. policy
igh enings o global economic ac i i y.
The emainde o he pape p oceeds as ollows. Sec ion 2 discusses he empi ical s a egy
and speci ica ion. In Sec ion 3, we p esen he empi ical esul s, including a ious ex ensions, he
he e ogenei y analysis, he coun e ac ual exe cises, and a ba e y o sensi i i y checks. Finally,
Sec ion 4 concludes.
2. Empi ical Me hodology
We empi ically examine he in e na ional spillo e s o U.S. mone a y policy using a medium-scale
TVP-VAR model ha allows o ime- a ia ion in he pa ame e s. On he domes ic le el, we
include he ede al unds a e, U.S. consume p ice index, and U.S. indus ial p oduc ion. This
in o ma ion se allows us o ack he domes ic ansmission channels o U.S. mone a y policy
shocks. Gi en ou in e es in in e na ional spillo e s, we u he include indica o s o global
economic ac i i y, ade, and he inancial cycle. We p oxy he global eal and ade cycles using
es o he wo ld (RoW, i.e., excluding he U.S.) indus ial p oduc ion and expo indices cons uc ed
by he Fede al Rese e Bank o Dallas (G ossman e al., 2014).5The global inancial cycle index is
de i ed om a dynamic ac o model cons uc ed om a comp ehensi e panel o isky asse p ices
aded wo ldwide (Mi anda-Ag ippino and Rey, 2020), summa izing global inancial condi ions.6
Consis en wi h he li e a u e, we s a iona ize he a iables be o e es ima ing he TVP-VAR. We
use he indica o s o U.S. p ices, U.S. indus ial p oduc ion, RoW indus ial p oduc ion, and RoW
expo s in log-di e ences o compu e he g ow h a e and keep he emaining a iables in le els.
Ou mon hly da ase co e s he ime span om 1980M1 o 2017M12, which we spli in wo pa s.
5We p oxy global ade wi h RoW expo s since i comple ely excludes U.S. p oduced goods. Howe e , we ind e y
simila esul s when conside ing RoW impo s.
6Aldaso o e al. (2023) show ha he global inancial cycle index as a p ice-based global ac o is ema kably simila
o a quan i y-based global ac o based on c oss-bo de capi al lows.
6
Figu e 3 con i ms he indings ob ained in he linea se ing, wi h a mone a y policy igh ening
being ollowed by con en ional nega i e demand- ype e ec s in he U.S. economy. A domes ic
con ac ion in indus ial p oduc ion goes along wi h a decline in p ices. The magni udes o hese
e ec s a e consis en wi h he linea speci ica ion. Some commen s, howe e , a e in o de . Fi s ,
o ob ain same-size shocks, he impac e ec o he shocks on he policy a e i sel diminishes
o e ime. This pa e n is consis en wi h a p og essi e decline o he long- un end o he U.S.
in e es a e, which eaches i s ough wi h he ze o lowe bound pe iod (2009-2015).16 Second,
ime-dependen pa e ns a ise in he esponses o U.S. agg ega es. Peak con ac ions in domes ic
indus ial p oduc ion agg a a e o e ime ( ising om −1.9% in 1993 o −4.4% in 2008) and only
s abilize a hei lowes le el wi h he onse o he G ea Recession ( his e olu ion is in line wi h he
indings o Paul (2020), who conside a di e en speci ica ion). A simila pa e n eme ges o U.S.
p ices, which con ac s onge o e ime. The impac in he ini ial pe iods o he sample is abou
−0.6%, which g ows in magni ude and eaches −1.1% du ing he G ea Recession.
Figu e 4 p esen s he dynamic esponses o RoW expo s, RoW indus ial p oduc ion, and he
global inancial cycle. Consis en wi h an ampli ied ole o he in e na ional ansmission channel
o U.S. mone a y policy shocks, we ind an inc ease in he ( ecessiona y) e ec s o U.S. shocks on
global ade, p oduc ion, and he inancial cycle. Th oughou he pe iod conside ed, all a iables
eac nega i ely (and mos o he ime s a is ically signi ican so) o a mone a y policy shock. The
peak e ec on RoW eal ac i i y s ongly inc eases o e ime, inc easing om −0.6% in 1993 o
−3.2% in 2008.17 In his ega d, he linea VAR seems o cap u e well he mean e ec o e ime,
masking hough he ime-speci ic he e ogenei y. Simila ly, he esponse o RoW expo s (as a
measu e o ade) is s ongly g owing in magni ude ( om −3.6% o −11.4%).18 The ime- a ia ion
in he impulse esponse o he global inancial cycle is mo e limi ed bu ye non-negligible om an
economic poin o iew, wi h an inc ease o e ime om −0.63 o −0.90.
Al hough he magni ude o he global e ec s has been g owing since he beginning o he
sample, he pace o his g ow h no ably accele a es in he ea ly 2000s. This in ensi ica ion o
in e na ional spillo e s coincides p ecisely wi h a pe iod cha ac e ized by ac o s such as he ade
boom, elaxed inancial egula ion and supe ision o banks (Shin, 2012), and a sha p ise in
16 Figu e C1 epo s he e olu ion o he long- un end along wi h he on impac esponse o he U.S. policy a e
unde lying ou speci ica ion. Acco dingly, he long- un end is ela i ely s able in 1990s, bu s a s dec easing in he
2000s, and becomes e en nega i e du ing he ze o lowe bound pe iod. See Appendix C o mo e de ails.
17 Gi en ha he shocks a e no malized o ha e he same magni ude, he e ec s o e ime a e compa able. Howe e ,
he shape o he policy a e’s esponses di e en . To elimina e any conce n, we epo in Figu e D1 he a ios o he
peak esponses o RoW s. U.S. indus ial p oduc ion (see Ho mann and Pee sman (2017) o a simila use o a io
impulse esponse unc ions). In each 𝑡, he shock hi ing he wo a iables is he same. We ob ain an inc ease in he
a io esponse o e ime, which again poin s owa ds signi ican ime- a ia ion in global spillo e s.
18 In un epo ed checks a ailable upon eques , we ind ex emely simila esul s when conside ing RoW impo s. This
is consis en wi h a symme ic con ac ion o RoW expo and impo in esponse o U.S. policy shocks (Gopina h
e al., 2020; Degaspe i e al., 2020).
13
Figu e 4: Time-Va ying Impulse Responses Func ions o Global Va iables.
No es: Responses o domes ic a iables o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin (1 pp, 100 basis poin s)
inc ease in he ede al und a e in 1993M1. Le column epo s he e olu ion o e ime o he median impulse- esponse unc ions. The igh column
epo s he peak e ec s o e ime o each a iable wi h 68% pos e io c edible se s. G ey do ed ho izon al lines: (cons an ) peak e ec s in he linea
VAR o a same-sized shock (which consis s o a 0.76 pp inc ease in he policy a e in he linea speci ica ion).
14
dolla -denomina ed c oss-bo de posi ions held by in e na ional ac o s (Rey, 2016). All hese
ac o s indica e a heigh ened ole o in e na ional linkages in ansmi ing U.S. mone a y policy
decisions ab oad, which is e iden in he da a. This downwa d end s abilizes only wi h he G ea
Recession. Se e al explana ions could accoun o his s abiliza ion. I may be a ibu ed o he
slowdown in ade and inancial in eg a ion esul ing om he inancial c isis, he e ec i eness
o in e na ional mac o-p uden ial policies (which educed banks’ isk- aking p opensi y and hei
ele ance in in e media ion), o he p esence o he ze o lowe bound pe iod, which could po en ially
impac ou es ima es. Ou empi ical model appea s o e ec i ely cap u e his economic na a i e.
The s onge global e ec s o U.S. mone a y policy shocks can also igge spillo e -spillback
loop e ec s. This means ha in e na ional ecessiona y e ec s spillback o he domes ic economy
and a ec i s economic agg ega es. E idence o an ac i e spillback mechanism is p o ided by
B ei enlechne e al. (2022), who ind ha spillbacks accoun o nea ly hal o he o e all e ec o
U.S. mone a y policy on domes ic eal ac i i y using coun e ac ual simula ions. In ou es ima ions,
such spillo e -spillback loops a e he likely explana ion o he ime- a ying e ec s ha we ind in
he esponse o U.S. domes ic a iables.
Finally, we in es iga e whe he ime- a ia ion is s a is ically signi ican by ocusing on pa icula
episodes in he sample. We conside he pe iods 1993M1 and 2008M8. We epo in Figu e 5 he
impulse esponse unc ions o he wo ime pe iods (le column) and he pos e io dis ibu ion
o he di e ences ( igh column). The p ima y dis inc ion be ween he wo examined pe iods lies
in he e ec s on RoW indus ial p oduc ion. On he one hand we ind no e idence in suppo
o a global down u n in eal ac i i y ollowing mone a y policy con ac ions in 1993, which is
somehow in line wi h he esul s in he linea model. On he o he hand, such nega i e e ec s
a e clea ly p esen in 2008. A simila pa e n a ises o RoW expo . In addi ion, as Figu e 5
indica es, di e ences o RoW expo s and RoW indus ial p oduc ion be ween hese ime pe iods
a e s a is ically signi ican ly di e en om ze o ( igh column). Time- a ia ion is a ele an pa e n
o global eal spillo e s. In sha p con as , he e idence o he global inancial cycle is weak:
While U.S. dis u bances gene a e signi ican inancial ic ions in bo h pe iods, he di e ence in he
impulse- esponse unc ions is no s a is ically di e en om ze o.
3.3 Wide P opaga ion Channels
To ge a be e unde s anding o he ime- a ying ansmission o U.S. mone a y policy shocks,
we analyze he e ec s on a ange o ele an mac oeconomic and inancial a iables. To compu e
he impulse esponses, we augmen he baseline VAR by one a iable a a ime, which esul s in
speci ica ions wi h a o al o se en a iables. Since he s a e-space would become oo la ge o
es ima e sensible esul s, we educe he numbe o lags o wo. Es ima ion and p io speci ica ion
a e kep unchanged (we e e o Appendix A o he exac a iable de ini ions and ans o ma ions).
15
Figu e 5: Di e ences in Impulse Responses: 1993M1 s. 2008M8
No es: Le column: median impulse- esponse unc ions and 68% pos e io c edible se s o he a iables conside ed a 1993M1 s. 2008M8.
Ve ical axis: pe cen age change; ho izon al axis: impulse esponse ho izon in mon hs. Righ column: di e ence in impulse esponses in such
pe iods (median and 68% pos e io c edibili y in e als a e epo ed).
We epo he esul s o hese ex ensions in Figu e 6 and Figu e 7. On he domes ic le el, we ind
ha U.S. mone a y igh enings a e ollowed i) by ab up inc eases in U.S. co po a e c edi sp eads -
p oxied by he excess bond p emium (EBP) o Gilch is and Zak ajšek (2012); ii) an app ecia ion o
he U.S. dolla e ec i e exchange; and iii) a ise in he U.S. expo impo a io. The sign o hese
e ec s a e as expec ed. Fi s , Calda a and He bs (2019) highligh he ole o inancial condi ions in
16
Figu e 6: Impulse Responses o Ex ensions o he Baseline Speci ica ion.
No es: Responses o addi ional a iables o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin (1 pp, 100 basis poin s)
inc ease in he ede al und a e in 1993M1. Le column epo s he e olu ion o e ime o he median impulse- esponse unc ions. The igh column
epo s he peak e ec s o e ime o each a iable wi h 68% pos e io c edible se s.
ansmi ing mone a y policy shocks. Simila o hei indings, an inc ease in he EBP is associa ed
wi h a igh ening o inancial condi ions as expec ed h ough he (domes ic) isk- aking channel
o mone a y policy.19 Second, he app ecia ion o he dolla is expec ed by he unco e ed in e es
a e pa i y. Thi d, he posi i e esponse o he U.S. expo impo a io implies no disce nible
expendi u e-swi ching channel, as expec ed in he dominan cu ency pa adigm (Gopina h e al.,
19 As men ioned ea lie , and in con as o Calda a and He bs (2019), ou speci ica ion emains obus o he inclusion
o c edi sp eads due o he choice o he ins umen (Mi anda-Ag ippino and Ricco, 2021; Mi anda-Ag ippino and
Ricco, 2023).
17
Figu e 7: Impulse Responses o Ex ensions o he Baseline Speci ica ion.
No es: Responses o addi ional a iables o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin (1 pp, 100 basis poin s)
inc ease in he ede al und a e in 1993M1. Le column epo s he e olu ion o e ime o he median impulse- esponse unc ions. The igh column
epo s he peak e ec s o e ime o each a iable wi h 68% pos e io c edible se s.
2020). While U.S. expo s decline in esponse o less agg ega e demand, U.S. impo s do no
ou weigh his o ce. The dolla app ecia ion leads in p inciple o an inc ease in he compe i i eness
o o eign goods and o a boos in impo s. Howe e , i mos o hese impo s a e al eady p iced
in dolla , his coun e ac ing o ce anishes (see e.g. Degaspe i e al., 2020 o simila esul s).
We unco e ha hese esponses, while being s a is ically signi ican h oughou he whole sample,
exhibi e y mild e idence o ime- a ia ion.
18
We now u n o he global esponses in Figu e 7. In line wi h he dominan cu ency pa adigm,
we ind ha U.S. igh enings esul in some in la iona y p essu es in he RoW ( ia he e alua ion
o he dolla and a widesp ead su ge in impo p ices). Pa icula ly in e es ing is he decline in
in la ion spillo e s o e ime ( om abou +2.1% o +0.1% a peak), which can be a ibu ed o
he ise o global alue chain pa icipa ion (Geo giadis e al., 2019).The in la iona y p essu es in
he RoW a e ackled wi h an endogenous inc ease in in e es a es om he majo cen al banks,
p oxied by he policy a e indica o o RoW economies.20 The g adual dec ease in he policy a e
esponse ha we ind in he da a is consis en wi h i) he inc easing e ec s on RoW p oduc ion and
ii) he diminishing in la ion spillo e s. Finally, we look a he esponses o he global s ock ma ke ,
measu ed h ough he RoW MSCI index. The indica o shows an ab up decline in esponse o
U.S. dis u bances, wi h peak esponses being s able o e ime (abou −15%). These indings a e in
line wi h he esponse o he global inancial ac o o Mi anda-Ag ippino and Rey (2020) in ou
benchma k speci ica ion: Again, we do no ind e idence o ime- a ia ion in he e ec s o U.S.
shocks on global inancial ma ke s.
3.4 He e ogenei y Analysis
We ha e documen ed a signi ican ime- a ia ion in he in e na ional spillo e s o U.S. mone a y
policy. In his and he ollowing sec ion, we conduc addi ional exe cises o explain he indings
and link i o di e en channels. As highligh ed by Kalemli-Özcan (2019), De Leo e al. (2022),
and Ca’Zo zi e al. (2023), mone a y policy spillo e s a e qui e asymme ical be ween ad anced
and eme ging economies (AEs and EMEs). Gi en ha EMEs ha e inc eased hei sha e o wo ld
economic ac i i y om app oxima ely 20% o 40% o e he pas h ee decades, pa o he obse ed
ime- a ia ion could be explained by mechanical composi ion e ec s. Hence, we look in o di e -
ences o spillo e s o AEs and EMEs. To s a disen angling he channels a play, we a e in e es ed
in inding c oss-sec ional a ia ion. To do so, we es ima e coun y-speci ic spillo e s o indus ial
p oduc ion. In he nex sec ion, we will inally combine hese es ima es wi h coun y-speci ic ade
and inancial da a o e alua e he e olu ion o he ansmission channels.
To look in o he di e ence be ween AEs and EMEs, we adap he baseline speci ica ion by
eplacing he RoW indus ial p oduc ion indica o and RoW expo indica o wi h he espec i e
indica o o AEs o EMEs.21 The esul s, shown in Figu e 8, e eal mo e p onounced ecessiona y
20 The comp ehensi e RoW policy a e indica o published by he Fede al Rese e o Dallas displays explosi e pa e ns
in he 1980s and 1990s, d i en by he me ging economies’ da a. This makes he es ima ion in easible. Hence, we
conside he index o RoW ad anced economies as a p oxy o RoW policy esponse.
21 Da a is again aken om he Da abase o Global Economic Indica o s o he Fede al Rese e Bank o Dallas. The
indus ial p oduc ion se ies o EMEs is a ailable om 1987M1. To es ima e he model s a ing om 1980M1, we
assume ha , om 1980M1 o 1986M12, he g ow h a e in indus ial p oduc ion o EMEs is equal o he one in
he comp ehensi e RoW se ies (we always obse e he ac ual EMEs se ies in he es ima ion sample). See he exac
ans o ma ions and lis in Appendix A.
19
Figu e 8: Compa ison o Ad anced and Eme ging Ma ke Economies.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin (pp) inc ease in he ede al und a e in 1993M1.
Uppe panel epo s he e olu ion o e ime o he median impulse- esponse unc ions. The lowe panel epo s he peak e ec s o e ime o each
a iable wi h 68% pos e io c edible se s.
e ec s in eme ging economies - consis en wi h exis ing s udies. The peak e ec on eme ging
(ad anced) coun ies’ indus ial p oduc ion is −1.6% (−0.5%)in 1993 and −3.2% (−2.4%)in
2008. In hose yea s, he peak esponse o RoW indus ial p oduc ion, which combines bo h
eme ging and ad anced economies, is −0.6% and −3.2%, espec i ely. While in he beginning o
he sample pe iod he esponse o RoW indus ial p oduc ion is s ongly il ed owa ds he one o
ad anced economies (eme ging economies ha e li le ele ance in he o e all index), he esponse
in 2008 aligns mo e wi h he e ec in EMEs (o cou se, es ima ion unce ain y mus be ake in o
accoun ). Fu he mo e, he dynamic esponses e eal ha he esponse o AEs’ indus ial p oduc ion
20
e u ns back o he ze o line ela i ely quickly, while he con ac ion in EMEs’ indus ial p oduc ion
is a mo e pe sis en . Since he composi ion o RoW indus ial p oduc ion be ween 1993 and 2008
has s ongly changed (in a ou o EMEs), composi ion e ec s can accoun o pa o he ising
spillo e s ha we obse e in he da a. Howe e , he ime- a ia ion in he esponses o bo h eme ging
and ad anced economies’ indus ial p oduc ion signals ha mechanical composi ion e ec s canno
ully explain he ising spillo e s, which mus depend on o he ac o s.
In o de o make p og ess on his issue, we a e in e es ed in e ie ing mo e c oss-sec ional
he e ogenei y wi h espec o in e na ional spillo e s. The e o e, we b eak down he esponse o
RoW indus ial p oduc ion in o i s coun y-speci ic componen s. We e-es ima e ou benchma k
VAR eplacing he agg ega e RoW indus ial p oduc ion wi h na ional-le el indices. This allows
us o examine he esponse a a mo e g anula le el. We conside a o al o 22 coun ies in ou
analysis. These coun ies we e selec ed based on wo c i e ia: i) hey a e included in he agg ega e
RoW p oduc ion measu e o Fede al Rese e Bank o Dallas and, ii) mon hly da a is a ailable om
1980 onwa ds. Ou sample includes: Aus ia, Belgium, B azil, Canada, Chile, Colombia, F ance,
Ge many, G eece, India, I aly, Sou h Ko ea, Japan, Malaysia, Mexico, Ne he lands, Pe u, Po ugal,
Sou h A ica, Spain, Sweden, UK.22
To sa e space, we epo each coun y’s median ime- a ying impulse esponse unc ions in
Appendix E (Figu e E1-Figu e E4). We also epo he e olu ion o coun y-speci ic mean e ec s
o e ime in Figu e 9. While we ha e so a conside ed peak e ec s as summa y s a is ics o impulse
esponse unc ions, we now shi o mean e ec s (i.e., we epo he mean e ec om ℎ=0, ..., 24
o each 𝑡) o accoun o he he e ogenei y in he shapes o he e ec s. Gi en ha we obse e
a posi i e sho - un esponse o indus ial p oduc ion in a signi ican sha e o coun ies - which
would no be conside ed when ocusing on peak e ec s, mean e ec s seem o be a mo e adequa e
measu e o he e ec s in a gi en coun y. (Howe e , he esul s a e e y simila when conside ing
peak e ec s.) The indings in Figu e 9 align wi h ou benchma k es ima ions and e eal a consis en
pa e n o inc easing spillo e e ec s. A U.S. policy igh ening gene a es ecessiona y e ec s in
mos coun ies conside ed, especially a e he ea ly 2000s. These e ec s end o in ensi y o e ime
un il he G ea Recession, a e which hey s abilize. The magni udes o he economic down u ns
a e in he ballpa k o ou es ima es o he agg ega e RoW p oduc ion. Addi ionally, wo sou ces o
c oss-sec ional he e ogenei y among coun ies a ise. Fi s ly, he a e age ele ance o ecessiona y
e ec s a ies ac oss coun ies: Ce ain coun ies, such as Canada and Mexico, his o ically exhibi a
g ea e suscep ibili y o U.S. shocks, whe eas o he s like he UK a e less a ec ed. This dimension
can be summa ized by aking he a e age o e ime o he peak e ec in coun y 𝑖. Secondly,
he inc ease in ecessiona y e ec s can be mo e o less subs an ial, and his dimension can be
22 We hank he au ho s o G ossman e al. (2014) o kindly sha ing hei da a wi h us. Da a sou ces a e desc ibed he e:
h ps://www.dallas ed.o g/ esea ch/in e na ional/dgei# ab2$
21
Figu e 9: Mean E ec s o Coun y-Speci ic Indus ial P oduc ion.
No es: Mean e ec s in he esponses o coun y-speci ic (non-U.S.) indus ial p oduc ion indices o a con ac iona y U.S. mone a y policy shock ha
induces a one pe cen age poin (pp) inc ease in he ede al und a e in 1993M1.
summa ized by compa ing he ecessiona y e ec s in coun y 𝑖a he end o he sample wi h hose
a he beginning. Time- a ia ion is pe asi e in Japan, Ge many, and Spain, while i is mo e limi ed
in he Ne he lands and UK.
3.5 E alua ion o he Channels
Ou empi ical analysis again poin s owa ds signi ican ime pa e ns in global spillo e s. Bu wha
a e he main d i e s o such dynamics? We shed ligh on his aspec by es ablishing a connec ion
be ween coun y-speci ic e ec s and coun y-speci ic in o ma ion on ade and inancial in eg a ion.
We de ine inancial in eg a ion as he a io o coun ies’ (dolla -denomina ed) ex e nal asse s and
liabili ies o GDP, while ade in eg a ion is de e mined by he a io o o al ade o GDP ( o a
simila choice, see Ca’Zo zi e al., 2023).23 Financial and ade in eg a ion a e hen compu ed
23 We e ie e he da a o inancial in eg a ion om Béné ix e al. (2019). Da a o ade in eg a ion is ins ead aken
om he Wo ld Bank (Wo ld De elopmen Indica o s). Gi en ha coun y-speci ic sha es o dolla in oiced ade
22
Re e ences
Aas ei , K A, Fu lane o, F and Lo ia, F (2023) ‘Has he Fed Responded o House and S ock
P ices? A Time-Va ying Analysis’. Re iew o Economics and S a is ics Vol. 105(5), pp. 1314–
1324.
Abba e, A, Eickmeie , S, Lemke, W and Ma cellino, M (2016) ‘The Changing In e na ional
T ansmission o Financial Shocks: E idence om a Classical Time-Va ying FAVAR’. Jou nal
o Money, C edi and Banking Vol. 48(4), pp. 573–601.
Aldaso o, I, A djie , S, Bo io, C and Disya a , P (2023) ‘Global and Domes ic Financial Cycles:
Va ia ions on a Theme’. In e na ional Jou nal o Cen al Banking Vol. 5, pp. 49–98.
Ami -Ahmadi, P, Ma hes, C and Wang, M.-C (2023) Unde s anding Ins umen s in Mac oeco-
nomics – A S udy o High-F equency Iden i ica ion. Wo king pape .
An olín-Díaz, J, Pe ella, I and Rubio-Ramí ez, J F (2021) ‘S uc u al Scena io Analysis wi h
SVARs’. Jou nal o Mone a y Economics Vol. 117(C), pp. 798–815.
Baue , M and Swanson, E T (2023a) ‘A Reassessmen o Mone a y Policy Su p ises and High-
F equency Iden i ica ion’. NBER Mac oeconomics Annual Vol. 37, pp. 87–155.
Baue , M D and Swanson, E T (2023b) ‘An Al e na i e Explana ion o he “Fed In o ma ion
E ec ”’. Ame ican Economic Re iew Vol. 113(3), pp. 664–700.
Boi in, J and Giannoni, M P (2006) ‘Has Mone a y Policy Become Mo e E ec i e?’ The Re iew
o Economics and S a is ics Vol. 88(3), pp. 445–462.
Boz, E, Casas, C, Geo giadis, G, Gopina h, G, Helena Le Mezo and, A M and Nguyen, T (2022)
‘Pa e ns o In oicing Cu ency in Global T ade: New E idence’. Jou nal o In e na ional
Economics Vol. 136(103604), pp. 1–16.
B ei enlechne , M, Geo giadis, G and Schumann, B (2022) ‘Wha Goes A ound Comes A ound:
How La ge A e Spillbacks om US Mone a y Policy?’ Jou nal o Mone a y Economics Vol.
131, pp. 45–60.
B uno, V and Shin, H S (2015) ‘C oss-Bo de Banking and Global Liquidi y’. The Re iew o
Economic S udies Vol. 82(2).
B äuning, F and She emi o , V (2023) ‘The T ansmission Mechanisms o In e na ional Business
Cycles: Ou pu Spillo e s h ough T ade and Financial Linkages’. IMF Economic Re iew Vol.
71((1)), 300–325.
29
Béné ix, A, Gau am, D, Ju enal, L and Schmi z, M (2019) C oss-Bo de Cu ency Exposu es.
New E idence Based on an Enhanced and Upda ed Da ase . IMF Wo king Pape No. 2019/299.
Calda a, D and He bs , E (2019) ‘Mone a y Policy, Real Ac i i y, and C edi Sp eads: E id-
ence om Bayesian P oxy SVARs’. Ame ican Economic Jou nal: Mac oeconomics Vol. 11(1),
pp. 157–92.
Cano a, F (2005) ‘The T ansmission o US Shocks o La in Ame ica’. Jou nal o Applied Econo-
me ics Vol. 20(2), pp. 229–251.
Cano a, F and Gambe i, L (2009) ‘S uc u al Changes in he US Economy: Is he e a ole o
mone a y policy?’ Jou nal o Economic Dynamics and Con ol Vol. 33(2), pp. 477–490.
Ca e , C K and Kohn, R (1994) ‘On Gibbs Sampling o S a e Space Models’. Biome ika Vol.
81(3), pp. 541–553.
Ca’Zo zi, M, Dedola, L, Geo giadis, G, Ja ocinski, M, S acca, L and S asse , G (2023)
‘Making Wa es: Mone a y Policy and I s Asymme ic T ansmission in a Globalized Wo ld’.
In e na ional Jou nal o Cen al Banking Vol. 19(2), pp. 95–144.
Cogley, T and Sa gen , T J (2005) ‘D i s and Vola ili ies: Mone a y Policies and Ou comes in he
Pos WWII US’. Re iew o Economic Dynamics Vol. 8(2), pp. 262–302.
Coibion, O (2012) ‘A e he E ec s o Mone a y Policy Shocks Big o Small?’ Ame ican Economic
Jou nal: Mac oeconomics Vol. 4((2)), pp. 1–32.
Cook, D and Pa el, N (2023) ‘Dolla In oicing, Global Value Chains, and he Business Cycle
Dynamics o In e na ional T ade’. Jou nal o In e na ional Economics Vol. 145, p. 103839.
De Leo, P, Gopina h, G and Kalemli-Özcan, (2022) Mone a y Policy Cyclicali y in Eme ging
Economies. NBER Wo king Pape No. 30458.
Dedola, L, Ri ol a, G and S acca, L (2017) ‘I he Fed Sneezes, Who Ca ches a Cold?’ Jou nal
o In e na ional Economics Vol. 108, pp. 23–41.
Dées, S and Galesi, A (2021) ‘The Global Financial Cycle and US Mone a y Policy in an In e -
connec ed Wo ld’. Jou nal o In e na ional Money and Finance Vol. 115, p. 102395.
Degaspe i, R, Hong, S and Ricco, G (2020) The Global T ansmission o US Mone a y Policy.
CEPR Discussion Pape No. 14533.
Degaspe i,RandRicco,G (2021) ‘In o ma ion and Policy Shocks in Mone a y Su p ises’. Wo king
Pape .
30
Del Neg o, M and P imice i, G E (2015) ‘Time Va ying S uc u al Vec o Au o eg essions and
Mone a y Policy: A Co igendum’. The Re iew o Economic S udies Vol. 82(4), pp. 1342–1345.
Fa hi, E and We ning, I (2014) ‘Dilemma no T ilemma? Capi al Con ols and Exchange Ra es
wi h Vola ile Capi al Flows’. IMF Economic Re iew Vol. 62(4), pp. 569–605.
Feldki che , M and Hube , F (2016) ‘The In e na ional T ansmission o US Shocks — E idence
om Bayesian Global Vec o Au o eg essions’. Eu opean Economic Re iew Vol. 81, pp. 167–
188.
F ühwi h-Schna e , S (1994) ‘Da a Augmen a ion and Dynamic Linea Models’. Jou nal o
Time Se ies Analysis Vol. 15(2), pp. 183–202.
Galí, J and Gambe i, L (2015) ‘The E ec s o Mone a y Policy on S ock Ma ke Bubbles: Some
E idence’. Ame ican Economic Jou nal: Mac oeconomics Vol. 7(1), pp. 233–57.
Geo giadis, G (2016) ‘De e minan s o Global Spillo e s om US Mone a y Policy’. Jou nal o
In e na ional Money and Finance Vol. 67(2), pp. 41–61.
Geo giadis, G, G äb, J and Khalil, M (2019) Global Value Chain Pa icipa ion and Exchange
Ra e Pass-Th ough. ECB Wo king Pape Se ies No. 2327.
Geo giadis, G, Mülle , G J and Schumann, B (2021) Global Risk and he Dolla . CEPR Discus-
sion Pape No. 16245.
Geo giadis, G, Mülle , G J and Schumann, B (2023) Dolla T ini y and he Global Financial
Cycle. CEPR Discussion Pape No. 18427.
Geo giadis, G and Schumann, B (2021) ‘Dominan -Cu ency P icing and he Global Ou -
pu Spillo e s om US Dolla App ecia ion’. Jou nal o In e na ional Economics Vol. 133,
p. 103537.
Ge ko, E and Rey, H (2017) ‘Mone a y Policy in he Capi als o Capi al’. Jou nal o he Eu opean
Economic Associa ion Vol. 15(4), pp. 721–745.
Ge le , M and Ka adi, P (2015) ‘Mone a y Policy Su p ises, C edi Cos s, and Economic Ac i -
i y’. Ame ican Economic Jou nal: Mac oeconomics Vol. 7(1), pp. 44–76.
Gilch is , S and Zak ajšek, E (2012) ‘C edi Sp eads and Business Cycle Fluc ua ions’. Ame ican
Economic Re iew Vol. 102(4), pp. 1692–1720.
Gi ai is, L, Kape anios, G and Ya es, T (2018) ‘In e ence on Mul i a ia e He e oscedas ic Time
Va ying Random Coe icien Models’. Jou nal o Time Se ies Analysis Vol. 39(2), pp. 129–149.
31
Goldbe g, L S and Tille, C (2008) ‘Vehicle Cu ency Use in In e na ional T ade’. Jou nal o
In e na ional Economics Vol. 76(2), pp. 177–192.
Gopina h, G (2015) The In e na ional P ice Sys em. Na ional Bu eau o Economic Resea ch
Wo king Pape 21646.
Gopina h, G, Boz, E, Casas, C, Díez, F J, Gou inchas, P.-O and Plagbo g-Mølle , M (2020)
‘Dominan Cu ency Pa adigm’. Ame ican Economic Re iew Vol. 110(3), pp. 677–719.
Gopina h,GandNeiman,B(2014) ‘T ade Adjus men and P oduc i i y in La ge C ises’. Ame ican
Economic Re iew Vol. 104((3)), pp. 793–831.
Gopina h, G and S ein, J C (2018) Banking, T ade, and he Making o a Dominan Cu ency.
NBER Wo king Pape No. 24485.
Gou inchas, P.-O and Rey, H (2007) ‘In e na ional Financial Adjus men ’. Jou nal o Poli ical
Economy Vol. 115(4), pp. 665–703.
Gou inchas, P.-O, Rey, H and Sauze , M (2019) ‘The In e na ional Mone a y and Financial
Sys em’. Annual Re iew o Economics Vol. 11, pp. 859–893.
G ossman, V, Mack, A and Ma ínez-Ga cía, E (2014) ‘Da abase o Global Economic Indica o s
(DGEI): A Me hodological No e’. Jou nal o Economic and Social Measu emen Vol. 39(3),
pp. 163–197.
Gü kaynak, R S, Sack, B P and Swanson, E T (2005) ‘Do Ac ions Speak Loude han Wo ds? The
Response o Asse P ices o Mone a y Policy Ac ions and S a emen s’. In e na ional Jou nal o
Cen al Banking Vol. 1(1), pp. 55–93.
Habib, M M and Vendi i, F (2019) The Global Capi al Flows Cycle: S uc u al D i e s and
T ansmission Channels. ECB Wo king Pape Se ies No. 2280.
Ho mann, B and Pee sman, G (2017) Mone a y Policy T ansmission and T ade-o s in he Uni ed
S a es: Old and New. BIS Wo king Pape s 649.
Iaco iello, M and Na a o, G (2019) ‘Fo eign E ec s o Highe US In e es Ra es’. Jou nal o
In e na ional Money and Finance Vol. 95, pp. 232–250.
Ilze zki, E and Jin, K (2021) ‘The Puzzling Change in he In e na ional T ansmission o US
Mac oeconomic Policy Shocks’. Jou nal o In e na ional Economics Vol. 130, p. 103444.
Ilze zki, E, Reinha , C M and Rogo , K S (2019) ‘Exchange A angemen s En e ing he Twen y-
Fi s Cen u y: Which Ancho Will Hold?’ The Qua e ly Jou nal o Economics Vol. 134(2),
pp. 599–646.
32
Ja ociński, M and Ka adi, P (2020) ‘Decons uc ing Mone a y Policy Su p ises – The Role o
In o ma ion Shocks’. Ame ican Economic Jou nal: Mac oeconomics Vol. 12(2), pp. 1–43.
Jo dà, Ò, Schula ick, M, Taylo , A M and Wa d, F (2019) ‘Global Financial Cycles and Risk
P emiums’. IMF Economic Re iew Vol. 67(1), pp. 109–150.
Kalemli-Özcan, S (2019) ‘U.S. Mone a y Policy and In e na ional Risk Spillo e s’. Jackson Hole
Symposium P oceedings.
Kalemli-Özcan, Şebnem and Unsal, F (2023) ‘Global T ansmission o FED Hikes: The Role o
Policy C edibili y and Balance Shee s’. B ookings Pape s on Economic Ac i i y (con e ence
d a ).
Ka lsson, S (2013) ‘Fo ecas ing wi h Bayesian Vec o Au o eg ession’. Handbook o Economic
Fo ecas ing Vol. 2, pp. 791–897.
Kim, S, Shepha d, N and Chib, S (1998) ‘S ochas ic Vola ili y: Likelihood In e ence and Com-
pa ison wi h ARCH Models’. The Re iew o Economic S udies Vol. 65(3), pp. 361–393.
Lane, P R (2019) ‘Globalisa ion and Mone a y Policy’. Speech a he Uni e si y o Cali o nia, Los
Angeles, 30 Sep embe .
Lane, P R and Milesi-Fe e i, G M (2018) ‘The Ex e nal Weal h o Na ions Re isi ed: In e na-
ional Financial In eg a ion in he A e ma h o he Global Financial C isis’. IMF Economic
Re iew Vol. 66(189), pp. 189–222.
Liu, L, Ma hes, C and Pe o a, K (2022) ‘Mone a y Policy ac oss Space and Time’. Ad ances
in Econome ics (Essays in Hono o Fabio Cano a) Vol. 44 Pa B.
Maćkowiak, B (2007) ‘Ex e nal Shocks, U.S. Mone a y Policy and Mac oeconomic Fluc ua ions
in Eme ging Ma ke s’. Jou nal o Mone a y Economics Vol. 54(8), pp. 2512–2520.
McKay, A and Wol , C (2023) ‘Wha Can Time Se ies Reg essions Tell Us Abou Policy Coun-
e ac uals?’ Econome ica.
Melosi, L (2017) ‘Signalling E ec s o Mone a y Policy’. The Re iew o Economic S udies Vol.
84(2), pp. 853–884.
Me ens, K and Ra n, M O (2013) ‘The Dynamic E ec s o Pe sonal and Co po a e Income Tax
Changes in he Uni ed S a es’. Ame ican Economic Re iew Vol. 103(4), pp. 1212–47.
Mi anda-Ag ippino, and Ricco, G (2023) ‘Iden i ica ion wi h Ex e nal Ins umen s in S uc u al
VARs’. Jou nal o Mone a y Economics Vol. 135, pp. 1–19.
33
Mi anda-Ag ippino, S and Rey, H (2020) ‘US Mone a y Policy and he Global Financial Cycle’.
The Re iew o Economic S udies Vol. 87(6), pp. 2754–2776.
Mi anda-Ag ippino, S and Ricco, G (2021) ‘The T ansmission o Mone a y Policy Shocks’.
Ame ican Economic Jou nal: Mac oeconomics Vol. 13(3), pp. 74–107.
Mon iel Olea, J L, S ock, J H, Wa son, M W e al. (2021) ‘In e ence in S uc u al Vec o
Au o eg essions Iden i ied wi h an Ex e nal Ins umen ’. Jou nal o Econome ics Vol. 225(1),
pp. 74–87.
Mum az, H and Pe o a, K (2023) ‘Algo i hms o P oxy SVAR models wi h ime- a ying pa a-
me e s’. Jou nal o Money, C edi and Banking Vol. 55((2-3)), pp. 635–654.
Nakamu a, E and S einsson, J (2018) ‘Iden i ica ion in Mac oeconomics’. Jou nal o Economic
Pe spec i es Vol. 32(3), pp. 59–86.
Noh, E (2019) ‘Impulse- esponse analysis wi h p oxy a iables’. Wo king Pape .
Obs eld, M (2020) ‘Global Dimensions o U.S. Mone a y Policy’. In e na ional Jou nal o Cen al
Banking, pp. 73–132.
Omo i, Y, Chib, S, Shepha d, N and Nakajima, J (2007) ‘S ochas ic Vola ili y wi h Le e age:
Fas and E icien Likelihood In e ence’. Jou nal o Econome ics Vol. 140(2), pp. 425–449.
Paul, P (2020) ‘The Time-Va ying E ec o Mone a y Policy on Asse P ices’. Re iew o Economics
and S a is ics Vol. 102(4), pp. 690–704.
P imice i, G E (2005) ‘Time Va ying S uc u al Vec o Au o eg essions and Mone a y Policy’.
The Re iew o Economic S udies Vol. 72(3), pp. 821–852.
Rey, H (2015) Dilemma no T ilemma: The Global Financial Cycle and Mone a y Policy Independ-
ence. NBER Wo king Pape No. 2162.
Rey, H (2016) ‘In e na ional Channels o T ansmission o Mone a y Policy and he Mundellian
T ilemma’. IMF Economic Re iew Vol. 64(1), pp. 6–35.
Shin, H S (2012) ‘Global Banking Glu and Loan Risk P emium’. IMF Economic Re iew Vol.
60(2), pp. 155–192.
Sims, C A and Zha, T (2006) ‘Does Mone a y Policy Gene a e Recessions?’ Mac oeconomic
Dynamics Vol. 10(2), pp. 231–272.
S ock, J H and Wa son, M W (2012) Disen angling he Channels o he 2007-2009 Recession.
NBER Wo king Pape No. 18094.
34
Wu, J C and Xia, F D (2016) ‘Measu ing he Mac oeconomic Impac o Mone a y Policy a he
Ze o Lowe Bound’. Jou nal o Money, C edi and Banking Vol. 48(2-3), pp. 253–291.
35
A. Da a
All se ies we e ga he ed om he sou ces lis ed below, pa icula ly he FRED and se e al pape s.
In Table A1, we de ine he exac ans o ma ions o he a iables used in he es ima ion o he
TVP-VAR. No e ha we use mon h-on-mon h g ow h a es.
In speci ic cases (indica ed by s a s ∗in he Table A1), we a e missing da a poin s in he p e-
sample pe iod. In o de o es ima e he TVP-VAR on he same es ima ion sample ( o which all da a
poin s a e a ailable), we pu sue he ollowing s a egy. I we ha e in o ma ion on a simila se ies
o he same ime pe iod, we assume he same g ow h scena io. Speci ically, we obse e indus ial
p oduc ion o EMEs only om 1987M1 o 2017M12. Hence, we assume ha indus ial p oduc ion
in EMEs g ows as he co esponding se ies in he RoW du ing he ime ame 1980M1 o 1986M12.
This allows us o ex end he se ies backwa ds.
36
Table A1: Va iable De ini ions.
Va iable T ans o ma ion in TVP-VAR De ails
𝑓 𝑓 𝑟𝑡FFR𝑡FFR𝑡is he ede al unds a e (F ed), eplaced
wi h he shadow a e o Wu and Xia (2016)
du ing he ZLB pe iod (2008M12-2015M12).
𝑐𝑝𝑖𝑈𝑆
𝑡100 ×hln CPI𝑈𝑆
𝑡−ln 𝐶𝑃𝐼𝑈𝑆
𝑡−1i CPI𝑈𝑆
𝑡is U.S. consume p ice index (F ed).
𝑖𝑝𝑈𝑆
𝑡100 ×hln IP𝑈𝑆
𝑡−ln IP𝑈𝑆
𝑡−1i IP𝑈𝑆
𝑡is U. S. indus ial p oduc ion (F ed).
𝑥𝑖𝑚𝑈𝑆
𝑡XIM𝑈𝑆
𝑡XIM𝑈𝑆
𝑡is he a io be ween U.S. expo and
impo in goods and se ices (F ed), in e pol-
a ed om qua e ly o mon hly using a shape-
p ese ing piecewise cubic in e pola ion.
𝑒𝑏𝑝𝑈𝑆
𝑡EBP𝑈𝑆
𝑡EBP𝑈𝑆
𝑡is he U.S. excess bond p emium by
Gilch is and Zak ajšek (2012).
𝑔 𝑓 𝑐𝑡GFC𝑡GFC𝑡is he global inancial cycle indica o by
Mi anda-Ag ippino and Rey (2020).
𝑟𝑒𝑒𝑟𝑡100 ×ln REER𝑈𝑆
𝑡REER𝑈𝑆
𝑡is he U.S. eal e ec i e exchange a e
om he BIS.
𝑐𝑝𝑖𝑅𝑜𝑊
𝑡100 ×hln CPI𝑅𝑜𝑊
𝑡−ln CPI𝑅𝑜𝑊
𝑡−1i CPI𝑅𝑜𝑊
𝑡is RoW consume p ices (excluding
he U.S.) G ossman e al. (2014).
𝑖𝑝𝑅𝑜𝑊
𝑡100 ×hln IP𝑅𝑜𝑊
𝑡−ln IP𝑅𝑜𝑊
𝑡−1i IP𝑅𝑜𝑊
𝑡is he RoW indus ial p oduc ion (ex-
cluding he U.S.) by G ossman e al. (2014).
𝑚𝑠𝑐𝑖𝑅𝑜𝑊
𝑡100 ×hln MSCI𝑅𝑜𝑊
𝑡−ln MSCI𝑅𝑜𝑊
𝑡−1i MSCI𝑅𝑜𝑊
𝑡is RoW MSCI (excluding he U.S.).
𝑒𝑥𝑅𝑜𝑊
𝑡100 ×hln EX𝑅𝑜𝑊
𝑡−ln EX𝑅𝑜𝑊
𝑡−1i EX𝑅𝑜𝑊
𝑡is RoW expo s (excluding he U.S.)
by G ossman e al. (2014).
𝑖𝑝𝐴𝐸
𝑡100 ×ln IP𝐴𝐸
𝑡−ln IP𝐴𝐸
𝑡−1IP𝐴𝐸
𝑡is he ad anced economies’ indus ial
p oduc ion (excluding he U.S.) by G ossman
e al. (2014).
𝑒𝑥𝐴𝐸
𝑡100 ×ln EX𝐴𝐸
𝑡−ln EX𝐴𝐸
𝑡−1EX𝐴𝐸
𝑡is AE expo s (excluding he U.S.) by
G ossman e al. (2014).
𝑖𝑝𝐸 𝑀𝐸
𝑡100 ×ln IP𝐸 𝑀𝐸
𝑡−ln IP𝐸𝑀 𝐸
𝑡−1IP𝐸𝑀 𝐸
𝑡is he eme ging ma ke econom-
ies’ indus ial p oduc ion by G ossman e al.
(2014).∗
𝑒𝑥𝐸 𝑀 𝐸
𝑡100 ×ln EX𝐸 𝑀𝐸
𝑡−ln EX𝐸𝑀 𝐸
𝑡−1EX𝐸𝑀 𝐸
𝑡is he eme ging ma ke economies ex-
po s by G ossman e al. (2014).
No es: RoW is sho o es -o -wo ld. All a iables a e a ailable o e he p e-sample and es ima ion sample pe iod, anging om 1980M5 o 2017M12.
Excep ions is: EME IP𝑡1987M1 o 2017M12.
37
B. TVP-VAR: P io Se ings and Es ima ion
We desc ibe he de ails on he p io densi y choice and hype pa ame e calib a ion. We ha e o se
p io densi ies o he ini ial alues o 𝜽, which we deno e wi h 𝜽0. Fo all hese ini ial alues, we
speci y Gaussian dis ibu ions. We also ha e o speci y p io densi ies o he co a iance ma ices
𝑸, whe e we use an in e se-Wisha p io densi y.26 Las , we also need a p io dis ibu ion o
he co a iance ma ix o he VAR 𝚺, which is again ollowing an in e se-Wisha dis ibu ion. To
calib a e he p io dis ibu ions, we use he i s 13 yea s as a aining sample. This esul s in a
aining sample anging om 1980M1 o 1992M12 o leng h 𝜏=156. We ob ain es ima es using
O dina y Leas Squa es (OLS) om his aining sample and assume he ollowing p io densi ies
o he coe icien s in he TVP-VAR
𝜽0∼ N ˆ
𝜽𝑂𝐿𝑆,4∗𝑉(ˆ
𝜽𝑂𝐿𝑆),
𝑸∼𝑖𝑊 𝜅2
𝑄∗𝜏∗𝑉(ˆ
𝜽𝑂𝐿𝑆, 𝜏),
(B.1)
whe e he subsc ip 𝑂𝐿𝑆 e e s o he OLS es ima o o he espec i e coe icien . Fo he ini ial
alues, we use he OLS poin es ima es and ou imes i s a iance. Fo he co a iance ma ix 𝑸, he
scaling ma ix is chosen o be a ac ion o he co esponding OLS es ima es (mul iplied wi h he
co esponding deg ees o eedom). Finally, we ha e o choose a alue o he hype pa ame e 𝜅2
𝑄,
go e ning he ime- a ia ion in he s a e equa ion. In pa icula , we assume 𝜅2
𝑄=0.015 ( ollowing
Paul, 2020). We use a a he conse a i e alue o his hype pa ame e such ha he ime- a ia ion
is no in la ed by ou p io . An addi ional no e is in o de : We obse e he high- equency su p ises
no un il 1991M1. Following he p ocedu e o Paul (2020), we plug in ze os o he obse a ions
p io o his pe iod. This should no cause a bias in OLS as long as hose ze os a e om a andom
sample. This should be indeed he case o he mone a y policy su p ises. Ne e heless, we es ima e
hose coe icien s wi h less p ecision.
Las , we discuss he p io densi y on he co a iance ma ix 𝚺, which is de ined as ollows
𝚺∼𝑖𝑊 (𝑰𝑀, 𝑀 +1),
whe e he scaling ma ix is se o an iden i y ma ix and he deg ees o eedom a e se o 𝑀+1, as
ecommended by Ka lsson (2013).
Rega ding he es ima ion p ocedu e, we se up he Gibbs sample along he lines o Del Neg o
and P imice i (2015) o ob ain pos e io dis ibu ions. In pa icula , we use Kalman il e ing
echniques o ob ain he unobse able s a es in 𝜽𝑇=(𝜽′
1, . . . , 𝜽′
𝑇)′. This esul s in a Gaussian s a e
space model, whe e s anda d Bayesian me hods o he Kalman il e can be applied (Ca e and
Kohn, 1994; F ühwi h-Schna e , 1994). The emaining pos e io quan i ies a e a he s anda d
and in e ence is conduc ed ia an MCMC algo i hm.
26 We deno e by 𝑖𝑊 (𝑆, 𝑑)an in e se-Wisha dis ibu ion wi h deg ees o eedom 𝑑and scale ma ix 𝑆.
38
Figu e E3: Coun y-Speci ic Indus ial P oduc ion Impulse Responses.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in 1993M1.
Coun ies: Spain, Sou h A ica, B azil, Chile, Columbia, Mexico.
45
Figu e E4: Coun y-Speci ic Indus ial P oduc ion Impulse Responses.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in 1993M1.
Coun ies: Pe u, Sou h Ko ea, Malaysia, India.
46
F. Addi ional Resul s o he Sensi i i y Analysis
Figu e F1: Robus ness: Ze o Lowe Bound.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in 1993M1.
Baseline model in blue; ed lines co espond o he model excluding he ZLB pe iod (ending in 2008M11).
47
Figu e F2: Robus ness: S ochas ic Vola ili y.
No es: Peak Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in
1993M1. Baseline model s. ex ended model ha allows o s ochas ic ola ili y speci ica ion.
De ails on s ochas ic ola ili y: As in ou benchma k analysis and in P imice i (2005), we use OLS es ima es on a p e-sample (1980M1-1992M12) o
calib a e he p io dis ibu ions. In addi ion, a p io belie on he ex en o ime- a ia ion in 𝑨−1
𝑡and ℎ𝑖𝑡 (𝑖=1, . . . , 𝑛)mus be speci ied. Using he
no a ion o P imice i (2005), his boils down o a selec ion choice on h ee pa ame e s: 𝜅𝑄(go e ning ime- a ia ion o au o eg essi e coe icien s),
𝜅𝑊( a iance o he esiduals), 𝜅𝑆(co a iance o he esiduals). We se 𝜅𝑄=0.015 (as be o e), 𝜅𝑊=0.001, and 𝜅𝑆=0.001. The alue o 𝜅𝑊
is among he ones conside ed by P imice i (2005), while ou 𝜅𝑆is ela i ely igh e ( o a oid ill beha io s in ou ela i ely sho e sample). We
es ima e he s ochas ic ola ili y model as in Kim e al. (1998) bu wi h he e inemen o Omo i e al. (2007).
48
Figu e F3: Robus ness: Di e en Mone a y Policy Ins umen s.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in 1993M1.
Baseline model in blue (ins umen o Mi anda-Ag ippino and Ricco, 2021); ed lines co espond o he es ima es using he high- equency mone a y
policy ins umen s o Ge le and Ka adi (2015) (GK) and Ja ociński and Ka adi (2020) (JK).
Figu e F4: Robus ness: P io Calib a ion.
No es: Responses o a con ac iona y U.S. mone a y policy shock ha induces a one pe cen age poin inc ease in he ede al und a e in 1993M1.
Baseline model in blue; ed lines co esponds o he es ima es using a igh e (𝜅2
𝑄=0.01) and wide (𝜅2
𝑄=0.02) p io agains he baseline p io
(𝜅2
𝑄=0.015).
49
