Dynamic impact of foreign exchange trading volume on foreign exchange volatility

Kang, Jong Woo; Cabae o, Ca los
Wo king Pape
Dynamic impac o o eign exchange ading olume on
o eign exchange ola ili y
ADB Economics Wo king Pape Se ies, No. 768
P o ided in Coope a ion wi h:
Asian De elopmen Bank (ADB), Manila
Sugges ed Ci a ion: Kang, Jong Woo; Cabae o, Ca los (2025) : Dynamic impac o o eign exchange
ading olume on o eign exchange ola ili y, ADB Economics Wo king Pape Se ies, No. 768, Asian
De elopmen Bank (ADB), Manila,
h ps://doi.o g/10.22617/WPS250025-2
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/322304
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/
ASIAN DEVELOPMENT BANK
ASIAN DEVELOPMENT BANK
6 ADB A enue, Mandaluyong Ci y
1550 Me o Manila, Philippines
www.adb.o g
DYNAMIC IMPACT OF FOREIGN
EXCHANGE TRADING VOLUME
ON FOREIGN EXCHANGE
VOLATILITY
Jong Woo Kang and Ca los Cabae o
ADB ECONOMICS
WORKING PAPER SERIES
NO. 768
Feb ua y 2025
Dynamic Impac o Fo eign Exchange T ading Volume on Fo eign Exchange Vola ili y
Fo eign exchange (FX) ading olume is a key ac o in ola ili y. This pape in es iga es he e ec
o ading olume on ola ili y using high- equency da a. Es ima ion esul s om econome ic models
e eal a signi ican impac o hi d-pa y ade olumes on he ola ili ies o o iginal cu ency pai s.
Though he Uni ed S a es dolla (USD) exe s sizeable e ec h ough hi d-pa y channels, cu ency pai s
wi hou USD linkages also ha e impac , calling enewed a en ion o u ilizing egional coope a ion
in mi iga ing ola ili y as compa ed wi h majo FX ading pa ne s.
Abou he Asian De elopmen Bank
ADB is commi ed o achie ing a p ospe ous, inclusi e, esilien , and sus ainable Asia and he Paci ic,
while sus aining i s e o s o e adica e ex eme po e y. Es ablished in 1966, i is owned by 69 membe s
—49 om he egion. I s main ins umen s o helping i s de eloping membe coun ies a e policy dialogue,
loans, equi y in es men s, gua an ees, g an s, and echnical assis ance.
ASIAN DEVELOPMENT BANK
The ADB Economics Wo king Pape Se ies
p esen s esea ch in p og ess o elici commen s
and encou age deba e on de elopmen issues
in Asia and he Paci ic. The iews exp essed
a e hose o he au ho s and do no necessa ily
e lec he iews and policies o ADB o
i s Boa d o Go e no s o he go e nmen s
hey ep esen .
ADB Economics Wo king Pape Se ies
Dynamic Impac o Fo eign Exchange T ading Volume
on Fo eign Exchange Vola ili y
Jong Woo Kang and Ca los Cabae o
No. 768 | Feb ua y 2025
Jong Woo Kang ([email protected] g) is he di ec o o
Regional Coope a ion and In eg a ion Di ision and
Ca los Cabae o (ccabae o.consul an @adb.o g)
is a consul an a he Economic Resea ch
and De elopmen Impac Depa men , Asian
De elopmen Bank.
C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
© 2025 Asian De elopmen Bank
6 ADB A enue, Mandaluyong Ci y, 1550 Me o Manila, Philippines
Tel +63 2 8632 4444; Fax +63 2 8636 2444
www.adb.o g
Some igh s ese ed. Published in 2025.
ISSN 2313-6537 (p in ), 2313-6545 (PDF)
Publica ion S ock No. WPS250025-2
DOI: h p://dx.doi.o g/10.22617/WPS250025-2
The iews exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews and policies
o  he Asian De elopmen Bank (ADB) o i s Boa d o Go e no s o he go e nmen s hey ep esen .
ADB does no gua an ee he accu acy o he da a included in his publica ion and accep s no esponsibili y o any
consequence o hei use. The men ion o speci ic companies o p oduc s o manu ac u e s does no imply ha hey
a e endo sed o ecommended by ADB in p e e ence o o he s o a simila na u e ha a e no men ioned.
By making any designa ion o o e e ence o a pa icula e i o y o geog aphic a ea in his documen , ADB does no
in end o make any judgmen s as o he legal o o he s a us o any e i o y o a ea.
This publica ion is a ailable unde he C ea i e Commons A ibu ion 3.0 IGO license (CC BY 3.0 IGO)
h ps://c ea i ecommons.o g/licenses/by/3.0/igo/. By using he con en o his publica ion, you ag ee o be bound
by he e ms o his license. Fo a ibu ion, ansla ions, adap a ions, and pe missions, please ead he p o isions
and e ms o use a h ps://www.adb.o g/ e ms-use#openaccess.
This CC license does no apply o non-ADB copy igh ma e ials in his publica ion. I he ma e ial is a ibu ed
oano he sou ce, please con ac he copy igh owne o publishe o ha sou ce o pe mission o ep oduce i .
ADB canno be held liable o any claims ha a ise as a esul o you use o he ma e ial.
Please con ac pubsma ke [email protected] g i you ha e ques ions o commen s wi h espec o con en , o i you wish
oob ain copy igh pe mission o you in ended use ha does no all wi hin hese e ms, o o pe mission o use
heADB logo.
Co igenda o ADB publica ions may be ound a h p://www.adb.o g/publica ions/co igenda.
ABSTRACT
Fo eign exchange (FX) ading olume is a key ac o in exchange a e ola ili y. Gi en he
impo an ole o ola ili y in economic g ow h and s abili y, his pape in es iga es he dynamic
na u e o exchange ading olume on exchange a e ola ili y using hou ly high- equency da a.
The es ima ion esul s om o dina y leas squa es, ixed e ec s and he gene al au o eg essi e
condi ional he e oskedas ici y model poin o a signi ican impac o hi d-pa y o eign exchange
ade olumes on he FX ola ili ies o o iginal cu ency pai s. The Uni ed S a es dolla (USD), as
he dominan cu ency, exe s sizeable e ec h ough his hi d-pa y channel and he magni ude
o he o eign exchange ading olume u ns ou o be a c ucial ac o o his e ec . Howe e ,
hi d-pa y cu ency pai s wi hou USD linkages also exe non-negligible impac , calling o
enewed a en ion o he e ec i eness o egional inancial coope a ion in mi iga ing exchange
a e ola ili y as compa ed wi h majo o eign exchange ading pa ne s, no only h ough di ec
ansac ion mechanisms bu h ough hi d pa y cu ency channels.
Keywo ds: FX ola ili y, hi d pa y channel, GARCH model
JEL codes: F31, G15, G18

I. In oduc ion
The impac o exchange a e mo emen on economic g ow h, de elopmen , and s abili y is well-
documen ed. Much o he li e a u e posi s ha exchange a es a ec in e na ional ade
pe o mance, including h ough expo /impo compe i i eness, olumes, and p ices, as well as
businesses’ ex e nal inancing cos s. Consequen ly, changes in exchange a es a ec o eign and
domes ic consump ion, p oduc i i y, and in es men . Gi en hese signi ican impac s, p i a e and
public en i ies alike mus ca e ully moni o and p epa e o exchange a e mo emen s.
The li e a u e has shown he impac o exchange a es on mul iple de elopmen and
mac oeconomic indica o s. Eicheng een (2007) ames o eign exchange a es as a i al
acili a ing condi ion o economic g ow h. De elopmen expe iences in high-g ow h economies,
such as in Eas Asia, and de eloping economies demons a e ha compe i i e exchange a es
a e c i ical in jumps a ing g ow h. An e icien exchange a e mechanism encou ages e icien
edeploymen o esou ces o p oduc i e sec o s, hus unlocking gains in p oduc i i y. S udies
co e ing mul iple economies ac oss he globe a i m his, pa icula ly ha unde alua ion o a
cu ency agains o eign coun e pa s is o en accompanied by g oss domes ic p oduc (GDP)
g ow h (Rod ik 2008, Se aj and Coskune 2021). In a s udy ocusing on India, Shaik and Rao
(2020) s a e ha he dep ecia ion o he Indian upee is associa ed wi h an inc ease in he
coun y’s o eign exchange ese es and in eal GDP. Zhao (2020) shows ha exchange a es
ha e a di ec e ec on he p ices and cos s o commodi ies, also impac ing expo s o o eign
ma ke s. Fu he esea ch shows ha exchange a es impac p oduc i i y and o eign ou ism.
Fluc ua ions in he exchange a e may in luence economic policy, pa icula ly in economies
adop ing in la ion- a ge ing egimes, while ha ming economic g ow h.
Exchange a es ha e signi ican impac on in e na ional ade, as no ed. Exchange a es ha e a
sizable e ec on he impo and expo cos s o p oduc s. When an economy’s cu ency is alued
highe , p oduc s om o eign ma ke s become cheape , hus encou aging g ea e impo a ion.
Con e sely, unde alued cu encies acili a e lowe p ices o an economy’s commodi ies and
lead o g ea e p oduc expo a ion. Thus, mo emen s in he exchange a e also a ec an
economy’s ade balance; a highe exchange a e mo es owa ds a nega i e ade accoun
balance, while a lowe exchange a e leads o a posi i e balance, al hough he g owing
complexi ies associa ed wi h deepening global alue chains end o compound his linea
ela ionship. The s eng h o his ela ionship has been widely s udied in he li e a u e, wi h a ying
esul s. Resea ch on de eloped and de eloping coun ies (Kang 2016) a e he global inancial
c isis show ha he e ec o cu ency de alua ion on expo g ow h pos - c isis has no been as
2
s ong as be o e he c isis. Fu he esea ch posi s ha an inc ease in he numbe o economies
ha engage in deepe cu ency de alua ion may lead o sluggish g ow h in in e na ional ade.
Thus, g ea impo ance has been placed in unde s anding exchange a e mo emen s, pa icula ly
h ough he concep o exchange a e ola ili y. Exchange a e ola ili y is de ined as he isk
associa ed wi h unexpec ed mo emen s in exchange a es (Oz u k 2006). The ep esen a i e
indica o o exchange a e ola ili y is he deg ee o a iance in an exchange a e o e a ce ain
pe iod. P ice mo emen s o igina e om an e en s-based app oach, such as poli ical-economic
news and announcemen s, ha in o ms decisions o economic ac o s, bo h public and p i a e.
O he ac o s like compa a i e in la ion and in e es a e di e en ials be ween economies likewise
lead o exchange a e ola ili y.
Aside om mac oeconomics and in e na ional ade, exchange a e ola ili y is consequen ial o
i ms and ade s, which could be exposed o sizeable exchange a e isks in hei inancing cos s
and inancial managemen . Vola ile o eign exchange ma ke s lead o g ea e ma ke unce ain y,
which impac s cos s and e enues o i ms and in u n in o ms hedging and in es men s a egies.
Fu he mo e, a ola ile exchange a e inc eases unce ain y in he o eign exchange ma ke and
discou ages isk-a e se ade s om engaging in in es men s, leading o changes in in es o
po olio lows (Flo es-Sosa, A iles-Ochoa, and Me igo 2023).
Gi en he in luence o exchange a es on economic g ow h and ade, main aining a s able and
p edic able exchange a e has been a cons an p io i y o so e eign inancial au ho i ies ac oss
he globe. Bo h he sou ces and implica ions o exchange a e ola ili y ha e been key a eas o
economic esea ch. Findings gene ally show ha highe exchange a e ola ili y leads o highe
cos s o isk-a e se ade s, as well as lowe o eign ade. This is due o changes in he alue o
he exchange a e upon he ag eemen o a con ac e sus i s ac ual paymen and
implemen a ion. When exchange a es become ola ile, unce ain y ises in p edic ing he cos s
and he eby he p o i s om ansac ions, which disincen i izes ade (Hoope and Kolhagen
1978). Despi e his, o he esea che s ha e been less de ini e abou he impac o exchange a e
ola ili y on in e na ional ade. De G auwe (1988), o example, posi s ha dominance o income
e ec s o e subs i u ion e ec s may lead o a posi i e ela ionship be ween ola ili y and ade. In
his heo y, an inc ease in exchange a e ola ili y could aise he ma ginal u ili y o expo e enue
in he eyes o su icien ly isk-a e se expo e s and could induce inc eased expo s. De G auwe
hus sugges s ha he e ec o exchange a e ola ili y depends on he deg ee o isk a e sion o
ma ke playe s.
3
Nume ous e o s ha e been made o s udy exchange a e ola ili y. As i is no a di ec ly
obse able phenomenon, ex ensi e esea ch has been ca ied ou o p edic mo emen s in he
exchange a es, as well as o iden i y causes, and possible indica o s o exchange a e ola ili y.
The li e a u e iden i ies economic undamen als, such as in la ion, in e es a es, and balance o
paymen s as sou ces o exchange a e ola ili y, especially as hese ac o s hemsel es ha e
become mo e ola ile since he 1980s. Fu he mo e, ac o s such as capi al accoun libe aliza ion,
echnological inno a ion, and cu ency specula ion all con ibu e o inc eased c oss-bo de lows
and ade olumes, adding o exchange a e ola ili y (Hook and Boon 2000). One o he key
ac o s a ec ing exchange a e ola ili y is o eign exchange ading olume, which is used as a
measu e o he s a e o he o eign exchange ma ke . Fo eign exchange ading is essen ial o
engaging in in e na ional ade as i allows ade s and i ms alike o con e domes ic cu ency
in o o eign cu ency and ice e sa o be able o ansac wi h ex e nal ma ke s. T ade s o en
s udy mo emen in he o ex ma ke and ca y ou ades based on hei alua ion o a ious
cu encies o make a p o i . Thus, luc ua ions in o ex ading olume a e also used o s udy and
p edica e he deg ee o o eign exchange ola ili y be ween cu encies. This places a clea
impe us in unde s anding he ela ionship be ween o eign exchange ola ili y and ading
olumes. Fo eign exchange ades can also be in luenced by he i m’s mo i es o minimize losses
om exchange a e ola ili y in de e mining he iming o exchanges o o eign bo owing o
epaymen , and he incen i es o hedge agains exchange a e ola ili y isks.
The ela ionship be ween o eign exchange ola ili y and o eign exchange ading olumes has
been explo ed in he li e a u e. Fo eign exchange ading olume is ega ded as a p oxy o
unobse able ma ke condi ions, such as ela i e liquidi y and p i a ely in o med ading
(Ga gano, Riddiough, and Sa no 2018). Vola ili y ends o mo e in conjunc ion wi h ading
olumes, in ha a s eep inc ease in ading olumes o en coincides wi h mo e ola ile o eign
exchange cu encies (Figu e on page 4). Va ious heo e ical explana ions aim o explain his.
Copeland (1976, 1997) p esen ed he model o “sequen ial in o ma ion a i al,” whe ein ading
pa icipan s eac o in o ma ion on he inancial ma ke indi idually. Thei eac ion o he a i al o
he news he eby shi s hei demand cu e o a pa icula cu ency. These ades hen ac as a
p oxy o ade s’ changing demand o a pa icula cu ency, and hus coincide wi h inc eased
ola ili y in he o eign exchange ma ke . Ano he explana ion o his phenomenon is he “mix u e
o dis ibu ions hypo hesis” p oposed by Cla k (1973). Unde his heo y, ola ili y and olume a e
de e mined by a common, unobse able ac o ha e lec s he a i al o new in o ma ion in he
o eign exchange ading ma ke . How ade s in e nalize his in o ma ion changes he p icing o
a pa icula cu ency, hus encou aging a highe numbe o ades. These ades he e o e signi y
4
disag eemen s be ween ade s on he p icing o a pa icula o eign exchange cu ency and mo e
ola ile p ice mo emen .
Figu e: USD-JPY Vola ili y and T ading Volumes o May 2023
USD-JPY = US dolla -Japanese yen cu ency pai .
No es: Values abo e a e he daily a e age o he hou ly ola ili y a es. The hou ly ola ili y a es a e calcula ed as he
absolu e sum o he 5-minu e in e al p ice change o a cu ency pai wi hin an hou , e lec ed as a pe cen age o he
exchange a e.
Sou ces: Asian De elopmen Bank calcula ions using da a om Bloombe g and he CLS FX da abases.
Sensoy and Se dengec y (2019) explo e he iabili y o he Mixed Dis ibu ion hypo hesis by
in es iga ing he ela ionship be ween USD-TRY (Tu kish li a) ola ili y and o eign exchange
ade olumes by cu ency ade and coun e pa . Thei esul s used a gene alized me hod o
momen s amewo k o es ablish a posi i e con empo aneous ela ionship be ween USD-TRY
exchange ola ili y and ade olumes in he spo ma ke . The esea ch u he showed ha he
dispe sion o ade belie s on he u u e USD-TRY exchange a e signi ican ly inc eases he
posi i e ela ionship be ween ola ili y and ading olume, s eng hening he hypo hesis ha he
join mo emen o wo a iables a e explained by ade unce ain y and disag eemen in o eign
exchange a e p edic ions. Gala i (2000) likewise explo es he e ec o local cu encies in
de eloping economies using o dina y leas squa es (OLS) eg ession wi h a gene al
au o eg essi e condi ional he e oskedas ici y (GARCH) componen , dis inguishing expec ed and
unexpec ed changes in ade olume. The esul s showed a posi i e signi ican ela ionship
be ween o eign exchange ola ili y and olumes o ou ou o he six cu encies examined.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
2
4
6
8
10
12
Billions (USD-JPY)
A e age daily ola ili y
A e age o ola ili y A e age o olume, USD-JPY
11
ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖1 = 𝛽𝛽𝑜𝑜+ 𝛽𝛽1ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖1𝑡𝑡 + 𝛽𝛽2ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖2𝑡𝑡 + 𝛽𝛽3ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖3𝑡𝑡
+ 𝛽𝛽4𝑣𝑣𝑣𝑣𝑙𝑙𝑙𝑙𝑒𝑒𝑑𝑑𝑣𝑣𝑜𝑜𝑣𝑣𝑣𝑣𝑡𝑡𝑖𝑖𝑣𝑣𝑖𝑖𝑡𝑡𝑣𝑣𝑖𝑖1𝑡𝑡−1 +𝛽𝛽5𝑣𝑣𝑣𝑣𝑙𝑙𝑙𝑙𝑒𝑒𝑑𝑑𝑣𝑣𝑜𝑜𝑣𝑣𝑣𝑣𝑡𝑡𝑖𝑖𝑣𝑣𝑖𝑖𝑡𝑡𝑣𝑣𝑖𝑖1𝑡𝑡−2 + 𝛽𝛽6𝑣𝑣𝑣𝑣𝑒𝑒𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣 + 𝛽𝛽7𝑤𝑤𝑒𝑒𝑑𝑑𝑤𝑤𝑒𝑒𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣
+ 𝛽𝛽8𝑣𝑣ℎ𝑣𝑣𝑢𝑢𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣 + 𝛽𝛽9𝑓𝑓𝑢𝑢𝑣𝑣𝑑𝑑𝑣𝑣𝑣𝑣 +𝜀𝜀𝑖𝑖𝑡𝑡
whe e i and co espond o he cu ency pai and hou , espec i ely.
We hen use a ixed e ec s model, whe ein he hou o each ading day has an unobse ed e ec
on i s espec i e o eign exchange ola ili y. This model is used o obse e he ela ionship
be ween ola ili y and ading olume conside ing any e ec s b ough abou by he hou s wi hin a
ading day.
ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖𝑡𝑡 = 𝛽𝛽𝑜𝑜+ 𝛽𝛽1ln _𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖1𝑡𝑡 + 𝛽𝛽2ln_𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖2𝑡𝑡 + 𝛽𝛽3ln_𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑒𝑒𝑖𝑖3𝑡𝑡
+ 𝛽𝛽4𝑣𝑣𝑣𝑣𝑙𝑙𝑙𝑙𝑒𝑒𝑑𝑑_𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖1𝑡𝑡−1 +𝛽𝛽5𝑣𝑣𝑣𝑣𝑙𝑙𝑙𝑙𝑒𝑒𝑑𝑑_𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖1𝑡𝑡−2 + 𝛽𝛽6𝑣𝑣𝑣𝑣𝑒𝑒𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣
+ 𝛽𝛽7𝑤𝑤𝑒𝑒𝑑𝑑𝑤𝑤𝑒𝑒𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣 + 𝛽𝛽8𝑣𝑣ℎ𝑣𝑣𝑢𝑢𝑡𝑡𝑑𝑑𝑣𝑣𝑣𝑣 + 𝛽𝛽9𝑓𝑓𝑢𝑢𝑣𝑣𝑑𝑑𝑣𝑣𝑣𝑣 +𝛽𝛽10ℎ𝑣𝑣𝑣𝑣𝑢𝑢𝑡𝑡 + 𝜀𝜀𝑡𝑡
whe e i and co espond o he cu ency pai and hou , espec i ely.
We u he adop he model by Epaph a (2017), whe ein he ARCH and GARCH models a e used
o iden i y ola ili y clus e ed in o eign exchange a es. I is well-es ablished in he li e a u e ha
o eign exchange a es end o beha e like inancial da a, and hus a e also examinable by models
ha aim o accoun o ime- ela ed e ec s in ola ili y. Fo his analysis, he adop ed GARCH
model employs he ollowing es ima ion p ocedu e.
Fi s , ola ili y and ade olumes a e log-di e enced o sa is y he condi ion o non-s a iona i y,
as con i med by he Augmen ed Dickey Fulle -Tes . Second, he basic eg ession equa ion is i ed
as ollows o each cu ency pai :
𝑑𝑑𝑣𝑣𝑤𝑤_𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑖𝑖𝑡𝑡 = 𝛽𝛽𝑜𝑜+ 𝛽𝛽2dln_ ade ol𝑖𝑖1𝑡𝑡 + 𝛽𝛽3dln_ ade ol𝑖𝑖2𝑡𝑡 + 𝛽𝛽4dln_ ade ol𝑖𝑖3𝑡𝑡 +𝜀𝜀𝑖𝑖𝑡𝑡
whe e i and co espond o he cu ency pai and hou , espec i ely.
The model consis s o he log-di e enced ade olumes wi hin each ipa i e ela ionship as
eg esso s. Following he Box-Jenkins me hod o speci ying models, he co elog am,
Au oco ela ion (ACF) and Pa ial Au oco ela ion (PACF) o he basic eg essions a e es ed o
iden i y he app op ia e amoun o Au o eg essi e (AR) and Mo ing A e age (MA) componen s o
add o he equa ion. The ollowing es s a e i al o ensu e ha he e is no se ial au oco ela ion
in he ime se ies da a, ha is, ha he e a e no co ela ions be ween he e o e ms o di e en
pe iods wi hin he ime se ies. Au o eg essi e Mo ing A e age (ARMA) componen s a e added
and emo ed based on epea ed es ing o he ACF and PACF o he model. When he ARMA

12
model is sa is ied, he Au o eg essi e Condi ional He e oskedas ici y (ARCH) es is applied on
he squa ed esiduals o he model o see i he e a e ARCH e ec s in he da a. The p esence o
an ARCH e ec in he ARMA model would show ha he e is a endency o he da a o exhibi
pe iods o high ola ili y, hen ollowed by e en highe le els o ola ili y, and ice- e sa. This
means ha he model also needs o be con olled o i s condi ional a iance. Fo his analysis,
he GARCH (gene al ARCH) model is u ilized o cap u e his condi ional a iance. The GARCH
model is an ex ension o he ARCH model, which models he condi ional a iance o ime se ies
da a as a linea unc ion o pas squa ed obse a ions. GARCH builds upon his by allowing mo e
lexibili y by modelling he condi ional a iance as a linea unc ion o pas squa ed obse a ions,
as well as pas condi ional a iances. Thus, he GARCH model adds in bo h sho - e m and long-
e m memo y in he ola ili y analysis.
The s anda d GARCH (1,1) model is hen i ed o he equa ion, and he ARMA componen s a e
modi ied un il he esiduals a e no au oco ela ed h ough he ACF and PACF es . The inal
equa ion o he model is as ollows:
whe e i and espec i ely co espond o he cu ency pai and hou , x e e s o he numbe o
lags, z e e s o a whi e noise p ocess wi h mean ze o and a iance 1, and αo e e s o he cons an
in he a iance.
III. Es ima ion Resul s
3.1. OLS es ima ion
The OLS es ima ion i mly es ablishes he signi ican posi i e impac o a cu ency pai ’s ade
olumes on he cu ency pai ’s o eign exchange ola ili y. The model also demons a es he
po ency o USD- ela ed ade olume as a de e minan o ola ili y ac oss mul iple cu ency pai s.
Table 5 models o eign exchange ola ili y in he USD-JPY-AUD ila e al, which shows ha
bila e al exchange a e ola ili y is signi ican ly a ec ed by he ading olume o co esponding
cu ency pai s. A 1% inc ease in AUD-USD ading olume leads o a 0.3% inc ease in AUD-USD
13
exchange a e ola ili y, he same in AUD-JPY ading olumes leads o a 0.09% inc ease in he
ola ili y o co esponding cu ency exchange a e, and he same in USD-JPY leads o a 0.33%
inc ease in USD-JPY exchange a e ola ili y. AUD-JPY ola ili y, howe e , is also a ec ed by
ades o non-co esponding cu encies, i.e., he ading olumes o AUD-USD and USD-J P Y.
In e es ingly, AUD-JPY ola ili y is mo e s ongly in luenced by AUD-USD ading olume. These
phenomena a es o he c ucial ole o USD in a ec ing he exchange ola ili y o non-USD ela ed
cu ency pai s. While he 1 hou lagged dependen a iable shows signi ican , posi i e impac on
exchange a e ola ili y, he 2-hou lagged dependen a iable has nega i e impac , indica ing
sho -li ed pe sis ency in exchange a e ola ili y.
Table 6 likewise models he ola ili y in he USD-JPY-NZD ila e al. This model also p esen s
signi ican and posi i e impac s o co esponding ading olumes on he exchange a e ola ili y
o cu ency pai s simila o he esul s o USD-JPY-AUD ila e al. In his case, he impac o
ades o non-co esponding cu ency pai s u ns ou o be e en s onge , h ough USD linked
ade channels. Fo example, NZD-JPY ola ili y is mo e s ongly a ec ed by USD-JPY ading
olume han he ading olume o i s own co esponding cu ency pai (0.14% s. 0.09%). In bo h
ila e als, non-USD- ela ed cu ency pai ola ili ies a e mo e a ec ed by he ades o hi d
cu ency pai s, in pa icula USD- ela ed ones.
Table 5: Reg essions Resul s o Vola ili y, T ila e al Rela ionship (USD-JPY-AUD)
Reg esso s
Vola ili y
AUDUSD
AUDJPY
USDJPY
Cons an
-5.60
***
-3.22
***
-5.42
***
Volume (AUDUSD)
0.30
***
0.10
**
-0.02
Volume (AUDJPY)
0.06
0.09
**
0.03
Volume (USDJPY)
0.01
0.07
**
0.33
***
Tuesday
0.01
0.04
-0.01
Wednesday
0.04
0.05
-0.01
Thu sday
0.07
*
0.16
***
0.04
F iday
0.01
0.03
-0.02
Lagged Dependen Va iable (1 hou )
0.20
***
0.21
***
0.29
***
Lagged Dependen Va iable (2 hou s)
-0.09
***
-0.11
***
-0.11
***
R-squa ed
0.60
0.51
0.58
No es: * o signi icance below 10%; ** o signi icance below 5%; *** o signi icance below 1%.
Sou ces: Au ho s’ calcula ions using da a om he Bloombe g and CLS FX Da abase
14
Table 6: Reg essions o Vola ili y, by T ila e al Rela ionship (USD-JPY-NZD)
Reg esso s
Vola ili y
NZDJPY
NZDUSD
USDJPY
Cons an
-4.36
***
-6.58
***
-6.86
***
Volume (NZDJPY)
0.09
***
0.12
***
0.10
Volume (NZDUSD)
0.09
***
0.15
***
0.03
Volume (USDJPY)
0.14
***
0.14
***
0.37
***
Tuesday
0.04
0.02
0.01
Wednesday
0.1
**
0.11
-0.00
Thu sday
0.15
***
0.88
**
0.05
F iday
0.04
0.05
-0.22
Lagged Dependen Va iable (1 hou )
0.15
***
0.10
**
0.20
***
Lagged Dependen Va iable (2 hou s)
-0.11
***
-0.08
***
-0.12
***
R-squa ed
0.58
0.58
0.61
No es: * o signi icance below 10%; ** o signi icance below 5%; *** o signi icance below 1%
Sou ces: Au ho s’ calcula ions using da a om Bloombe g and CLS FX Da abase.
The OLS models om he wo ila e al ela ionships a i m he s eng h o he USD- ela ed ade
lows in in luencing o eign exchange ola ili ies, and show he possibili y o a hi d egional
cu ency impac ing exchange a e ola ili y. Ano he poin wo h no ing is ha he size o he
o eign exchange ading ma ke , especially in compa ison o o he cu ency pai s wi hin each
ila e al ela ionship, could d i e he s eng h o he e ec o a pa icula cu ency ading olume.
As an example, he USD-JPY ading ma ke is signi ican ly la ge han NZD-JPY and NZD-USD
ma ke s, which leads o he USD-JPY ading olume ha ing he s onges e ec on cu ency pai
ola ili ies among he h ee bila e al pai s. The ac ha non-USD cu ency pai s AUD-JPY and
NZD-JPY show di e en dynamics, in ha he o me is a ec ed s onges by USD-AUD ades
while he la e is a ec ed s onges by USD-JPY ades, also pa ly e lec s he ela i ely bigge
size o he USD-AUD exchange ma ke han he USD-NZD.
3.2. Fixed e ec es ima ion
The abo e indings a e u he suppo ed by he ime- ixed e ec model es ima ions o bo h
ila e al ela ionships. Table 7 summa izes he impac o ading olumes on he USD-JPY-AUD
ila e al (see Appendix 2 o he ull model). The esul s show ha he ola ili y o each cu ency
pai is a ec ed signi ican ly by he ading olume o co esponding cu ency pai s. The inc ease
in USD-JPY ading olume accoun s o a 0.29% inc ease in USD-JPY exchange a e ola ili y.
A he same ime, howe e , he ola ili y o coun y pai exchange a es is also signi ican ly
a ec ed by he ade o non-co esponding coun y pai s. This hi d cu ency ading channel
wo ks s ongly o he AUD-JPY ola ili y, which is mo e s ongly a ec ed by USD-AUD ading
15
olume han by i s own co esponding ading olume. This esul is consis en wi h OLS
es ima ion esul s. Mo eo e , he inclusion o he ime- ixed e ec a iable b oadly es ablishes a
signi ican ela ionship be ween ola ili y and speci ic hou s wi hin he ading day, al hough his
is no as p e alen in he case o USD-JPY ola ili y. Fu he mo e, he 2-hou lagged ola ili y
u ns ou o be no longe signi ican while he posi i e impac o 1-hou lagged ola ili y is la gely
main ained, and he impac o ading in he middle o he week is only signi ican in one o he
h ee cu ency pai s.
Table 7: Fixed-E ec Reg essions on USD-JPY-AUD T ila e al
Cu ency
T ading Volumes
USDJPY (Majo Pai )
USDAUD (Majo Pai )
AUDJPY (Mino Pai )
USDJPY
0.29%***
0.08%**
0.01%
USDAUD
0.06%***
0.33%***
0.03%
AUDJPY
0.11%***
0.17%***
0.09%***
No es: * o signi icance below 10%; ** o signi icance below 5%; *** o signi icance below 1%.
Majo pai e e s o cu ency pai s in ol ing widely used cu encies pai ed wi h he USD. Mino pai s e e o widely
used cu encies, excluding USD.
Sou ces: Au ho s’ calcula ions using da a om Bloombe g and CLS FX Da abase.
Somewha di e en esul s a e es ima ed o he case o USD-JPY-NZD, as illus a ed by Table 8
(see Appendix 3 o he ull model). Exchange a e ola ili y is a ec ed by he co esponding
cu ency pai s signi ican ly. Howe e , he magni ude o impac u ns ou o be s onges om he
non-co esponding cu ency pai s excep o USD-JPY ola ili y. Consis en wi h OLS es ima ion
esul s, NZD-JPY ola ili y seems o be mo e s ongly a ec ed h ough he USD-linked hi d
cu ency ading channel, i.e., USD-JPY in ou model. Di e en om OLS es ima ion esul s, USD-
NZD ola ili y is also mo e s ongly a ec ed by he USD-linked hi d cu ency ading channel, i.e.,
USD-JPY ins ead o i s own co esponding pai . USD-JPY exchange a e ola ili y is s ill a ec ed
by USD-JPY ading olume he mos , wi h 1% inc ease in USD-JPY ading olumes leading o
0.43% inc ease in USD-JPY exchange a e ola ili y. In he case o USD-JPY ola ili y, his di ec
e ec is dominan , and he indi ec impac o o he cu ency pai s u n ou o be insigni ican . The
ime- ixed e ec a iable in he USD-JPY-NZD ila e al a e mos ly signi ican ly co ela ed wi h
ola ili y, al hough i is wo h no ing ha no all hou s ha e a signi ican e ec on USD-JPY
ola ili y. The inclusion o he ime- ixed e ec also ende s he 2-hou lagged ola ili y as
insigni ican , while in ading day a iables, Wednesdays and Thu sdays a e signi ican o NZD-
JPY ola ili y, and Wednesdays o NZD-USD ola ili y.
16
Table 8: Fixed-E ec Reg ession on USD-JPY-NZD T ila e al
Cu ency
T ading Volumes
USDJPY (Majo Pai )
USDNZD (Majo Pai )
NZDJPY (Mino Pai )
USDJPY
0.43%***
0.01%
0.002%
USDNZD
0.17%***
0.13***
0.10***
NZDJPY
0.22%***
0.09***
0.08%***
No es: * o signi icance below 10%; ** o signi icance below 5%; *** o signi icance below 1%
Sou ces: Au ho s’ calcula ions using da a om Bloombe g and CLS FX Da abase.
O e all, he weake in luence o NZD- ela ed ades on he ola ili y o hei own co esponding
cu ency pai may be a ibu ed o he AUD-USD and AUD-JPY ading ma ke s being la ge han
he NZD-JPY and NZD-USD ma ke s. AUD-USD olume, o example, is second only o USD-
JPY ading olumes in e ms o size. Rega dless, he ixed e ec models show ha hi d cu ency
pai ading olumes can ha e an impac on exchange a e ola ili y. While his hi d-pa y cu ency
pai channel is s onge h ough USD ela ed ades, non-USD ela ed hi d pa y cu ency ade
can also exe signi ican impac on he exchange ola ili y, as he es ima ion esul s indica e.
3.3. GARCH es ima ion
GARCH analysis is used o u he explo e he e ec s o ading olumes on o eign exchange
ola ili y, con olling o ime-speci ic componen s.
The di e enced ime se ies a e hen con e ed in o log a iables and i ed wi h hei espec i e
log-di e enced ading olumes o each ipa i e ela ionship, as well as an ARMA componen o
con ol o ime-speci ic p ope ies. An ARCH es is hen pe o med o in es iga e possible
he e oskedas ici y wi hin each ime se ies. Table 9 summa izes he esul s o he ARCH es on
each cu ency ola ili y. The es shows ha o eign exchange ola ili ies o he AUD-USD, AUD-
JPY and NZD-JPY ARMA models show he e oskedas ici y, and hus necessi a e he addi ion o
an ARCH componen o cap u e u he ime e ec s. In con as , NZD-USD and USD-JPY o bo h
AUS and NZD ipa i e ARMA models a e al eady su icien ly explained by he ARMA model, and
do no necessa ily need an addi ional ARCH componen . Ne e heless, o analyze he models on
a uni o m g ound, he GARCH (1,1) componen is adop ed all models.

17
Table 9: ARCH Tes on Cu ency Vola ili ies (24 lags)
Vola ili y
F s a is ic
Obs R2
P ob F(24,478)
P ob Chi2 (24)
AUDUSD
3.56
76.21
0
0
AUDJPY
2.71
60.35
0
0
NZDUSD
0.56
13.78
0.96
0.95
NZDJPY
1.65
38.32
0.03
0.03
USDJPY (AUD)
0.69
17
0.86
0.85
USDJPY (NZD)
0.68
16.6
0.87
0.86
Sou ces: Au ho s’ calcula ions using da a om Bloombe g and CLS FX Da abase.
The indi idual GARCH (1,1) models o each cu ency pai all show ha mos cu ency ade
olumes exe a posi i e signi ican e ec on he espec i e cu ency ola ili ies. The ime-speci ic
e ec s a y pe model.
Resul s om he USD-JPY-AUD ipa i e analysis (see Table 10) show ha USD-AUD and AUD-
JPY ola ili ies all ha e posi i e co ela ions wi h he a ious cu ency pai ading olumes. In he
case o AUD-USD, he ading olume wi h he s onges e ec is i s own pai (0.33), ollowed by
USD-JPY (0.10) and AUD-JPY (0.05). Likewise, AUD-JPY ola ili y is mos in luenced by USD-
JPY (0.19), ollowed by AUD-JPY (0.11) and USD-AUD (0.08). In con as wi h hese cu ency
pai s, USD-JPY ola ili y is only in luenced by he USD-JPY ading olume (0.38). I is no ewo hy
ha hese ela ionships hold e en wi h he inclusion o ime se ies a iables in he model. All h ee
o he ola ili y pai s a e posi i ely co ela ed wi h he 24-hou AR componen a simila le els,
which sugges ha ola ili y alues end o be abo e he mean ola ili y i i s 24-hou lagged alue
was also abo e he mean. Since he coe icien s egis e a a ound 0.33–0.34, he pas alue’s
in luence on p esen alues, hough signi ican , a e no e y s ong. The 1-hou MA componen
o all h ee ola ili ies a e signi ican , la ge and nega i ely co ela ed wi h p esen ola ili y alues.
The MA (1) coe icien s o AUD-USD and AUD-JPY a e bo h highe in magni ude han in he
USD-JPY; his means ha he e o e ms o 1-hou lagged alues end o exe a s onge
in luence on he p eceding wo cu ency pai s. A nega i e co ela ion sugges s ha a highe e o
e m o shock om he p e ious hou ends o pull cu en ola ili y alue lowe (and ice e sa).
The MA (2) componen s a e also nega i ely co ela ed wi h cu en ola ili y alues, hough he
magni ude o hei coe icien s is much lowe . This implies ha he shocks o e o e ms om 1-
hou lagged alues ha e a s onge in luence o e cu en ola ili y han 2-hou lagged alues.
The MA (2) coe icien magni ude o USD-JPY ola ili y is also ma ginally highe han he o he
pai s, sugges ing ha he USD-JPY cu ency is mo e in luenced by i s pas 2-hou e o e ms.
The ARCH and GARCH componen s o he model o e an idea o whe he ola ili y alues a e
18
mo e in luenced by sho - e m o long- e m shocks. The ARCH (1) coe icien s o USD-JPY and
AUD-USD a e bo h insigni ican , while he magni ude o AUD-JPY’s, hough signi ican , is e y
small. This sugges s ha ha sho - e m shocks do no ha e a pe sis en e ec on cu en ola ili y
alues o all cu ency pai s. Meanwhile, GARCH (1) componen s a e all signi ican and la ge in
magni ude ac oss all cu ency pai s, which implies ha olde ola ili y alues ha e a mo e long-
e m and pe sis en e ec on cu en ola ili y.
Table 10: GARCH Model Es ima ions o USD-JPY-AUD
Reg esso s
Vola ili y
AUDUSD
AUDJPY
USDJPY
Cons an
-0.00
***
-0.00
***
-0.00
**
Volume (AUDUSD)
0.33
***
0.08
***
0.02
Volume (AUDJPY)
0.05
***
0.11
***
0.03
Volume (USDJPY)
0.10
***
0.19
***
0.38
***
AR (24)
0.35
***
0.34
***
0.34
***
MA (1)
-0.75
***
-0.76
***
-0.68
***
MA (2)
-0.19
***
-0.14
***
-0.29
***
ARCH (1)
-0.02
-0.01
*
0.00
GARCH (1)
0.85
***
0.87
***
0.92
***
R-squa ed
0.53
0.47
0.56
Aike in o c i e ion
-0.03
-0.19
0.13
No es: 24-hou lagged ola ili y alues a e in loga i hmic o m, while all ade olume a iables a e in loga i hmic o m,
di e enced by 1 hou . The ARMA model o ARMA-GARCH model o AR (24) MA (1) MA (2) ARCH (1) GARCH (1) a e
used o all models.
Sou ces: Au ho s’ calcula ions using da a om Bloombe g and CLS FX Da abase.
Resul s om he USD-JPY-NZD ipa y (see Table 11) e lec a ied pa e ns om he p e ious
ipa i e analysis. NZD-USD ola ili y is a ec ed by all ading olumes. The magni ude o he
alues is mo e e en han he cu encies in he p e ious analysis, whe ein NZD-USD ola ili y is
mos a ec ed by i s own olume (0.16), ollowed by USD-JPY (0.14), and NZD-JPY (0.10). NZD-
JPY ola ili y is also in luenced by all ading olumes, wi h he la ges in luence om USD-JPY
(0.21), ollowed by USD-NZD (0.10), and NZD-JPY (0.06). USD-JPY ola ili y is hea ily in luenced
by i s own ading olume (0.31), hen dis an ly ollowed by NZD-JPY (0.05). The AR (1)
componen o NZD-JPY is posi i ely signi ican a a low magni ude (0.24), which means ha i s
ola ili y is mildly in luenced by i s 1-hou lagged alue. The AR (24) coe icien s o all ola ili ies
a e likewise posi i ely signi ican , whe ein USD-JPY (0.31) and USD-NZD (0.35) ha e highe
magni ude han NZD-JPY (0.24). This would show ha he 24-hou lagged ola ili ies o he
p eceding cu encies ha e a sligh ly la ge impac on cu en alues. The MA (1) coe icien s o
all h ee cu encies a e nega i ely signi ican . The analysis shows ha he MA (1) componen o
19
NZD-JPY is highes (-0.99), ollowed by NZD-USD (-0.89), and USD-JPY (-0.61). This would imply
ha cu en ola ili y alues a e a iedly in luenced by immedia e pas e o e ms and shocks.
The MA (2) componen s a e also nega i ely co ela ed o USD-NZD and USD-JPY bu ha e
as ly di e en magni udes. The analysis would show ha shocks om a 2-hou window s ill ha e
a s ong e ec on USD-NZD bu ha e signi ican ly ape ed o USD-JP Y. The ARCH and GARCH
componen s a e bo h signi ican o NZD-USD, wi h a la ge magni ude o he GARCH coe icien
implying ha much p e ious ola ili y alues ha e a mo e pe sis en e ec on cu en alues. In
compa ison, only he ARCH coe icien is signi ican o USD-JPY and NZD-JPY, which means
ha hese ola ili y alues a e mo e in luenced by ecen shocks.
Table 11: GARCH Model Es ima ions o USD-JPY-NZD
Reg esso s
Vola ili y
NZDJPY
NZDUSD
USDJPY
Cons an
-0.00
-0.00
**
-0.00
Volume (NZDJPY)
0.06
***
0.10
***
0.05
***
Volume (NZDUSD)
0.10
***
0.16
***
0.01
Volume (USDJPY)
0.21
***
0.14
***
0.31
***
AR (1)
0.24
***
NA
NA
AR (24)
0.24
***
0.35
***
0.31
***
MA (1)
-0.99
***
-0.89
***
-0.61
***
MA (2)
NA
-0.8
***
-0.34
***
ARCH Residual
0.23
**
-0.01
***
0.10
**
GARCH
-0.03
0.58
**
0.10
R-squa ed
0.46
0.54
0.58
Aike in o c i e ion
-0.01
0.23
0.24
No es: 24-hou lagged ola ili y alues a e in loga i hmic o m, while all ade olume a iables a e in loga i hmic o m,
di e enced by 1 hou . NA implies ha a pa icula ARMA componen was omi ed in he analysis o he cu ency pai .
The omission o hese componen s o pa icula models was done o ensu e ha he indi idual model exhibi ed no
se ial au oco ela ion o p o ide a clea e analysis.
Sou ces: Au ho s’ calcula ions using da a om he Bloombe g and CLS FX Da abase.
While he GARCH analysis o bo h ipa i e ela ionships shows ha ime- ela ed coe icien s
ha e a ying e ec s on each indi idual ime-se ies, a pa e n eme ges ha , o mos o he
cu ency pai s, all ading olumes ha e a posi i e and signi ican co ela ion wi h hei ading
olumes, con olling o hese ime elemen s (see Tables 12 and 13). While NZD-JPY and AUD-
JPY ola ili y a e bo h mo e in luenced by USD-JPY ading olumes, o he cu ency pai s such
as USD-AUD and USD-NZD a e mos s ongly de e mined by hei own cu ency pai ola ili ies.
In all hese ou cases, he hi d-pa y cu ency also has a signi ican in luence on hei espec i e
ola ili ies. As expec ed, USD-JPY ola ili y is in luenced mos ly by i s own ading olume; in he
USD-JPY-AUD ipa i e, i is in ac he sole in luence o USD-JPY ola ili y. The USD-JPY-NZD
20
ipa i e, howe e , sugges s ha e en small cu ency pai s like NZD-JPY also ha e a signi ican ,
i minu e, e ec on USD-JPY ola ili y.
Table 12: Pe cen age E ec s o FX ola ili y, by Fo eign Exchange T ading Volume Pai
(USD, JPY, AUD)
Cu ency
T ading Volumes
USDJPY (Majo Pai )
USDAUD (Majo Pai )
AUDJPY (Mino Pai )
USDJPY
0.38%***
0.02%
0.03%
USDAUD
0.10%***
0.33%***
0.05%***
AUDJPY
0.19%***
0.08%***
0.11%***
No es: Da a is aken om GARCH (1,1) eg essions on ma ching each indi idual cu ency wi h ading olume pai s.
*** deno es signi icance a 0.01, ** deno es signi icance a 0.05, * deno es signi icance a 0.10. Majo pai e e s o
cu ency pai s in ol ing widely used cu encies pai ed wi h he USD. Mino pai s e e o widely used cu encies,
excluding USD.
Sou ces: Au ho s’ calcula ions om da a om Bloombe g and CLS Da abase.
Table 13: Pe cen age E ec s o Fo eign Exchange Vola ili y by T ading Volume Pai
(USD, JPY, NZD)
Cu ency
T ading Volumes
USDJPY (Majo Pai )
USDNZD (Majo Pai )
NZDJPY (Mino Pai )
USDJPY
0.31%***
0.01%
0.05%***
USDNZD
0.14%***
0.16***
0.10***
NZDJPY
0.21%***
0.10***
0.06%***
No es: Da a is aken om GARCH (1,1) eg essions on ma ching each indi idual cu ency wi h ading olume pai s.
*** deno es signi icance a 0.01, ** deno es signi icance a 0.05, * deno es signi icance a 0.10
Sou ces: Asian De elopmen Bank calcula ions om da a om Bloombe g and CLS Da abase.
IV. Conclusions
US dolla dominance as a global ese e cu ency has signi ican implica ions o he impac o
USD- ela ed cu ency ade olumes on exchange a e ola ili y. This pape has in es iga ed
whe he hi d-pa y o eign exchange ade olumes can ha e any signi ican e ec on he o eign
exchange ola ili ies o o iginal cu ency pai s despi e he c ucial ole o he US dolla as he
dominan cu ency. While he s udy demons a es he signi ican e ec o exchange ade olumes
on he ola ili y o he co esponding cu ency pai s, hi d pa y cu ency ade olumes also exe
signi ican impac on he ola ili y o he co esponding cu ency pai s. This hi d pa y ading
channel is qui e s ong o he USD- ela ed cu ency ades. Howe e , non-USD- ela ed hi d pa y
ades also signi ican ly a ec he ola ili y o co esponding cu ency pai s. These indings a e
in es iga ed h ough h ee ypes o models ha explo e he ela ionship be ween o eign
exchange ola ili y and ade olumes. The OLS eg essions poin ou ha , indi idually, mos
cu ency ola ili ies a e a ec ed by all cu ency olumes in he ipa i e ela ionships, including
he indi ec , hi d-pa y cu ency ading channels. The ime- ixed e ec model adds ha , e en
27
Appendix 3: Time-Fixed E ec Model o USD-JPY-NZD T ila e al
(NZDJPY)
(NZDUSD)
(USDJPY)
VARIABLES
Vola ili y
Vola ili y
Vola ili y
Volume (NZDJPY)
0.0782***
0.102***
0.00283
(0.0140)
(0.0175)
(0.0150)
Volume (NZDUSD)
0.0909***
0.132***
0.0118
(0.0151)
(0.0189)
(0.0164)
Volume (USDJPY)
0.219***
0.168***
0.433***
(0.0238)
(0.0296)
(0.0277)
Lagged Dependen Va iable (1 hou )
0.111***
0.118***
0.195***
(0.0376)
(0.0379)
(0.0375)
Lagged Dependen Va iable (2 hou s)
0.0239
0.0212
0.00274
(0.0352)
(0.0358)
(0.0351)
F iday
0.000770
0.0184
-0.0400
(0.0317)
(0.0397)
(0.0344)
Thu sday
0.0820**
0.0399
-0.00216
(0.0324)
(0.0397)
(0.0347)
Tuesday
-0.00363
-0.00785
-0.0208
(0.0288)
(0.0359)
(0.0314)
Wednesday
0.0509*
0.0667*
-0.0323
(0.0300)
(0.0376)
(0.0324)
1.hou
-0.181**
-0.579***
0.476***
(0.0721)
(0.0918)
(0.0744)
2.hou
-0.384***
-0.665***
0.0732
(0.0703)
(0.0880)
(0.0778)
3.hou
-0.453***
-0.604***
-0.0361
(0.0800)
(0.100)
(0.0880)
4.hou
-0.455***
-0.512***
-0.312***
(0.0814)
(0.102)
(0.0881)
5.hou
-0.569***
-0.669***
-0.430***
(0.0786)
(0.0993)
(0.0845)
6.hou
-0.624***
-0.737***
-0.359***
(0.0752)
(0.0952)
(0.0809)
7.hou
-0.681***
-0.825***
-0.250***
(0.0735)
(0.0928)
(0.0802)
8.hou
-0.578***
-0.687***
-0.181**
(0.0742)
(0.0930)
(0.0818)
9.hou
-0.439***
-0.472***
-0.123
(0.0802)
(0.100)
(0.0880)
10.hou
-0.387***
-0.539***
-0.0399
(0.0813)
(0.103)
(0.0877)
11.hou
-0.405***
-0.528***
-0.222***
(0.0807)
(0.102)
(0.0859)
12.hou
-0.544***
-0.635***
-0.163**
(0.0775)
(0.0974)
(0.0824)
13.hou
-0.532***
-0.645***
-0.144*
(0.0747)
(0.0943)
(0.0801)
14.hou
-0.503***
-0.529***
-0.0699
(0.0757)
(0.0954)
(0.0823)
15.hou
-0.487***
-0.462***
0.130
(0.0839)
(0.105)
(0.0917)
Con inued on he nex page

28
(NZDJPY)
(NZDUSD)
(USDJPY)
VARIABLES
Vola ili y
Vola ili y
Vola ili y
16.hou
-0.501***
-0.568***
-0.150
(0.0855)
(0.108)
(0.0937)
17.hou
-0.522***
-0.618***
-0.188*
(0.0901)
(0.114)
(0.0982)
18.hou
-0.741***
-0.842***
-0.400***
(0.0863)
(0.109)
(0.0927)
19.hou
-0.621***
-0.822***
-0.251***
(0.0785)
(0.0992)
(0.0846)
20.hou
-0.557***
-0.729***
-0.126
(0.0733)
(0.0923)
(0.0794)
21.hou
-0.508***
-0.638***
-0.103
(0.0717)
(0.0899)
(0.0779)
22.hou
-0.430***
-0.602***
-0.194**
(0.0727)
(0.0913)
(0.0786)
23.hou
-0.677***
-0.885***
-0.385***
(0.0724)
(0.0905)
(0.0790)
Cons an
-5.397***
-5.509***
-7.882***
(0.455)
(0.569)
(0.512)
Obse a ions
546
546
546
R-squa ed
0.696
0.687
0.752
No es: S anda d e o s in pa en heses *** p<0.01, ** p<0.05, * p<0.1
Sou ces: ADB calcula ions using Bloombe g and CLS da abase.
29
REFERENCES
Cla k. (1973). A subo dina ed s ochas ic p ocess model wi h ini e a iance o specula i e.
Econome ica.
Copeland. (1976). “A model o asse ading unde he assump ion o sequen ial in o ma ion
a i al”. Jou nal o Finance, 1149-1168.
Copeland. (1997). “A p obabili y model o asse ading”. Jou nal o Financial and Quan i a i e
Analysis.
De G auwe, P. (1988). Exchange a e a iabili y and he slowdown in g ow h o in e na ional
ade. Re ie ed om In e na ional Mone a y Fund: h ps://www.js o .o g/s able/3867277
Eicheng een, B. (2007). The Real Exchange Ra e and Economic G ow h. Social and Economic
S udies.
Epaph a, M. (2017). Modeling exchange a e ola ili y: Applica ion o he GARCH and EGARCH
models. Jou nal o Ma hema ical Finanace.
Flo es-Sosa, M., A iles-Ochoa, E., and Me igo , J. (2020). Exchange a e and ola ili y: A
bibliome ic e iew. Re ie ed om Wiley Online Lib a y:
h ps://onlinelib a y.wiley.com/doi/epd /10.1002/ij e.2223
Gala i, G. (2000). T ading olumes, ola ili y and sp eads in FX ma ke s:. Bank o In e na ional
Se lemen s.
Ga gano , A., Riddiough, S., and Sa no, L. (2018). The alue o olume in o eign exchange.
e ie ed om h p://www.olsenda a.com/da a_p oduc s/clien _pape s/pape s/201804-
An onioGa gano-Valueo Volume.pd
Hoope , P. and Kohlhagen, S e en. (1978). The e ec o exchange a e unce ain y on he
p ices and olume o in e na ional ade. Jou nal o In e na ional Economics, 483-511.
Kang, J. W. (2016). In e na ional T ade and Exchange Ra e. Re ie ed om Asian De elomen
Bank: h ps://www.adb.o g/publica ions/in e na ional- ade-and-exchange- a e
Khemi i, R. (2012). Volume and ola ili y in o eign exchange ma ke mic os uc u e: a Ma ko
swi ching app oach.
Me cado, R., Jacildo, R., and Basu Das, S. (2022). US dolla dominance in Asia’s ade
in oicing. Asian De elopmen Bank.
Mougoue, M., and Agga wal, R. (2011). T ading olume and exchange a e ola ili y: E idence
o he sequen ial a i al o in o ma ion hypo hesis. Jou nal o Banking and Finance,
2690-2703.
Oz u k, I. (2006). Exchange Ra e Vola ili y and T ade: A Li e a u e Su ey. In e na ional Jou nal
o Applied Econome ics and Quan i a i e S udies, 85-102.
Pa k, D., Tian, S., Qu eshi, I., and Villa uel, M. (2022). Impac o Mone a y Policy Unce ain y .
Asian De elopmen Bank.
30
Rajbongshi, G., and Su esh, P. S. (2023). Exchange a e ola ili y: A e iew o global
expe ience. IUP Jou nal o Applied Economic, 52-72.
Rod ik, D. (2008). The eal exchange a e and economic g ow h. Ha a d Uni e si y. Re ie ed
om h ps://muse.jhu.edu/pub/11/a icle/261232/summa y
S.L., H., and Boon, H. (2000). Real exchange a e ola ili y and Malaysian expo i s majo
ading pa ne s. Uni e si y Pu a Malaysia.
Sensoy, A., and Se dengec i, S. (2019). In aday olume- ola ili y nexus in he FX ma ke s:
E idence om an eme ging ma ke . In e na ional Re iew o Financial Analysis.
Se aj, M., and Coskune , C. (2021). Real exchange a e e ec on economic g ow h: compa ison
o undamen al equilib ium exchange a e and Balassa–Samuelson based Rod ik
app oach. Jou nal o Applied Economics, 541-554. Re ie ed om
h ps://www. and online.com/doi/ ull/10.1080/15140326.2021.1977083
Shaik, K., and Rao, B. (2020). Does exchange a e ha e any impac on economic g ow h in
India? An empi ical analysis. Theo e ical and Applied Economics, 223-234. Re ie ed
om h ps://s o e.ec ap. o/a icole/1485.pd
Vienna Ini ia i e. (2024). Vienna Ini ia i e. Re ie ed om Vienna Ini ia i e: h ps://www. ienna-
ini ia i e.com/abou /
Zhao, Y. (2020, Janua y 2020). The in luence and impac o he exchange a e on he economy.
Re ie ed om
h ps://www. esea chga e.ne /publica ion/347413240_The_In luence_and_Impac _o _ he
_Exchange_Ra e_on_ he_Economy
Zhao, Y. (2020). The in luence and impac o he exchange a e on he economy . Re ie ed
om E3S Web o Con e ences: h ps://www.e3s-
con e ences.o g/a icles/e3scon /pd /2020/74/e3scon _ebldm2020_03007.pd
ASIAN DEVELOPMENT BANK
ASIAN DEVELOPMENT BANK
6 ADB A enue, Mandaluyong Ci y
1550 Me o Manila, Philippines
www.adb.o g
DYNAMIC IMPACT OF FOREIGN
EXCHANGE TRADING VOLUME
ON FOREIGN EXCHANGE
VOLATILITY
Jong Woo Kang and Ca los Cabae o
ADB ECONOMICS
WORKING PAPER SERIES
NO. 768
Feb ua y 2025
Dynamic Impac o Fo eign Exchange T ading Volume on Fo eign Exchange Vola ili y
Fo eign exchange (FX) ading olume is a key ac o in ola ili y. This pape in es iga es he e ec
o ading olume on ola ili y using high- equency da a. Es ima ion esul s om econome ic models
e eal a signi ican impac o hi d-pa y ade olumes on he ola ili ies o o iginal cu ency pai s.
Though he Uni ed S a es dolla (USD) exe s sizeable e ec h ough hi d-pa y channels, cu ency pai s
wi hou USD linkages also ha e impac , calling enewed a en ion o u ilizing egional coope a ion
in mi iga ing ola ili y as compa ed wi h majo FX ading pa ne s.
Abou he Asian De elopmen Bank
ADB is commi ed o achie ing a p ospe ous, inclusi e, esilien , and sus ainable Asia and he Paci ic,
while sus aining i s e o s o e adica e ex eme po e y. Es ablished in 1966, i is owned by 69 membe s
—49 om he egion. I s main ins umen s o helping i s de eloping membe coun ies a e policy dialogue,
loans, equi y in es men s, gua an ees, g an s, and echnical assis ance.