Ünlü, E im Zeynep
A icle
T ade in he ace o e o : Examining Tu key's expo
dynamics
Peace Economics, Peace Science and Public Policy (PEPS)
P o ided in Coope a ion wi h:
De G uy e B ill
Sugges ed Ci a ion: Ünlü, E im Zeynep (2025) : T ade in he ace o e o : Examining Tu key's expo
dynamics, Peace Economics, Peace Science and Public Policy (PEPS), ISSN 1554-8597, De G uy e ,
Be lin, Vol. 31, Iss. 1, pp. 91-122,
h ps://doi.o g/10.1515/peps-2024-0023
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/333339
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/4.0/
E im Zeynep Ünlü*
T ade in he Face o Te o : Examining
Tu key’s Expo Dynamics
h ps://doi.o g/10.1515/peps-2024-0023
Recei ed May 8, 2024; accep ed No embe 20, 2024; published online Decembe 13, 2024
Abs ac : In he p eceding 11 yea s, Tu key has wi nessed 925 e o is inciden s,
esul ing in he unexpec ed passing o 1,439 people. Concu en ly, he neighbou ing
na ions, Sy ia, and I aq, ha e also expe ienced a mul i ude o e o is ac i i ies. This
s udy’s main objec i e is o explo e how hese e en s ha e affec ed Tu key’s expo s
and in a-indus y ade. Thein es iga ion is cen ed onTu key’s ade ela ionships
wi h 71 coun ies, ep esen ing i s p ima y ading pa ne s, employing a one-way
ade model. U ilizing he Poisson pseudo maximum likelihood (PPML) es ima o , he
s udy spans 2012 o 2022. The findings unde sco e he ad e se impac o domes ic
e o is inciden s, casual ies in o he coun ies, and he global e o ism index on
Tu key’s expo s. Su p isingly, he in es iga ion finds a posi i e connec ion be ween
Tu key’s expo s and e o is inciden s in bo h neighbou ing and non-bo de ing
coun ies. Fu he mo e, he empi ical analysis sheds ligh on Tu key’s engagemen
in in a-indus y ade conce ning manu ac u ing indus y and GDP simila i ies.
Howe e , he scena io changes o in o ma ion and communica ion echnology (ICT)
expo s, e ealing an in e - ype o indus y ade.
Keywo ds: empi ical s udies o ade; in e na ional conflic s; poli ical economy
JEL Classifica ion: F14; F51; F50
1 In oduc ion
Te o ism, as defined by Ende s and Sandle (2000), is he in en ional use o deadly
iolence o h ea s o u he pa isan goals, ins illing ea o in imida ion among a
sizable po ion o he popula ion. A key elemen o his defini ion is he poli ical
mo i e behind e o is ac ions, dis inguishing hem om o dina y c iminal ac i -
i ies ha lack such mo i es. Addi ionally, e o ism is cha ac e ized using ex eme
iolence, o en in ended o a ac media a en ion.
*Co esponding au ho : E im Zeynep Ünlü, Independen Resea che , Is anbul, Tü kiye,
E-mail: [email p o ec ed]. h ps://o cid.o g/0000-0002-2045-5702
Peace Econ. Peace Sci. Pub. Pol. 2025; 31(1): 91–122
Open Access. © 2024 he au ho (s), published by De G uy e . This wo k is licensed unde he
C ea i e Commons A ibu ion 4.0 In e na ional License.
Te o ism has long been ecognized as a significan dis up o o economic s a-
bili y and g ow h. Globally, e o is ac i i ies ha e been shown o unde mine
in es o confidence, educe o eign di ec in es men , and dis up ade flows,
leading o b oade economic ins abili y (Abadie and Ga deazabal 2003; Gaibulloe
and Sandle 2009, 2019). The economic consequences o e o ism ex end a beyond
immedia e des uc ion. Te o ism os e s a clima e o unce ain y, leading o
beha iou al changes among economic agen s. This can esul in educed in es men ,
dec eased consume spending, and dis up ions o in e na ional ade. Addi ionally,
he inc eased cos s associa ed wi h e o ism, such as highe shipping cos s and
insu ance p emiums, can be passed on o consume s, u he impac ing economic
ac i i y (Bandyopadhyay 2017).
The ela ionship be ween e o ism and in e na ional ade has been o
pa icula in e es in ecen yea s, wi h s udies indica ing ha e o ism can
significan ly educe ade flows, dis up supply chains, and lead o he ealloca ion o
ade ou es (Bandyopadhyay 2016). These effec s a e o en compounded in egions
wi h high le els o geopoli ical ins abili y, whe e e o ism wo sens exis ing
economic ulne abili ies (Bandyopadhyay, Sandle , and Younas 2014).
Te o ism no only imposes di ec human cos s bu also gene a es significan
economic dis up ions, pa icula ly h ough i s impac on in e na ional ade.
Resea ch has shown ha e o ism can c ea e spillo e effec s, whe e a e o is
e en in one coun y ad e sely affec s he bila e al ade o neighbou ing na ions
(Pham and Doucouliagos 2017). These spillo e effec s a e long-las ing, o en pe -
sis ing o up o fi e yea s a e he e en and can a ise e en om inciden s wi h a
small dea h oll. The mechanisms h ough which e o ism impac s ade include
inc eased ade cos s due o s ic e secu i y measu es, heigh ened egula o y bu -
dens, and psychological dis ess among economic agen s (Pham and Doucouliagos
2017).
Gi en Tu key’s geopoli ical posi ion, unde s anding hese spillo e effec s is
i al. The po en ial inc ease in ade cos s and esul ing ic ions a e pa icula ly
ele an o Tu key, which sha es bo de s wi h se e al coun ies ha ha e expe-
ienced significan e o is ac i i y. This s udy aims o build on he exis ing li e a-
u e by examining how e o ism, bo h wi hin Tu key and in neighbou ing
coun ies, affec s Tu key’s ade dynamics, wi h a specific ocus on in a-indus y
ade.
Al hough he economic impac s o e o ism a e well-documen ed, esea ch
specifically examining i s effec on in a-indus y ade, especially in de eloping
economies like Tu key, emains limi ed. Tu key’s unique posi ion, ac oss bo h
Eu ope and he Middle Eas , and i s expe ience wi h bo h domes ic and ansna ional
e o ism, make i a c i ical case s udy o unde s anding hese dynamics (Bilgel and
Ka ahasan 2019).
92 E. Z. Ünlü
While some s udies examined he effec s o e o is inciden s on diffe en
a eas, o ins ance, on he Tu kish s ock ma ke (Aksoy and Demi alay 2019; Gok,
Demi dogen, and Topuz 2020), on o eign di ec in es men (FDI) (Çe in, Kese , and
Ay 2019; Ib ahim and A i 2021) and on domes ic bank loans (Ila slan and Yildiz 2022).
To he bes o my knowledge, no s udy has explo ed he ela ionship be ween ade,
in a-indus y ade, and e o ism, especially ega ding Tu key.
In a-indus y ade allows wo ke s and businesses o imp o e hei skills and
expe ise in specific p oduc s, encou aging inno a ion and le e aging economies o
scale. Consequen ly, a close analysis o how e o ism impac s ade, pa icula ly
in a-indus y ade, is essen ial o unde s anding he unique challenges ha could
affec Tu key.
In a-indus y ade allows wo ke s and businesses o imp o e hei skills and
expe ise in specific p oduc s, encou aging inno a ion and le e aging economies o
scale. Consequen ly, a close analysis o how e o ism impac s ade, pa icula ly
in a-indus y ade, is essen ial o unde s anding he unique challenges ha could
affec Tu key. Despi e he significance o in a-indus y ade, e y ew s udies ha e
employed g a i y models o examine his ype o ade, such as (Lei ão, Faus ino, and
Yoshida 2010; Wakasugi 2007; Yoshida, Lei ão, and Faus ino 2009). Howe e , hese
s udies did no ocus on he impac o e o ism, pa icula ly in de eloping econo-
mies like Tu key. This gap in he li e a u e u he unde sco es he impo ance o he
p esen s udy.
To achie e he esea ch goals, six unique one-way g a i y models we e imple-
men ed, employing e o ism da a sou ced om he Ins i u e o Economics and
Peace (IEP) (2023). These models we e designed o add ess he ollowing inqui ies: (i)
Wha impac do he global e o index (GTI), he equency o e o ism inciden s,
and he numbe o a ali ies among Tu key’s ading pa ne s ha e on Tu key’s
expo s? (ii) In ligh o e o is a acks, how do Tu key’s expo s be impac ed by he
economic simila i ies wi h i s ade pa ne s? (iii) Wha effec does Tu key’s
manu ac u ing indus y simila i ies wi h hose o he coun ies i ades wi h ha e
on Tu key’s expo s in he case o e o is inciden s? (i ) How does e o ism affec
Tu key’s in a-indus y ade, in oducing a no el a iable- ela i e endowmen
(RELEN) o in o ma ion and communica ion echnology (ICT) se ice expo s
compa ed o popula ion size o cap u e he in a-indus y ade con ex ?
This s udy has wo majo con ibu ions o he li e a u e. Fi s , i conside s h ee
dis inc e o ism a iables: he global e o index (GTI), he numbe o inciden s
(INC), and he numbe o a ali ies (FTL) in each inciden . Second, he s udy add esses
he e o - ade ela ionship wi h a one-way g a i y model ailo ed o Tu key and i s
op 71 expo des ina ion coun ies.
The s udy is a anged as ollows: Sec ion 2 p o ides comp ehensi e defini ions
ele an o he esea ch con ex . Sec ion 3 includes an in-dep h e iew o he
T ade and Te o : Tu key’s Expo Dynamics 93
li e a u e, cu en esea ch, and ele an hypo heses abou he ela ionship be ween
in e na ional ade and e o ism. Sec ion 4 ou lines he empi ical s a egy and
explain da a sou ces. Sec ion 5 is de o ed o he model es ima ion p ocess and he
p esen a ion o he es ima ion esul s. Sec ion 6 discusses he findings in he con ex
o exis ing li e a u e, while Sec ion 7 summa izes he key insigh s and sugges s
a enues o u u e esea ch.
2 Theo e ical F amewo k
The g a i y model d aws i s heo e ical basis om New on’s law o uni e sal g a -
i a ion. This model p oposes ha he olume o ade be ween wo coun ies is
di ec ly ela ed o hei economic sizes and in e sely ela ed o he dis ance be ween
hem. This idea o igina es om Tinbe gen (1962), who applied New on’s g a i y
heo y; he model ma hema ically exp esses he o al ade (TT) be ween coun ies ‘i’
and ‘j’as a unc ion o hei g oss domes ic p oduc s (GDP), dis ance, and an e o
e m.
This can be exp essed ma hema ically,
TTij =β0⎛
⎝GDPβ1
iGDPβ2
j
Dβ3
ij ⎞
⎠εij (1)
While he heo e ical ounda ion o he g a i y model is widely accep ed, mode n
applica ions ha e inco po a ed se e al adjus men s o add ess economic and
econome ic challenges, pa icula ly hose aised by Ande son and Wincoop (2003)
in ela ion o Mul ila e al Resis ance Te ms (MRT). MRT e e s o he concep ha
ade be ween any wo coun ies depends no only on bila e al ac o s (such as
dis ance o a iffs) bu also on hei ade ela ionships wi h all o he coun ies. This
in oduces he need o ime- a ying impo e and expo e fixed effec s o con ol
o he unobse ed mul ila e al esis ance affec ing ade flows. In s anda d g a i y
models, ailu e o accoun o hese MRTs can lead o biased es ima es.
To add ess he econome ic challenges p esen ed by Sil a and Ten ey o (2006),
specifically, he issue o he e oskedas ici y and ze o ade flows – he Poisson pseudo
maximum likelihood (PPML) es ima o has become a widely adop ed solu ion. Unlike
he adi ional log-linea iza ion o he g a i y equa ion, which can p oduce incon-
sis en es ima es when he e a e ze o ade flows, he PPML es ima o handles he
mul iplica i e na u e o he ade model in i s o iginal o m, mi iga ing biases caused
by ze o ade obse a ions and he e oskedas ici y. The au ho employed he Poisson
pseudo maximum likelihood high dimensional fixed effec s (PPMLHDFE) command,
as de eloped by Co eia, Guima ães, and Zylkin (2020), o i s abili y o handle
94 E. Z. Ünlü
high-dimensional fixed effec s efficien ly. This S a a command builds upon he
Poisson pseudo maximum likelihood (PPML) es ima o and enhances compu a ional
pe o mance, pa icula ly when dealing wi h la ge da ase s and nume ous dummy
a iables.
The p ima y mo i a ion o employing PPML es ima o in his s udy is i s abili y
o effec i ely add ess bo h he ze o- ade flow p oblem and he nume ous dummy
a iable p oblem, while also accoun ing o mul ila e al esis ance ia impo e
fixed effec s. The decision o exclude expo e fixed effec s in his case is jus ified by
he ac ha Tu key is he sole expo e in he model, hus nega ing he need o such
effec s, as he e is no a ia ion in Tu key’s cha ac e is ics ac oss obse a ions.
Gi en he p esence o la ge dependen a iable alues (i.e. ade da a epo edin
millions), he PPML es ima o is p e e ed o e al e na i es like he Pooled O dina y
Leas Squa es (POLS) o Fixed Effec s (FE) es ima o s. POLS, o ins ance, ails o
accoun o he e oskedas ici y, while FE app oaches may in oduce collinea i y
issues when mul iple dummy a iables a e included.
The model specifica ion is exp essed as ollows:
TTij =exp[β0+β1ln(GDPi)+β2ln(GDPj)−β3ln(Dis anceij)+μij]εij (2)
In his o mula ion, he inclusion o high-dimensional fixedeffec s–such as impo e
fixed effec s –add esses mul ila e al esis ance and con ols o ime- a ying ac o s
ha affec ade flows. Impo an ly, he model’s abili y o handle he ze o- ade issue
and mi iga e po en ial biases s emming om he e oskedas ici y makes PPML he
app op ia e choice.
3 Li e a u e Re iew
Sandle and Ende s (2008, 17–47) unde sco ed he di e se economic ins and ou s o
e o ism and iden ified how a ge ed na ions bea financial bu dens. These impli-
ca ions span ade es ic ions, in as uc u e des uc ion, he di e sion o Fo eign
Di ec In es men (FDI), and changes in public in es men owa d secu i y. No e-
wo hy esea ch has s udied he impac o e o ism on a ious economic a iables,
including i s impac on FDI (A i , Rawa , and Khan 2020; Dimi o a, T iki, and Val-
en ino 2022; Lanoua and Shahzad 2021; Polyxeni and Theodo e 2019). A comp e-
hensi e unde s anding o hese connec ions is essen ial o assessing he b oade
impac on he g oss domes ic p oduc (GDP) and o e all economic g ow h.
Simul aneously, e o is a acks impose nume ous cos s on na ions, causing
angible damages o c i ical in as uc u e and p ope y, as well as in angible losses
like human li es, labou , and inc eased insu ance isk p emiums. The consequences
T ade and Te o : Tu key’s Expo Dynamics 95
ex end o educed business p ofi abili y, a decline in ou is isi s, and an inc eased
demand o enhanced secu i y measu es. These mul i ace ed economic implica ions
highligh he necessi y o comp ehending he in e connec edness o e o ism and
a ious economic sides. Fo example, Bilgel and Ka ahasan (2017) in es iga ed he
impac o e o ism on Tu kish expo s. They e ealed ha be ween 1988 and 2001
when he s udy was conduc ed, income pe pe son in he coun y’s eas e n egions
declined by 6.6 %.
While nume ous s udies ha e explo ed he ela ionships be ween GDP, impo s,
expo s, FDI, and e o ism, none comp ehensi ely unde s and he dynamics su -
ounding ade in he same indus y o he impac s o e o ism. Haq, Ullah, and
Iqbal (2018) s udied he effec s o e o inciden s on bila e al ade, u ilizing a
g a i y model ac oss 50 de eloping and de eloped na ions wi h in ense e o
inciden s om 1990 o 2013. Thei esea ch e ealed ha e o ism exe s an
un a ou able effec on impo s and expo s, pa icula ly when de eloping na ions
ade wi h weal hie coun e pa s.
Simila ly, Ba onchelli, Ca uso, and Ricciu i (2022) used a g a i y model o
in es iga e he impac o a ms emba goes on he ade o Small A ms and Ligh
Weapons (SALW) om 1990 o 2017. Thei findings showed ha while o al
emba goes educed SALW ade, EU emba goes we e mo e effec i e han UN em-
ba goes, which had no measu able impac . This s udy demons a es how ex e nal
es ic ions, akin o he p essu es e o ism exe s on ade flows, can be effec i ely
analysed using g a i y models.
Bandyopadhyay, Sandle , and Younas (2016) ex ended his in es iga ion o
include he analysis o e o inciden s’effec s on bila e al ade. Thei findings
indica ed ha compa ed o domes ic e o ism, e o is ac i i ies ab oad ha e a
mo e p onounced nega i e impac on o e all manu ac u ed ade, including
expo s, impo s, and o he ca ego ies. In e es ingly, his nega i e effec is mo e
pe cei able on impo s han on expo s.
Chaudhu y and Sinha (2019) examined he effec s o e o is ac i i ies on GDP
and in e na ional comme ce by using a panel ec o au o eg essi e (VAR) model o
examine 44 indus ialized and de eloping na ions be ween 1991 and 2015. Thei
findings demons a ed ha while e o a acks ha e a mo e significan nega i e
impac on o eign ade in unde de eloped coun ies, hei influence is less
no iceable in ich coun ies.
Julius, Azu, and Muhammad (2019) con ibu ed o he discou se by examining
he implica ions o e o inciden s on comme ce wi hin he Sou h A ican De el-
opmen Communi y (SADC) coun ies, u ilizing he Poisson pseudo maximum like-
lihood (PPML) es ima o . Thei findings sugges ed ha e o ism has an ex emely
low effec on bila e al ade, wi h low- and middle-income coun ies expe iencing
96 E. Z. Ünlü
posi i e effec s. In con as , high-income coun ies obse e nega i e impac s om
e o ac i i ies.
In a depa u e om he di ec impac o e o ism on ade, Oka o and Piesse
(2018) shi ed he ocus owa ds unde s anding he ac o s con ibu ing o e o ism.
Thei examina ion spanned om 2005 o 2014, in ol ing 38 coun ies om a ious
egions wi h he highes anks on he F agile S a es Index (FSI). Thei esul s indi-
ca ed ha ac o s such as he p opo ion o e ugees, you h unemploymen , and
coun ies wi h uns able economies all posi i ely influence e o ism. Addi ionally,
o eign di ec in es men (FDI) was iden ified as ha ing an ad e se effec on
e o ism.
Pham and Nguyen (2024) in es iga ed he impac o e o is inciden s in one
coun y on a neighbou ing coun y. Acco ding o he esul s o panel da a analysis
wi h mo e han 160 coun ies, ade wi h coun ies neighbou ing he coun y wi h
e o is inciden s dec eased, and his effec was mainly in he Sub-Saha an A ica
egion.
Meie ieks and Schneide (2021) examined how e o ism affec ed global eco-
nomic policy o 170 na ions. They indica e ha go e nmen s eac o e o ism by
adop ing mo e es ic i e o eign economic policies o p e en he o ma ion and
unding o e o ism, es ic capi al fligh , s abilize s a e finances, and demons a e
poli ical de e mina ion. Mo e impo an ly, hey a gue ha he impac o e o ism is
especially c i ical o smalle na ions.
Some schola s s udied he GDP- e o ela ionship wi h diffe en econome ic
me hodologies. Ende s, Hoo e , and Sandle (2016), in hei s udy, es ablished a
Lo enz cu e o e o ism o demons a e he highe concen a ion o bo h
domes ic and in e na ional e o is s ikes in middle-income na ions, poin ing o a
nonlinea link be ween income and e o ism. Thei s udy is significan in he sense
ha hey assume he ela ionship o be nonlinea . They also employed di e se da a
om The In e na ional Te o ism: A ibu es o Te o is E en s (ITERATE) and he
Global Te o ism Da abase (GTD). They also shed ligh on po e y and e o ism
li e a u e and answe ed ques ions abou whe he po e y s imula es e o ism.
Se e al egional s udies ha e e ealed he link be ween economic indica o s
and e o ism. Sezgin (2019) de ised a no el index o Tu key spanning he yea s
1970–2016, showing he g ow h in ade olume ela i e o he g ow h o e o
inciden s. Despi e a lack o empi ical e idence, he s udy challenges he p e ailing
no ion ha e o is a acks nega i ely impac in e na ional ade. No ably, Sezgin
(2019) in es iga ion ailed o iden i y any subs an ial e idence suppo ing he
nega i e influence o e o ism on Tu key’s ade.
Ruiz Es ada, Pa k, and Khan (2018) in es iga ed how Tu key’s economic pe -
o mance was impac ed by e o ism. Employing a e o is a ack ulne abili y
e alua ion model co e ing e o inciden s, GDP dis ibu ion, and income
T ade and Te o : Tu key’s Expo Dynamics 97
dis ibu ion, hey a gued ha economic decline escala ed be ween 1990 and 2016.
Addi ionally, Bilgel and Ka ahasan (2019) ocused on he PKK e o g oup’sinflu-
ence on he Tu kish economy, u ilizing he syn he ic con ol me hod o he pe iod
1955–2008. Thei findings sugges ed ha Tu key’s GDP pe capi a could ha e been
13.8 % highe i no o he exposu e o PKK e o ism.
While exis ing s udies ha e ex ensi ely explo ed he connec ions be ween
e o ism and b oade economic indica o s, such as GDP, impo s, expo s, and
o eign di ec in es men , he e is a gap in he li e a u e conce ning he nuanced
effec o e o ism on in a-indus y ade. Limi ed schola ly a en ion has been
gi en o unde s anding how e o ism influences he in a-indus y ade pa e ns o
na ions.
This s udy aims o add ess he significan gap in he unde s anding o he link
be ween e o and in a-indus y ade by explo ing p e iously undisco e ed
a eas. This s udy is he fi s o ocus on his issue, offe ing a dis inc i e iewpoin on
he financial effec s o e o ism. By analysing he da a h ough he lens o he p oxy
ela i e endowmen (RELEN) calcula ed om ICT se ice expo s pe capi a and
in oducing no el a iables such as GDP simila i y and manu ac u ing ou pu sim-
ila i y, he aim is o en ich he unde s anding o he in e ac ion be ween e o ism
and he b oade economic s uc u e, albei a a b oade le el hanspecific indus ies.
This app oach con ibu es o he exis ing ade- e o ela ionship li e a u e
and sheds ligh on he po en ial implica ions o policymaking. This s udy seeks o
b oaden he discou se on e o ism’s economic impac , pa icula ly in in a-indus y
ade dynamics, offe ing esh insigh s ex ending beyond he cu en scope o
schola ly in es iga ion.
Helpman and K ugman (1987) p oposed using diffe ences in GDP pe capi a
among coun ies and economic simila i y as p oxies o analysing in a-indus y
ade. Subsequen s udies ha e adop ed a simila app oach; employing hese
a iablesasp oxies ounde s andin a-indus y dynamics. Some examples
include (An onucci and Manzocchi 2006; Ka acan and Ko kmaz 2022; Ülengin e al.
2015).
The li e a u e e iew p o ides in o ma ion on complex ela ionships be-
ween e o ism, economic indica o s, and ade dynamics. While exis ing s udies
ha e made significan addi ions in unde s anding he mac oeconomic impac s o
e o ism, a no able gap emains in explo ing he ade in he same indus y
specifically. This wo k aims o close his gap by ocusing on Tu key’sexpo s o71
coun ies, employing one-way g a i y model and con ibu ing no el insigh s in o
he in e play be ween e o ism and ade, pa icula ly a he in a-indus y
le el.
98 E. Z. Ünlü
Ci ies wi h he highes numbe o inciden s include Hakka i, Şı nak, Diya bakı ,
Ma din, Van, and Tunceli. No ably, majo u ban cen es like Is anbul and Anka a
wi nessed 59 and 22 inciden s, espec i ely. Con e sely, he Medi e anean and
Aegean coas al a eas gene ally epo ed ewe occu ences.
This spa ial pa e n aligns wi h he p esence o es ablished e o is o ganiza-
ions nea he I aq-Sy ia bo de . As no ed by he Republic o Tü kiye Minis y o
Fo eign Affai s (2022a), he PKK e o is g oup, main ains aining camps in he
Qandil Moun ains, enabling hem o infil a e Tu key and launch a acks (Republic o
Tü kiye Minis y o Fo eign Affai s 2022b). We see a decline in PKK- ela ed inciden s
ollowing coun e - e o ism ope a ions by he Tu kish A med Fo ces. The numbe o
inciden s a ibu ed o he PKK epo edly dec eased om 147 in 2016 o 10 in 2020.
3
This figu e p esen s he spa ial dis ibu ion o e o ism- ela ed a ali ies ac oss
Tu key om 2012 o 2020. The map isualizes cumula i e a ali ies in a ious egions,
highligh ing ci ies wi h a ying le els o impac .
The Eas e n Ana olia Region expe ienced he mos significan concen a ion o
a ali ies, wi h ci ies like Hakka i, Şı nak, Diya bakı , Ma din, and Gazian ep
epo ing pa icula ly high numbe s. Addi ionally, Is anbul and Anka a, majo
u ban cen es, eco ded a o al o 191 and 248 a ali ies, espec i ely. These inciden s
a ge ed no only secu i y pe sonnel bu also ci ilians, unde sco ing he indisc im-
ina e na u e o e o is a acks.
A pa icula ly de as a ing e en was he Anka a ain s a ion bombings o
Oc obe 2015, which claimed he li es o 109 ci ilians. This a ack emains he
deadlies e o is inciden in Tu key’s his o y.
In 2016, Fe hullah Gülen Te o is O ganiza ion (FETO) had a coup a emp . Two
housand people we e inju ed, and 246 people died du ing he coup a emp . The
Tu kish Pa liamen and he P esiden ial Palace we e wo o he se e al go e nmen
s uc u es ha we e bomba ded. This coup a emp was hwa ed, and he Tu kish
people upheld democ acy and he ule o law: he P ime Minis e , he Go e nmen ,
he Membe s o he G and Na ional Assembly, and he P esiden , as s a ed by he
Republic o Tü kiye Minis y o Fo eign Affai s (2022b).
A e conduc ing g aphical analysis o explo e he ela ionship be ween e o
inciden s and a ali ies, a en ion u ns o he desc ip i e s a is ics p esen ed in
Table 2. This able p o ides a summa y o he da a used o explo e he ela ionship
be ween ade and e o ism.
Table 2 p o ides a summa y o e o ism inciden s, associa ed a ali ies and
o he a iables among Tu key’s ading pa ne s. The da a e eals a wide ange o
3Fo ch onological calcula ion he ollowing da a a e used: “START (Na ional Conso ium o he
S udy o Te o ism and Responses o Te o ism). (2022). Global Te o ism Da abase, 1970–2020 [da a
file]. h ps://www.s a .umd.edu/g d”.
T ade and Te o : Tu key’s Expo Dynamics 105
inciden s, om ze o o 2,974 (in I aq, 2015). Tu key i sel expe ienced inciden s
anging om 10 o 247. Fa ali ies also a ied significan ly, wi h I aq epo ing a
maximum o 4,807 (in 2015) and Tu key’s highes being 591.
Rega ding expo s, Tu key’s la ges ading pa ne s included Ge many
(US$20.90 billion), he Uni ed S a es, he Uni ed Kingdom, I aq, and I aly. In con as ,
Bah ain, Mau i ania, and Venezuela ecei ed ela i ely low expo s om Tu key.
Following he p esen a ion o Table 2, which shows he desc ip i e s a is ics o
he da a used in his s udy, Table 3 p esen s he es ima ion esul s. Six dis inc
models we e conduc ed o analyse he ela ionship be ween e o ism, Tu key’s
expo s, and in a-indus y ade. The PPML es ima o was employed in bo h models,
and he co ec ness o he model specifica ions was assessed using he Rese Tes .
The es ima ion esul s o he six econome ic models a e p esen ed in Table 3.
Models 1 and 2 co espond o Equa ions (6) and (7), espec i ely, ocusing on he
impac o e o ism on Tu key’s expo s using he “Global Te o Index”(GTI) as he
p ima y a iable. Models 3 and 4, de i ed om Equa ions (8) and (9), inco po a e
“ e o inciden s”(INC) as a key measu e. Finally,Models 5 and 6, basedon Equa ions
(10) and (11), in oduce “ a ali ies”(FTL) as an addi ional explana o y a iable, wi h
Model 6 assessing he combined effec s o GTI and FTL. The esul s unde sco e he
significan influence o e o ism on Tu key’s expo pe o mance, whe he he
h ea is domes ic o o igina es om he ading pa ne .
Models 1, 2, and 6 highligh he Global Te o Index (GTI) as a c i ical explana o y
ac o . In hese models, he GTI o pa ne coun ies exe s a posi i e and significan
Table :Desc ip i e s a is ics.
Va iable N Mean S d De Min Max
Expo
ij
,, ,, ,. ,,
Global Te o Index (GTI) o pa ne . .
Global Te o Index (GTI) o Tu key . . . .
Inciden s (INC) in pa ne . . ,
Inciden s (INC) in Tu key . .
Fa ali ies (FTL) in pa ne . . ,
Fa ali ies (FTL) in Tu key . .
Sum o GDPs ,,a,,a,a,,a
GDP simila i y . . . .
Manu ac u ing simila i y . . . .
Rela i e endowmen (ICT pe capi a) . . . .
Popula ion o ading pa ne ,, ,, ,, ,,,
Popula ion o Tu key ,, ,, ,, ,,
Dis ance
ij
,. ,. . ,.
aRep esen ed in housands.
106 E. Z. Ünlü
Table :G a i y model es ima ion esul s using PPML es ima o .
Va iables Model () Model () Model () Model () Model () Model ()
Global Te o Index (GTI) o
pa ne
.c
(.)
.c
(.)
.b
(.)
Global Te o Index (GTI) o
Tu key
−.c
(−.)
−.c
(−.)
−.
(−.)
Inciden s (INC) in pa ne .
(.)
.a
(.)
Inciden s (INC) in Tu key −.b
(−.)
−.b
(−.)
Fa ali ies (FTL) in pa ne −.
(−.)
.
(.)
Fa ali ies (FTL) in Tu key −.b
(−.)
.
(.)
Sum o GDPs .c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
GDP simila i y .c
(.)
.c
(.)
.c
(.)
Manu ac u ing simila i y .c
(.)
.c
(.)
.c
(.)
Rela i e endowmen (ICT pe
capi a)
.c
(.)
.
(.)
.c
(.)
.b
(.)
.c
(.)
.b
(.)
Popula ion o ading pa ne −.c
(−.)
−.
(−.)
−.c
(−.)
−.c
(−.)
−.c
(−.)
−.c
(−.)
Popula ion o Tu key .
(.)
.
(.)
.b
(.)
.c
(.)
.a
(.)
.
(.)
Dis ance
ij
−.c
(−.)
−.c
(−.)
−.c
(−.)
−.b
(−.)
−.c
(−.)
−.c
(−.)
Bo de
ij
.b
(.)
.b
(.)
−.c
(−.)
−.c
(−.)
−.c
(−.)
−.c
(−.)
Colony
ij
.c
(.)
.
(.)
.b
(.)
.
(.)
.c
(.)
.c
(.)
Cus oms union .c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
Bo de *Inciden s .c
(.)
.c
(.)
Bo de *Fa ali ies .c
(.)
GlobalTe o Index*Fa ali ies −.a
(−.)
Bo de *GlobalTe o Index .c
(.)
Cons an −.
(−.)
−.
(−.)
−.
(−.)
−.c
(−.)
−.
(−.)
−.a
(−.)
T ade and Te o : Tu key’s Expo Dynamics 107
effec on Tu key’s expo s. Fo example, he coefficien s in Model 1 (0.265) and Model
2 (0.352) a e significan a he 1 % le el, sugges ing ha heigh ened e o ism isks in
pa ne coun ies a e associa ed wi h inc eased ade olumes o Tu key. In
con as , Tu key’s own GTI consis en ly shows a nega i e impac on expo s, wi h
coefficien s o −0.863 (Model 1) and −1.097 (Model 2), bo h significan a he 1 % le el,
indica ing ha ele a ed e o ism isks wi hin Tu key educe i s expo pe o -
mance. In Model 6, he in e ac ion be ween GTI and bo de p oximi y exhibi s a
significan posi i e effec (coefficien o 3.540), sugges ing ha geog aphical p ox-
imi y o high- isk a eas can pa adoxically enhance Tu key’s expo s.
Models 3 and 4 examine he influence o e o ism inciden s (INC) on Tu key’s
expo s. The findings e eal ha an inc ease in he numbe o inciden s in pa ne
coun ies has a modes ly posi i e impac on expo s, wi h a coefficien o 0.0349 in
Model 4, significan a he 10 % le el. Con e sely, e o ism inciden s wi hin Tu key
ha e a consis en ly nega i e effec on expo s, wi h coefficien s o −0.040 (Model 3)
and −0.047 (Model 4), bo h significan a he 5 % le el. These esul s indica e ha
while ising e o ism inciden s in pa ne coun ies may ma ginally boos ade,
inc eased inciden s wi hin Tu key lead o educed expo pe o mance. Fu he -
mo e, he in e ac ion be ween bo de p oximi y and inciden s is significan and
posi i e in bo h models, wi h coefficien s o 0.501 (Model 3) and 0.598 (Model 4), bo h
significan a he 1 % le el, sugges ing ha inciden s nea bo de s may encou age
c oss-bo de ade, possibly due o secu i y- ela ed comme ce.
Models 5 and 6 ocus on he effec o a ali ies (FTL) on Tu key’s expo s. The
analysis shows ha a ali ies in pa ne coun ies do no ha e a significan impac on
Tu key’s expo s, as e idenced by he insignifican coefficien s in bo h models.
Howe e , a ali ies wi hin Tu key exhibi a significan nega i e impac on expo s
in Model 5, wi h a coefficien o −0.041. In e es ingly, his ela ionship becomes
insignifican in Model 6. The in e ac ion be ween bo de p oximi y and a ali ies in
Model 5 is posi i e and significan (coefficien o 0.329), implying ha a ali ies nea
bo de s may s imula e c oss-bo de ade.
Table :(con inued)
Va iables Model () Model () Model () Model () Model () Model ()
N
Pseudo R. . . . . .
Fixed effec s Yes Yes Yes Yes Yes Yes
Rese es [.][.][.][.][.][.]
(), [] deno e -s a is ics and p- alues espec i ely. (a), (b), (c) indica e significance a %, % and % espec i ely.
PPMLHDFE command is used o he es ima es.
108 E. Z. Ünlü
Economic indica o s such as he sum o GDPs and GDP simila i y eme ge as
consis en and significan d i e s o ade ac oss all models. The sum o GDPs yields
coefficien s anging om 0.484 in Model 1 o 1.221 in Model 4, while GDP simila i y
has coefficien s anging om 1.310 in Model 5 o 1.508 in Model 1, aligning wi h
economic heo y ha sugges s la ge and economically simila coun ies end o
ade mo e wi h each o he . Manu ac u ing simila i y is also significan in Models 2,
4, and 6, wi h coefficien s be ween 0.235 and 0.493, indica ing ha g ea e simila i y
in manu ac u ing s uc u es os e s inc eased ade. Addi ionally, he ela i e
endowmen o ICT pe capi a is significan in mos models, wi h coefficien s
eflec ing ha coun ies wi h be e ICT capabili ies engage mo e in ade. The
coefficien sign o ICT se ice expo pe capi a is posi i e in all six models indica ing
in e - ype o ade. Popula ion dynamics show mixed effec s: while he ading
pa ne ’s popula ion consis en ly has a nega i e impac on expo s, wi h coefficien s
anging om −0.046 o −0.281, he significance o Tu key’s popula ion a ies ac oss
models. Bo h economic simila i ies (SIMGDP) and manu ac u ing indus y simila -
i ies (SIMMANUF) coefficien s a e posi i e, indica ing in a-indus y ade. This
implies ha Tu key and i s pa ne coun ies a e engaging in in a-indus y ade,
co obo a ing he findings o Ka acan and Ko kmaz (2022).
Geog aphical dis ance be ween ading pa ne s emains a significan
cons ain on ade ac oss all models, wi h coefficien s anging om −0.848
o −0.986. Bo de p oximi y and colonial ies gene ally enhance ade, al hough hei
effec s a y be ween models. The p esence o a cus oms union consis en ly shows a
significan and posi i e impac on ade, wi h coefficien s be ween 0.280 and 0.657,
highligh ing i s ole in acili a ing c oss-bo de comme ce.
All models include high-dimensional fixed effec s and pass he RESET es o
unc ional o m misspecifica ion, indica ing ha he models a e well-specified. The
pseudo-R-squa ed alues ange om 0.7282 o 0.9073, sugges ing a good fi o he
ade da a.
The empi ical findings o his s udy align wi h exis ing li e a u e, which
consis en ly highligh s he nega i e impac o e o ism on ade. Doucouliagos
(2017) epo ed ha an addi ional e o is a ack in a bo de ing coun y could
educe bila e al ade by app oxima ely 0.013 %, leading o a loss o oughly $6.4
million USD in o al ade alue. Addi ionally, a ali ies esul ing om e o is
e en s ha e a diminishing effec on bila e al ade. Simila ly, Shah, Hasna , and
Sa a h (2020) ound ha a 1 % inc ease in e o inciden s led o a 0.091 % educ ion
in Pakis an’s expo s. Compa able esul s we e ob ained by Egge and Gassebne
(2015), who demons a ed ha e o is ac i i ies, bo h his o ical and ongoing, in
expo ing and impo ing coun ies ypically esul in a decline in ade. Howe e , he
ex en o his impac a ies depending on he iming o he a acks and he cha ac-
e is ics o he coun ies in ol ed. Fu he mo e, hei analysis unde sco es he
T ade and Te o : Tu key’s Expo Dynamics 109
po en ial o ime agg ega ion bias, which could affec he p ecision o he findings.
Ba dwell and Iqbal (2021) u he no ed ha 74 % o e o - ela ed a ali ies occu ed
in Sy ia, I aq, A ghanis an, Nige ia, and Pakis an, leading o a 22 % dec ease in
A ghanis an’s GDP in 2018 and a 27 % dec ease in I aq’s GDP in 2014.
While Tu key is a de eloping na ion, i aces challenges ela ed o e o ism, a
pe asi e issue in many egions (Luca 2021). sugges s ha de eloped economies a e
o en be e equipped o mi iga e he economic consequences o e o ism due o
hei s onge secu i y capabili ies and as e esponse mechanisms. Con e sely,
unde de eloped na ions like Sy ia, I aq, and A ghanis an ha e bo ne a disp opo -
iona e economic bu den, wi h e o ism accoun ing o o e 50 % o GDP in some
cases. The findings o his s udy a e consis en wi h hose epo ed by (Çe in, Kese ,
and Ay 2019), who obse ed ha e o ism exe s a nega i e influence on key
mac oeconomic indica o s, including GDP, expo s, and o eign in es men , pa ic-
ula ly in he Middle Eas . Like he es ima ion esul s in his s udy, hey ound ha he
economic impac o e o ism is especially p onounced in Tu key, whe e i signifi-
can ly inc eases public expendi u e. Thei use o spa ial lag es ima ion u he
suppo s he idea ha he effec s o e o ism a e no confined o he coun ies
di ec ly expe iencing he a acks bu ha e b oade egional implica ions due o
in e na ional in e ac ions. This aligns wi h he findings ha highligh he subs an ial
impac o e o ism on Tu key’s expo s, pa icula ly when conside ing bo de
p oximi y and egional ins abili y.
In line wi h he b oade li e a u e on he impac o iolence, including e o ism
and conflic , he findings o his s udy echo he conclusions d awn by Dinçe and
Yüksel (2019), who emphasize ha conflic isks lead o significan economic and
social dis up ions in affec ed coun ies. Like how e o ism nega i ely impac s GDP,
expo s, and public expendi u e, conflic s also esul in lowe economic g ow h and
heigh ened unemploymen , pa icula ly in he Middle Eas . The o e lap be ween he
economic consequences o e o ism and conflic is e iden in he des abilizing
effec s obse ed in coun ies like Sy ia, I aq, and Tu key. Fu he mo e, he social
and poli ical epe cussions o iolence, such as inc eased co up ion and educed
go e nmen effec i eness, align wi h he obse ed nega i e impac s o e o ism on
mac oeconomic s abili y.
6 Robus ness Checks
6.1 Es ima ion wi h Yea and Impo e Fixed Effec s
To es he obus ness o he findings p esen ed in Table 3, I e-es ima e he models
using he PPML es ima o while inco po a ing yea and impo e fixed effec s. The
110 E. Z. Ünlü
esul s o hese es ima ions a e displayed in Table 4. This app oach allows us o
con ol o any unobse ed ime- a ian o coun y-specific ac o s ha may influ-
ence Tu key’s expo s, p o iding a mo e comp ehensi e unde s anding o he
ela ionships be ween e o ism, economic indica o s, and ade pe o mance.
Models 1 and 2 in Table 4 include he Global Te o Index (GTI) as he p ima y
explana o y a iable, while Models 3 and 4 ocus on e o inciden s (INC), and
Models 5 and 6 inco po a e a ali ies (FTL) as key measu es. These esul s a e
compa ed wi h he baseline models (wi hou fixed effec s) o assess he consis ency
and obus ness o he es ima es.
Unlike he p e ious es ima ions, whe e he GTI o pa ne coun ies had a
posi i e and significan effec on Tu key’s expo s, in he obus ness models, he
effec is ei he insignifican o ma ginally significan . Fo ins ance, in Models 1 and 2,
Table :G a i y model es ima ion esul s using PPML es ima o wi h “yea ”and “impo e fixed effec s”.
Va iables Model () Model () Model () Model () Model () Model ()
Global Te o Index
(GTI) o pa ne
−.
(−.)
−.
(−.)
.a
(.)
Inciden s (INC) in
pa ne
.b
(.)
.a
(.)
Fa ali ies (FTL) in
pa ne
−.
(−.)
−.
(−.)
Sum o GDPs .c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
.c
(.)
GDP simila i y .b
(.)
.b
(.)
.b
(.)
Manu ac u ing
simila i y
.b
(.)
.b
(.)
.a
(.)
Rela i e endowmen
(ICT pe capi a)
.
(.)
.
(.)
.
(.)
.c
(.)
.a
(.)
. (.)
Popula ion o ading
pa ne
−.c
(−.)
−.c
(−.)
−.b
(−.)
. (.). (.)−.
(−.)
Cus oms union −.
(−.)
−.a
(−.)
−.a
(−.)
−.c
(−.)
−.c
(−.)
−.
(−.)
Cons an .c
(.)
.c
(.)
.
(.)
−.c
(−.)
−.c
(−.)
−.
(−.)
N
Pseudo R. . . . . .
Fixed effec s Impo e
yea
Impo e
yea
Impo e
yea
Impo e Impo e Impo e
yea
Rese es [.][.][.][.][.][.]
(), [] deno e z-s a is ics and p- alues espec i ely. (a), (b), (c) indica e significance a %, % and % espec i ely.
PPMLHDFE command is used o he es ima es.
T ade and Te o : Tu key’s Expo Dynamics 111
he coefficien s a e (−0.050) and (−0.054), espec i ely, wi h bo h being s a is ically
insignifican . Howe e , Model 6 shows a posi i e and ma ginally significan effec
(0.057, significan a he 10 % le el). This con as s wi h he esul s in Table 3, whe e
he GTI o pa ne coun ies consis en ly exhibi ed a posi i e and highly significan
effec .
The esul s o e o inciden s (INC) in pa ne coun ies in Models 3 and 4
emain consis en wi h hose in Table 3. The coefficien o inciden s is posi i e and
significan in bo h models, wi h he effec being s onge and mo e s a is ically
significan in Model 3 (0.025, significan a he 5 % le el). This aligns wi h he p e ious
finding ha inc easing e o ism inciden s in pa ne coun ies can pa adoxically
boos Tu key’s expo s.
Fa ali ies (FTL) in pa ne coun ies con inue o show no significan impac on
Tu key’s expo s ac oss Models 5 and 6. This finding emains consis en wi h he
es ima ion esul s p esen ed in Table 3, indica ing ha a ali ies in ading pa ne s
do no exe a di ec influence on ade pe o mance.
Economic ac o s such as he sum o GDPs and GDP simila i y emain highly
significan ac oss all models, consis en wi h he p e ious es ima ion. The sum o
GDPs con inues o exhibi s ong posi i e effec s on Tu key’s expo s, wi h
coefficien s anging om 0.603 (Model 1) o 1.624 (Model 5), all significan a he 1 %
le el. GDP simila i y also emains a posi i e and significan d i e o ade, hough
he coefficien s a e sligh ly lowe han in he p e ious models, anging om 0.830
(Model 3) o 0.952 (Model 1).
Like he p e ious es ima ed models, manu ac u ing simila i y con inues o exe
a posi i e and significan influence on expo s. The coefficien s a e smalle in
magni ude, anging om 0.215 o 0.277, bu emain significan in Models 2, 4, and 6,
indica ing he impo ance o in a-indus y ade.
The popula ion o ading pa ne s exhibi s a s ong nega i e effec in Models 1.
Howe e , in Models 4, 5, and 6, he significance o he popula ion a iable
diminishes, wi h coefficien s becoming ei he insignifican o changing signs. This
di e gence could be a ibu ed o he inclusion o fixed effec s, which con ol o
unobse ed coun y-specific cha ac e is ics ha may be co ela ed wi h popula ion.
The cus oms union a iable, which was p e iously posi i e and significan in
Table 3, now exhibi s nega i e o insignifican coefficien s ac oss mos models in
Table 4. The inclusion o fixed effec s appea s o al e he impac o cus oms unions
on Tu key’s expo s, sugges ing ha he posi i e associa ion obse ed ea lie may be
d i en by unobse ed ac o s cap u ed by he impo e o yea effec s.
The RESET es esul s ac oss all models indica e ha he models a e co ec ly
specified in e ms o unc ional o m. Only impo e fixed effec s a e included in
Model 4 and Model 5 because including yea effec s esul ed in a specifica ion e o ,
as iden ified by he RESET es . This ensu es ha he model emains well-specified
112 E. Z. Ünlü
and a oids o e fi ing by ocusing on he significan impo e -specific a ia ions
wi hou in oducing unnecessa y yea effec s.
When compa ing he esul s om Table 4 (wi h yea and impo e fixed effec s)
o hose in Table 3 (wi hou fixed effec s), se e al diffe ences eme ge:
–The posi i e impac o he Global Te o Index o pa ne coun ies obse ed in
Table 3 becomes weake o insignifican in Table 4, sugges ing ha con olling
o unobse ed impo e and yea -specific ac o s al e s he ela ionship
be ween e o ism and ade.
–The effec s o e o inciden s in pa ne coun ies emain consis en ac oss
bo h ables, ein o cing he finding ha an inc ease in inciden s may boos
expo s.
–Economic a iables such as he sum o GDPs and GDP simila i y emain obus
and significan ac oss bo h models, highligh ing hei s ong influence on ade
pe o mance.
–The influence o popula ion and cus oms unions a ies be ween he wo se s o
esul s, wi h he fixed-effec s models p o iding mo e nuanced insigh s in o hese
ela ionships.
Including yea and impo e fixed effec s p o ides a mo e efined unde s anding o
he ac o s affec ing Tu key’s expo s, pa icula ly in e ms o he ole o e o ism
and economic indica o s. While some esul s emain consis en wi h he no-effec s
models, o he s, such as he impac o he Global Te o Index and cus oms unions,
sugges ha unobse ed ac o s play a c i ical ole in shaping hese ela ionships.
6.2 Random Coefficien s Model
The applica ion o andom coefficien s models is pa icula ly sui ed o cap u ing he
inhe en complexi y and he e ogenei y in g a i y models. As no ed by Bal agi, Egge ,
and P affe may (2014), andom effec s models effec i ely handle he double o e en
iple indexing o coun y pai s, add essing he unobse ed he e ogenei y ha a ises
in bila e al ade ela ionships. In g a i y models, unobse ed ac o s such as cul-
u al affini y, his o ical ies, o pa icipa ion in ade ag eemen s o en a y ac oss
coun ies and a e no di ec ly measu able. By inco po a ing andom effec s, he
model in oduces flexibili y, allowing hese unobse ed ac o s o be ea ed as
andom, a he han fixed, ac oss coun y pai s. This enables he model o accoun
o a ia ions in he way diffe en coun ies in e ac in ade, wi hou he es ic i e
assump ions imposed by fixed-effec s models.
Ma hema ically, conside a s anda d g a i y model o ade whe e bila e al
ade flows Y
ij
be ween coun y i and coun y j a e influenced by obse ed ac o s X
ij
T ade and Te o : Tu key’s Expo Dynamics 113
(such as GDP, dis ance, and a iffs). A basic fixed-effec s model assumes ha he
unobse ed he e ogenei y ac oss coun y pai s is cons an :
Yij =Xijβ+αi+αj+εij (12)
whe e α
i
and α
j
ep esen coun y-specificfixed effec s. Howe e , his specifica ion
may no cap u e he ull complexi y o ade ela ionships, especially when he
unobse ed ac o s a e no cons an ac oss ading pa ne s. By con as , a andom
coefficien s model assumes ha he impac o some explana o y a iables a ies
ac oss coun ies:
Yij =Xij(β+uij)+αi+αj+εij (13)
whe e u
i
ep esen s he andom de ia ion om he a e age effec o X
ij
, allowing
o he e ogenei y in he coefficien s ac oss coun y pai s. This andom a ia ion
cap u es he idea ha , o ins ance, he effec o dis ance o a iffs on ade may diffe
depending on he specific ade ela ionship.
Egge and P usa (2014) u he jus i y he use o andom coefficien s models in
he p esence o measu emen e o o he e ogeneous esponses o a iables like
a iffs and dis ance. In ade models, hese a iables a e o en subjec o a ying
le elso accu acy o influence depending on hecoun ies in ol ed. Fo example, he
impac o he Global Te o ism Index (GTI) o GDP on ade may no be uni o m
ac oss all coun ies; some ading ela ionships may be mo e sensi i e o e o ism
o economic size han o he s. A andom coefficien s model cap u es his a ia ion by
allowing o diffe en sensi i i ies ac oss ading pa ne s, hus eflec ing he
nuanced na u e o in e na ional ade dynamics.
To o malize his, conside he andom coefficien o a a iable X
ij
in he con ex
o in e na ional ade. The model can be ep esen ed as:
Yij =Xijβ+Zij(γ+uij)+εij (14)
he e, Z
ij
ep esen s he se o a iables o which he coefficien s γa e allowed o
a y ac oss coun ies, wi h u
ij
∼N(0,σ2) ep esen ing he andom coefficien s. This
specifica ion in oduces flexibili y, as i accoun s o he coun y-specific o pai -
specific de ia ions om he a e age effec , p o iding a mo e accu a e ep esen a-
ion o he he e ogeneous na u e o ade flows.
In conclusion, he adop ion o a andom coefficien s model allows o a mo e
ealis ic and flexible app oach o modelling in e na ional ade, pa icula ly when
dealing wi h complex phenomena such as e o ism’s impac on ade. S anda d
fixed-effec s models impose uni o mi y on he ela ionships be ween a iables,
which may be o e ly es ic i e in a global con ex whe e economic, poli ical, and
social condi ions a y widely. By allowing key a iables such as he GTI and GDP o
ha e diffe en effec s ac oss coun ies, andom coefficien s models offe a mo e
114 E. Z. Ünlü
Dinçe , Hasan, and Se ha Yüksel. 2019. “Analyzing he Possibili y o Violen Conflic in he Middle Eas
Economies Using De e minan s o Global Conflic Risk Index wi h an In eg a ed Fuzzy Mul ic i e ia
Decision Making Model.”In The Impac o Global Te o ism on Economic and Poli ical De elopmen ,
edi ed by Ramesh Chand a Das, 155–66. Leeds: Eme ald Publishing Limi ed.
Doucouliagos, H is os. 2017. “An Inju y o One Is an Inju y o All: Te o ism’s Spillo e Effec s on Bila e al
T ade.”SSRN Schola ly Pape . Roches e , NY. h ps://doi.o g/10.2139/ss n.2998968.
Egge , Pe e , and Ma in Gassebne . 2015. “In e na ional Te o ism as a T ade Impedimen ?”Ox o d
Economic Pape s 67 (1): 42–62.
Egge , Pe e , and Jan P usa. 2014. “The De e minan s o T ade Cos s: A Random Coefficien s App oach.”
CESi o Wo king Pape Se ies, CESi o Wo king Pape Se ies. h ps://ideas. epec.o g//p/ces/ceswps/_
4859.h ml.
Ende s, Wal e , Ga y A. Hoo e , and Todd Sandle . 2016. “The Changing Nonlinea Rela ionship be ween
Income and Te o ism.”Jou nal o Conflic Resolu ion 60 (2): 195–225.
Ende s, Wal e , and Todd Sandle . 2000. “Is T ansna ional Te o ism Becoming Mo e Th ea ening? A
Time-Se ies In es iga ion.”Jou nal o Conflic Resolu ion 44 (3): 307–32.
Gaibulloe , Khus a , and Todd Sandle . 2009. “The Impac o Te o ism and Conflic s on G ow h in Asia.”
Economics & Poli ics 21 (3): 359–83.
Gaibulloe , Khus a , and Todd Sandle . 2019. “Wha We Ha e Lea ned abou Te o ism since 9/11.”Jou nal
o Economic Li e a u e 57 (2): 275–328.
Gok, Ib ahim Yasa , Ya uz Demi dogen, and Se a Topuz. 2020. “The Impac s o Te o ism on Tu kish
Equi y Ma ke : An In es iga ion Using In aday Da a.”Physica A: S a is ical Mechanics and I s
Applica ions 540 (Feb ua y): 123484.
Haq, Zahoo Ul, Zia Ullah, and Ja ed Iqbal. 2018. “Te o is Inciden s and T ade.”Global Social Sciences
Re iew 3 (2): 55–70.
Helpman, Elhanan, and Paul K ugman. 1987. Ma ke S uc u e and Fo eign T ade: Inc easing Re u ns,
Impe ec Compe i ion, and he In e na ional Economy. Camb idge, Massachuse s: MIT P ess.
Ib ahim, Bello, and A iA i. 2021. “The Impac o Te o ism on Fo eign Di ec In es men : The Case o
Tu key.”Üçüncü Sek ö Sosyal Ekonomi 56 (3): 1781–97.
Ila slan, Kenan, and Mune e e Yildiz. 2022. “The Effec s o Te o ism and Economic Indica o s on Bank
Loans o he P i a e Sec o : E idence om De eloping Coun ies.”Eme ging Ma ke s Finance and
T ade 58 (2): 329–41.
IMF. 2023. “IMF Di ec ion o T ade S a is ics.”Di ec ion o T ade S a is ics. h ps://da a.im .o g/?
sk=9D6028D4-F14A-464C-A2F2-59B2CD424B85.
Ins i u e o Economics & Peace (IEP). 2023. Global Te o ism Index. h ps://www. isiono humani y.o g/
public- elease-da a/.
Julius, S acy, Nnanna P. Azu, and Maimuna Y. Muhammad. 2019. “Assessing he Impac o Te o ism in
T ade De elopmen in he SADC Region: A G a i y Model App oach.”Asian Economic and Financial
Re iew 9 (10): 1147–59.
Ka acan, Semih, and Özge Ko kmaz. 2022. “Tu kish Expo s be o e and a e he 2001 Financial C isis: A
Panel G a i y Model.”Fo eign T ade Re iew 57 (1): 27–40.
Lanoua , Cha eddine, and Ume Shahzad. 2021. “Te o ism and Capi al Flows: The Missed Impac o
Te o ism in Big Ci ies.”Applied Economics Le e s 28 (18): 1626–33.
Lei ão, Nuno Ca los, Ho ácio C. Faus ino, and Yushi Yoshida. 2010. “F agmen a ion, Ve ical In a-Indus y
T ade, and Au omobile Componen s.”Economics Bulle in 30: 1006–15.
Linde , S. B. 1961. An Essay on T ade and T ans o ma ion. New Yo k: John Wiley & Sons.
Luca, Dumi u-Mihai. 2021. “The Global Economic Impac o Te o ism.”Eu opean Jou nal o
In e disciplina y S udies A icles 7: 19–29.
T ade and Te o : Tu key’s Expo Dynamics 121
Meie ieks, Daniel, and F ied ich Schneide . 2021. “Te o ism and In e na ional Economic Policy.”
Eu opean Jou nal o Poli ical Economy 69 (Sep embe ): 102011.
Oka o , Godwin, and Jeni e Piesse. 2018. “Empi ical In es iga ion in o he De e minan s o Te o ism:
E idence om F agile S a es.”De ence and Peace Economics 29 (6): 697–711.
Pham, Cong S., and Ch is Doucouliagos. 2017. An Inju y o One Is an Inju y o All: Te o ism’s Spillo e
Effec s on Bila e al T ade. h ps://www.iza.o g/publica ions/dp/10859/an-inju y- o-one-is-an-inju y-
o-all- e o isms-spillo e -effec s-on-bila e al- ade.
Pham, Cong S., and Hoa Nguyen. 2024. “Impac o Te o on In e na ional T ade in Financial Se ices:
Does he De elopmen o Financial Ins i u ions Ma e ?”The Wo ld Economy n/a (n/a): 1–39.
Polyxeni, Kechagia, and Me axas Theodo e. 2019. “An Empi ical In es iga ion o FDI Inflows in De eloping
Economies: Te o ism as a De e minan Fac o .”The Jou nal o Economic Asymme ies 20 (No embe ):
e00125.
Republic o Tü kiye Minis y o Fo eign Affai s. 2022a. An In o ma i e No e o he Minis y o Fo eign
Affai s on he Failed Coup A emp /Republic o Tü kiye Minis y o Fo eign Affai s. h ps://www.m a.
go . /an-in o ma i e-no e-o - he-minis y-o - o eign-affai s-on- he- ailed-coup-a emp .en.m a.
Republic o Tü kiye Minis y o Fo eign Affai s. 2022b. h ps://www.m a.go . /pkk.en.m a.
Ruiz Es ada, Ma io A u o, Donghyun Pa k, and Alam Khan. 2018. “The Impac o Te o ism on Economic
Pe o mance: The Case o Tu key.”Economic Analysis and Policy 60 (Decembe ): 78–88.
Sandle , Todd, and Wal e Ende s. 2008. Economic Consequences o Te o ism in De eloped and De eloping
Coun ies: An O e iew, edi ed by Philip Kee e , and No man Loayza, 1s ed., 17–47. Camb idge:
Camb idge Uni e si y P ess.
Sezgin, Şennu . 2019. “In e na ional T ade and Te o Rela ionship in Tu key.”Sosyoekonomi 27 (40):
55–66.
Shah, Syed Hasana , Ha sa Hasna , and Delpachi a Sa a h. 2020. “The Impac o Fo eign Di ec
In es men on T ade in Pakis an: The Mode a ing Role o Te o ism.”Jou nal o Economic S udies 47
(5): 1137–54.
Sil a,J.M. C. San os, and Sil ana Ten ey o. 2006. “The Log o G a i y.”The Re iewo Economics and S a is ics
88 (4): 641–58.
The Wo ld Bank. 2023. WDI. h ps://da abank.wo ldbank.o g/sou ce/wo ld-de elopmen -indica o s.
Tinbe gen, Jan. 1962. Shaping he Wo ld Economy; Sugges ions o an In e na ional Economic Policy. New
Yo k: Twen ie h Cen u y Fund.
Ülengin, Füsun, Bo a Çekyay, Pe al Tok aşPalu , Bu ç Ülengin, Özgü Kabak, Özay Özaydın, and
Şule Önsel Ekici. 2015. “Effec s o Quo as on Tu kish Fo eign T ade: A G a i y Model.”T anspo Policy
38 (C): 1–7.
Wakasugi, Ryuhei. 2007. “Ve ical In a-Indus y T ade and Economic In eg a ion in Eas Asia*.”Asian
Economic Pape s 6 (1): 26–39.
Yoshida, Yushi, Nuno Ca los Lei ão, and Ho ácio C. Faus ino. 2009. “Ve ical In a-Indus y T ade and
Fo eign Di ec In es men be ween Japan and Eu opean Coun ies.”A lan ic Economic Jou nal 37:
351–65.
122 E. Z. Ünlü
