Kös ekçi, Ahme ; Çelik, Ali
A icle
Modelling he ela ionship be ween public expendi u e,
ax e enue and economic g ow h in Tü kiye using he
AARDL app oach
Ekonomika
P o ided in Coope a ion wi h:
Vilnius Uni e si y P ess
Sugges ed Ci a ion: Kös ekçi, Ahme ; Çelik, Ali (2024) : Modelling he ela ionship be ween public
expendi u e, ax e enue and economic g ow h in Tü kiye using he AARDL app oach, Ekonomika,
ISSN 2424-6166, Vilnius Uni e si y P ess, Vilnius, Vol. 103, Iss. 2, pp. 90-108,
h ps://doi.o g/10.15388/Ekon.2024.103.2.5
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323149
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/
90
Ekonomika ISSN 1392-1258 eISSN 2424-6166
2024, ol. 103(2), pp. 90–108 DOI: h ps://doi.o g/10.15388/Ekon.2024.103.2.5
Modelling he Rela ionship Be ween
Public Expendi u e, Tax Re enue and
Economic G ow h in Tü kiye Using
he AARDL App oach
Ahme Kös ekçi
Fı a Uni e si y, Elaziğ, Tu key
Email: [email p o ec ed]
ORCID: h ps://o cid.o g/0000-0001-8485-887X
Ali Celik*
Is anbul Gelisim Uni e si y, Is anbul, Tu key
Wes e n Caspian Uni e si y in Baku, Aze baijan
Email: [email p o ec ed]
ORCID: h ps://o cid.o g/0000-0003-3794-7786
Abs ac . This s udy aims o in es iga e he mac oeconomic impac o iscal policy in Tü kiye, whe e iscal
policy aces se e al challenges. Using annual ime se ies da a om 1980 o 2021, we examine he impac o ax
and public expendi u e subcomponen s on GDP using he augmen ed au o eg essi e dis ibu ed lag (A-ARDL)
bound es app oach p oposed by Sam e al. (2019). The A-ARDL es esul s indica e ha ax e enue has a
posi i e impac on economic g ow h in he sho un, while ax e enue has a nega i e impac on economic
g ow h in he long un. Fu he mo e, we conclude ha inc eases in cu en and in es men expendi u es ha e
a posi i e impac on economic g ow h in he sho and long un, while inc eases in ans e expendi u es ha e
a nega i e impac on economic g ow h in he sho un.
Keywo ds: Fiscal policy, Tax e enue, Public expendi u e, Economic g ow h, Tü kiye
1. In oduc ion
His o ically, he e ha e been pe iods when he impo ance o iscal policy as a policy ins u-
men has inc eased o dec eased. Fo example, wi h he s ock ma ke c ash and he G ea
Dep ession, policymake s pushed o a mo e ac i e ole o he s a e in he economy (Ho on
& El-Ganainy, 2019). As a esul , ma ke ailu es allowed o he implemen a ion o iscal
policy, which was Keynes’s p ima y ool o he in e en ionis economic app oach. Howe e ,
a e he 1970 oil c isis, bo h ad anced and eme ging economies aced he p oblem o iscal
Recei ed: 16/02/2024. Re ised: 09/04/2024. Accep ed: 29/04/2024
Copy igh © 2024 Ali Celik, Ahme Kös ekçi. Published by Vilnius Uni e si y P ess
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use,
dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed.
Con en s lis s a ailable a Vilnius Uni e si y P ess
* Co esponden au ho .
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
91
imbalances and he deb - o-GDP a io ose abo e 100 pe cen in some coun ies ha inanced
hei budge de ici s by bo owing (Ka agöz & Keskin, 2016, p. 409). This la ge iscal shock
o ced policymake s o answe a ious ques ions abou how o sol e he p oblems. Du ing
his pe iod, when he concep o go e nmen ailu e began o be sc u inized in academic
ci cles (Buchanan, 1983; K uege , 1990), a nega i e pe spec i e on he implemen a ion o
iscal policy eme ged. The nega i e iew o iscal policy was no limi ed o he 1970s, bu
con inued un il he ea ly 2000s (Dullien, 2012, p. 7).
Despi e hese changes, which educed he ole and in luence o go e nmen in he
economy, many coun ies esumed mo e ac i e iscal policies when he global inancial
c isis h ea ened a global ecession (Ho on & El-Ganainy, 2019). Fo ins ance, du ing he
economic c isis o 2008, go e nmen s in e ened o suppo inancial sys ems, s imula e
economic g ow h, and mi iga e he impac o he c isis on ulne able popula ions. This
made he ole and objec i es o iscal policy pa icula ly impo an . Indeed, in he s a e-
men ollowing hei Ap il 2009 summi in London, G20 leade s decla ed ha hey had
emba ked on an unp eceden ed and coo dina ed iscal expansion (Ho on & El-Ganainy,
2019). As iscal s imulus expanded and was on he agenda o all coun ies, i was deba -
ed whe he ax cu s o spending inc eases would be a be e solu ion o hese coun ies
(Alesina & A dagna, 2009).
A e he 2008 c isis, he expansion o iscal s imulus was suppo ed no only by na-
ional and in e na ional mone a y and iscal au ho i ies bu also by economis s. Indeed, he
esul s o he policies implemen ed con i med he e ec i eness o iscal policy in s abi-
lizing agg ega e demand, especially du ing pe iods o ex ao dina y economic weakness
(Dullien, 2012). Simila ly, he changes in iscal policy due o he co ona i us pandemic
(COVID-19), which s a ed in 2019, we e no able (Weins ock, 2021). In esponse o pan-
demics ha ha e his o ically caused se e e ecessions, many coun ies and policymake s
ha e eso ed o ac i e iscal policies o o e ed a ious iscal policy packages (A onso
& Coelho, 2023; Boug e al., 2023; Chudik e al., 2021). As a esul , e en i he iscal
consolida ion app oach is insu icien , iscal policy has been ein igo a ed in he global
discou se o e he las qua e cen u y (Woldu & Kano, 2023) and inding an e ec i e
iscal policy mix is mo e impo an oday han e e (Donadelli & G üning, 2021). In his
subjec , in o ma ion on he economic e ec s o iscal policy is an impo an elemen o
guide he achie emen o iscal sus ainabili y and should be aken in o accoun in he de-
sign and ecommenda ion o public policies (Sos illa-Ri e o & Rubio-Gue e o, 2022).
The e o e, in his s udy, which aims o examine he e ec s o ax and expendi u e policies,
we conside Tü kiye as an in e es ing and impo an example. This is because Tü kiye,
which s uggled wi h high budge de ici s in he 1990s, managed o c ea e a s ong iscal
s uc u e by ensu ing iscal discipline in he ea ly 2000s. Howe e , due o he policies
implemen ed in ecen yea s, iscal policy in Tü kiye has aced a ious challenges and
he mac oeconomic e ec s o iscal policy ha e become an impo an opic o deba e.
Ou main hypo hesis is ha iscal policy has a signi ican e ec on GDP g ow h in
Tü kiye, bu his e ec di e s signi ican ly ac oss he subcomponen s o iscal policy
ins umen s. Wi hin he scope o his hypo hesis, he main mo i a ion and objec i e o
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
92
his s udy is o empi ically de e mine whe he iscal policy is e ec i e in s imula ing
economic ac i i y in Tü kiye and, i so, he s eng h and du a ion o hese e ec s. In his
sense, speci ic ques ions such as whe he he subcomponen s o iscal policy ins umen s
ha e di e en e ec s in p omo ing economic g ow h and which o he subcomponen s o
iscal policy ins umen s ha e s onge e ec s on economic g ow h cons i u e he esea ch
ques ions o ou s udy. The answe s o hese ques ions will con ibu e o he ongoing
deba e on he ole o iscal policy. Based on he basic in o ma ion p o ided, he sec ions
o he s udy a e he in oduc ion a e o ganized as ollows. The empi ical li e a u e e-
la ed o p e ious s udies is p esen ed in Sec ion 2. The da a se , model and me hodology
a e indica ed in Sec ion 3. Sec ion 4 p esen s he empi ical esul s. The pape ends wi h
conclusions and policy ecommenda ions.
2. Li e a u e Re iew
Fiscal policy, in which go e nmen s can in luence economic ac i i y by using public ex-
pendi u es, ax e enues o bo h (Al es & Palma, 2023; Ka aş & Ka aş, 2023; Weins ock,
2021), has ecen ly a ac ed a en ion. This is pa icula ly e iden du ing economic cycles
(Calda a & Kamps, 2017; Ramey & Zubai i, 2018; Te a e al., 2021). On he o he hand,
ega dless o business cycles, many coun ies eso o con ac iona y iscal policy meas-
u es o educe public deb le els ha inc ease as a esul o iscal expansion (Alesina &
A dagna, 2009; Alesina e al., 2015; A izala e al., 2021; Ch is elis e al., 2019; Rompuy,
2021; Woldu & Kano, 2023). Despi e he ole o iscal policy, he e is no consensus on he
size and e ec i eness o he subcomponen s o iscal ins umen s and he e is s ill deba e
on he implemen a ion o iscal policy om a mac oeconomic pe spec i e (Fukuda, 2023;
Goo jes & Haan, 2022; Heimbe ge , 2023; Mawejje e & Odhiambo, 2022).
The impac o iscal policy is assessed by Blancha d and Pe o i (2002) using da a on he
elas ici y o iscal a iables. They conclude ha expansiona y iscal shocks boos ou pu ,
ha e posi i e e ec s on p i a e consump ion, and nega i ely a ec p i a e in es men .
Cas o (2006), who analyzed he mac oeconomic e ec s o iscal policy in Spain wi h
da a o he pe iod 1980–2001, concluded ha inc eases in public expendi u es ha e a
posi i e e ec on GDP in he sho un. Howe e , his e ec was ound o be nega i e in
he medium and long un. Poin ing o he non-Keynesian e ec s o iscal policy, he a icle
showed ha a iscal consolida ion based on expendi u e cu s can ha e an expansiona y
e ec on economic ac i i y. Gio dano e al. (2007) ound in a ela ed s udy ha public
spending has a posi i e and long- e m impac on ou pu in I aly. The SVAR me hod is
used by Moun o d and Uhlig (2009) o analyze he e ec s o iscal policy in he Uni ed
S a es using qua e ly da a co e ing he yea s 1955–2000. Acco ding o he s udy’s p i-
ma y indings, ax cu s inanced by de ici s a e he bes iscal policy o imp o ing GDP
and s imula ing he economy. Howe e , his esea ch illus a es he neoclassical iew o
go e nmen in e en ion in he economy because o he long- e m cos s o iscal expan-
sion wi h public expendi u es. Con e sely, Jawaid e al. (2010) ind ha he Keynesian
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
93
pe spec i e is e lec ed in hese e ec s and ha iscal policy is a c ucial mac oeconomic
policy ool o suppo ing ou pu g ow h in Pakis an.
Pa kyn and Vehbi (2014) examine he impac o public expendi u e and e enue on
ou pu in New Zealand. They ind ha public expendi u e shocks ha e a small posi i e
e ec on ou pu in he sho un bu lead o lowe ou pu in he medium and long un.
Al hough he e ec o changes in ax e enue on GDP is less p onounced, i is ound o be
mode a e. The e o e, he esea che s ound ha he impac o disc e iona y iscal policy
on GDP is p o-cyclical. Using da a om 1970 o 2009 o 20 OECD coun ies, Yang e
al. (2015) conclude ha sho - e m iscal adjus men s always ha e a con ac iona y e ec
on economic ac i i y. Rega ding he unc ion o he composi ion o he iscal adjus men ,
hey ind ha ou pu educ ions om expendi u e-based iscal adjus men s a e smalle
han hose om ax-based iscal adjus men s. Howe e , heo e ical models sugges ha
households’ ma ginal p opensi y o consume may some imes ha e binding bo owing
cons ain s and may be asymme ic in he ace o empo a y changes in income o expend-
i u e (Sos illa-Ri e o & Rubio-Gue e o, 2022). In his con ex , Bunn e al. (2018) ind
ha households’ ma ginal p opensi y o consume is highe unde nega i e income shocks
han unde posi i e shocks. This implies ha a la ge pa o he expendi u e asymme y
can be explained by households’ balance shee cha ac e is ics such as high deb le els,
small liquidi y bu e s, conce ns abou he c edi , and he possibili y o low u u e income.
Pula and Elshani (2018) o Koso o economy and A in e al. (2019) o OECD
coun ies conduc ed simila s udies and ound esul s suppo ing he Keynesian iew. In
ano he s udy, A onso and Aubyn (2019) examined he mac oeconomic impac o public
and p i a e in es men spending based on da a om 17 OECD coun ies in 1960–2014,
using VAR analysis. Using impulse esponse unc ions, he esea che s ound ha public
in es men spending has a con ac iona y e ec in Finland, he UK, Sweden, Japan, and
Canada, and a posi i e e ec in mos o he coun ies. The s udy also ound empi ical e -
idence ha public in es men spending c owds ou p i a e in es men spending. In e ms
o expendi u es, cu en spending boos s economic g ow h, acco ding o he indings o
he Sel ana han e al. (2021) s udy. In con as o his s udy, Oni ade e al. (2020) ind ha
cu en expendi u es ha e s ong and opposi e e ec s on economic g ow h.
Sos illa-Ri e o and Rubio-Gue e o (2022) examined he sho - and long- e m sym-
me ic and asymme ic e ec s o iscal policy on ou pu . Using qua e ly da a om 1980
o 2020, hey ind ha while educ ions in public spending o inc eases in ne incomes
educe he sho - and long- e m e ec s on g ow h, inc eases in public spending and
dec eases in ne incomes con ibu e o Spain’s economic g ow h in he sho and long
un. Golpe e al. (2023), in hei s udy wi h qua e ly da a o he pe iod 1990–2019 o
12 EU coun ies, ound ha mone a y policy plays a leading ole in economic g ow h
in he complex economic sys em, while agg ega e public expendi u es a e an impo an
ins umen suppo ing he d i ing o ce o iscal policy. Am i e al. (2023) used a panel
da a se o 24 Indonesian p o inces om 2006 o 2015 o apply a dynamic GMM model
o es ima e he e ec o public expendi u e on g ow h. The esea che s ound e idence
ha public spending signi ican ly boos s economic g ow h. Howe e , i u ned ou ha
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
94
local ax e o s had a nega i e impac on g ow h in he economy. Howe e , he esul s
o he s udy conduc ed by Al es and Palma (2023) wi h he Mixed F equency Vec o
Au o eg essi e (MIDAS-VAR) model o B azil showed ha public expendi u es ha e
no signi ican e ec on eal GDP g ow h.
Tu ning o he case o Tü kiye, he e ha e been many s udies on he ela ionship be ween
iscal policy and GDP g ow h. Howe e , he limi ed numbe o s udies wi h subindica o s
o iscal policy ins umen s sugges s ha mo e s udies a e needed o mo e speci ic esul s.
Using da a o Tü kiye om 1998 o 2016, Ka ahan and Çolak (2019) show ha public
spending has a posi i e impac on economic g ow h in he sho un and a nega i e impac
in he long un. F om an economic policy pe spec i e, he e is signi ican e idence ha ex-
pansiona y iscal policy can suppo economic g ow h in he sho un. A e is e al. (2021)
conduc ed a di e en s udy and analyzed he compa a i e e ec i eness o mone a y and
iscal policy o Tü kiye. Al hough bo h iscal and mone a y policies can a ec ou pu g ow h
o di e en deg ees, empi ical e idence sugges s ha mone a y policy has a g ea e impac
on ou pu g ow h. The esul s on iscal policy show ha di ec and indi ec axes nega i ely
e ec ou pu g ow h in he sho and long e m, while public consump ion and in es men
expendi u es ha e a posi i e e ec on ou pu . Öze and Ka agöl (2018) analyzed he e ec
o mone a y and iscal policies on g ow h in he Tu kish economy o he pe iod 1998–2016
and ound ha mone a y policy has only a sho - e m e ec on g ow h, while iscal policy
a ec s g ow h in he long un and causes g ow h. Du an (2022), using he ARDL bounds es
o he pe iod 1975–2020, inds ha he e ec o subcomponen s o public expendi u e on
economic g ow h in Tü kiye is posi i e. In summa y, i is no ed ha he e is disag eemen
in Tü kiye’s empi ical li e a u e on he opic.
3. Da a se , model and me hodology
3.1. Da a and model
The s udy examines he e ec s o iscal policy in Tü kiye o e he pe iod 1980–2021.
Acco dingly, Table 1 shows he a iables used in he s udy and he sou ces om which he
da a we e compiled. The econome ic model used in he s udy is based on he Keynesian
expendi u e model. John Mayna d Keynes c ea ed a simpli ied e sion o he economy
known as he Keynesian Expendi u e Model o Agg ega e Expendi u e Model. The main
ocus o his model is he ela ionship be ween he le el o agg ega e expendi u e and
he le el o eal GDP. In an open economy, consump ion (C), in es men (I), go e nmen
spending (G), and ne expo s (NX) a e he basic expendi u e componen s in he model
(Keynes, 1936). The sum o go e nmen spending, ne expo s, in es men , and consump-
ion is he inal planned o al expendi u e in he economy. When planned o al expendi u es
(AE) and o al income (Y) a e equal, mac oeconomic equilib ium is eached. In his case,
o he equilib ium condi ion:
AE = Y = (C0 + c(Y – T)) + I0 + G0 + (X − M) (1)
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
95
In summa y, he equilib ium le el o eal GDP in he Keynesian Expendi u e Model is
ep esen ed by his equa ion. I indica es ha he equilib ium GDP is de e mined by au on-
omous consump ion (C0), ma ginal consump ion p opensi y (C), au onomous in es men
(I0), au onomous go e nmen spending (G0), axes (T), and ne expo s (NX). This means
ha a change in axes o au onomous expendi u es (like go e nmen spending, in es men ,
e c.) will esul in a change in na ional income. In his heo e ical amewo k, we gene a e
ou econome ic model in which he ne expo a iable is no u ilized as ollows:
GDP = α0 + α1TR + α2CE + α3IE + α4TE + u (2)
In Eq. (2), he subsc ip s a es he ime pe iod. In addi ion, α0 shows a cons an e m,
while u e e s o he e o e m. In addi ion, we employed G oss Domes ic P oduc (GDP)
as he dependen a iable, while using Tax Re enues (TR), Cu en Expendi u e (CE),
In es men Expendi u e (IE), and T ans e Expendi u e (TE) as independen a iables.
Table 1. The desc ip ion o a iables
Va iables Uni Sou ce
G oss Domes ic P oduc (GDP) Annual G ow h Ra e (%) The Wo ld Bank
Tax Re enues (TR) Sha e o in GDP (%) Di ec o a e o S a egy and
Budge
Cu en Expendi u e (CE) Sha e o in GDP (%)
In es men Expendi u e (IE) Sha e o in GDP (%)
T ans e Expendi u e (TE) Sha e o in GDP (%)
3.2. Me hodology
3.2.1. Uni oo es
As is known, adi ional uni oo es s such as Augmen ed Dickey–Fulle (ADF) (1979)
and Phillips–Pe on (PP) (1989) a e linea adi ional uni oo es s ha do no ake in o
accoun s uc u al b eaks. The Fou ie ADF es de eloped by Ch is opoulos and Leon
Ledesman (2010), which can be conside ed one o he cu en uni oo es s modelling
s uc u al b eaks in se ies wi h igonome ic e ms, was used. The main ad an age o his
es is ha i allows o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es
s a is ics a e calcula ed as ollows (Ch is opoulos & Leon-Ledesman, 2010):
o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es s a is ics a e calcula ed
as ollows (Ch is opoulos & Leon-Ledesman, 2010):
𝑦𝑦𝑡𝑡=𝛿𝛿0+𝛿𝛿1sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝛿𝛿2sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝑣𝑣𝑡𝑡 (3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T s a es
he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea de e minis ic
componen in he abo e model will be de e mined by choosing he op imal k among k alues
be ween 1 and 5, which will be he alue ha minimizes he esidual sum o squa es h ough
he leas squa es es ima o .
𝑣𝑣𝑡𝑡 =𝑦𝑦𝑡𝑡− [𝛿𝛿
0+𝛿𝛿
1𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)+𝛿𝛿
2𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)] (4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo . Th ee
sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
∆𝑣𝑣𝑡𝑡 = 𝛼𝛼1𝑣𝑣𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (5)
∆𝑣𝑣𝑡𝑡 = 𝜌𝜌𝑣𝑣𝑡𝑡−1(1−exp(−𝜃𝜃∆𝑣𝑣𝑡𝑡−𝑖𝑖
2))+∑𝛼𝛼𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (6)
∆𝑣𝑣𝑡𝑡 = 𝜆𝜆𝑣𝑣3𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (7)
He e θ>0 and 𝑢𝑢𝑡𝑡 ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o a uni
oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age is s a ed. I
he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y a ound he s uc u al
de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a iables
wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001). The ARDL
es equi es ha he dependen a iable I(1) and he independen a iables I(0) o I(1) be
s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL) me hod
in oduced by Sam e al. (2019) does no equi e he dependen a iable o be s a iona y a I(1).
To implemen he ARDL model in cases whe e he dependen a iable is s a iona y, McNown
e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged independen a iables.
The e o e, i can be u ilized especially when he dependen a iable is s a iona y a I(0) (Sam
e al., 2019). Th ee es s (F bounds es s a is ics, bounds es s a is ics, he Exogenous F-
Bound) a e employed o A-ARDL me hod. When hese h ee es s a e s a is ically signi ican
oge he , he e is a coin eg a ion ela ionship be ween he a iables. We es ablished he ARDL
model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡=𝛽𝛽0+∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝜋𝜋
𝑖𝑖=1 +∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽2𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽3𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +£1𝛥𝛥𝛥𝛥𝑡𝑡−1
+£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡 (8)
In Eq. (8), 𝑎𝑎0 deno es he cons an e m, while 𝛥𝛥 deno es he di e ence ope a o , and 𝑢𝑢𝑡𝑡
indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h. The null
and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
𝐻𝐻0: £1=£2=£3=£4= 0 (No coin eg a ion ela ionships exis ed be ween a iables)
(3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T
s a es he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea
de e minis ic componen in he abo e model will be de e mined by choosing he op imal
k among k alues be ween 1 and 5, which will be he alue ha minimizes he esidual
sum o squa es h ough he leas squa es es ima o .
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
96
o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es s a is ics a e calcula ed
as ollows (Ch is opoulos & Leon-Ledesman, 2010):
𝑦𝑦𝑡𝑡=𝛿𝛿0+𝛿𝛿1sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝛿𝛿2sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝑣𝑣𝑡𝑡 (3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T s a es
he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea de e minis ic
componen in he abo e model will be de e mined by choosing he op imal k among k alues
be ween 1 and 5, which will be he alue ha minimizes he esidual sum o squa es h ough
he leas squa es es ima o .
𝑣𝑣𝑡𝑡 =𝑦𝑦𝑡𝑡− [𝛿𝛿
0+𝛿𝛿
1𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)+𝛿𝛿
2𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)] (4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo . Th ee
sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
∆𝑣𝑣𝑡𝑡 = 𝛼𝛼1𝑣𝑣𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (5)
∆𝑣𝑣𝑡𝑡 = 𝜌𝜌𝑣𝑣𝑡𝑡−1(1−exp(−𝜃𝜃∆𝑣𝑣𝑡𝑡−𝑖𝑖
2))+∑𝛼𝛼𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (6)
∆𝑣𝑣𝑡𝑡 = 𝜆𝜆𝑣𝑣3𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (7)
He e θ>0 and 𝑢𝑢𝑡𝑡 ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o a uni
oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age is s a ed. I
he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y a ound he s uc u al
de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a iables
wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001). The ARDL
es equi es ha he dependen a iable I(1) and he independen a iables I(0) o I(1) be
s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL) me hod
in oduced by Sam e al. (2019) does no equi e he dependen a iable o be s a iona y a I(1).
To implemen he ARDL model in cases whe e he dependen a iable is s a iona y, McNown
e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged independen a iables.
The e o e, i can be u ilized especially when he dependen a iable is s a iona y a I(0) (Sam
e al., 2019). Th ee es s (F bounds es s a is ics, bounds es s a is ics, he Exogenous F-
Bound) a e employed o A-ARDL me hod. When hese h ee es s a e s a is ically signi ican
oge he , he e is a coin eg a ion ela ionship be ween he a iables. We es ablished he ARDL
model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡=𝛽𝛽0+∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝜋𝜋
𝑖𝑖=1 +∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽2𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽3𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +£1𝛥𝛥𝛥𝛥𝑡𝑡−1
+£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡 (8)
In Eq. (8), 𝑎𝑎0 deno es he cons an e m, while 𝛥𝛥 deno es he di e ence ope a o , and 𝑢𝑢𝑡𝑡
indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h. The null
and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
𝐻𝐻0: £1=£2=£3=£4= 0 (No coin eg a ion ela ionships exis ed be ween a iables)
(4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo .
Th ee sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es s a is ics a e calcula ed
as ollows (Ch is opoulos & Leon-Ledesman, 2010):
𝑦𝑦𝑡𝑡=𝛿𝛿0+𝛿𝛿1sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝛿𝛿2sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝑣𝑣𝑡𝑡 (3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T s a es
he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea de e minis ic
componen in he abo e model will be de e mined by choosing he op imal k among k alues
be ween 1 and 5, which will be he alue ha minimizes he esidual sum o squa es h ough
he leas squa es es ima o .
𝑣𝑣𝑡𝑡 =𝑦𝑦𝑡𝑡− [𝛿𝛿
0+𝛿𝛿
1𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)+𝛿𝛿
2𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)] (4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo . Th ee
sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
∆𝑣𝑣𝑡𝑡 = 𝛼𝛼1𝑣𝑣𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (5)
∆𝑣𝑣𝑡𝑡 = 𝜌𝜌𝑣𝑣𝑡𝑡−1(1−exp(−𝜃𝜃∆𝑣𝑣𝑡𝑡−𝑖𝑖
2))+∑𝛼𝛼𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (6)
∆𝑣𝑣𝑡𝑡 = 𝜆𝜆𝑣𝑣3𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (7)
He e θ>0 and 𝑢𝑢𝑡𝑡 ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o a uni
oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age is s a ed. I
he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y a ound he s uc u al
de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a iables
wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001). The ARDL
es equi es ha he dependen a iable I(1) and he independen a iables I(0) o I(1) be
s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL) me hod
in oduced by Sam e al. (2019) does no equi e he dependen a iable o be s a iona y a I(1).
To implemen he ARDL model in cases whe e he dependen a iable is s a iona y, McNown
e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged independen a iables.
The e o e, i can be u ilized especially when he dependen a iable is s a iona y a I(0) (Sam
e al., 2019). Th ee es s (F bounds es s a is ics, bounds es s a is ics, he Exogenous F-
Bound) a e employed o A-ARDL me hod. When hese h ee es s a e s a is ically signi ican
oge he , he e is a coin eg a ion ela ionship be ween he a iables. We es ablished he ARDL
model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡=𝛽𝛽0+∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝜋𝜋
𝑖𝑖=1 +∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽2𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽3𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +£1𝛥𝛥𝛥𝛥𝑡𝑡−1
+£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡 (8)
In Eq. (8), 𝑎𝑎0 deno es he cons an e m, while 𝛥𝛥 deno es he di e ence ope a o , and 𝑢𝑢𝑡𝑡
indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h. The null
and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
𝐻𝐻0: £1=£2=£3=£4= 0 (No coin eg a ion ela ionships exis ed be ween a iables)
(5)
o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es s a is ics a e calcula ed
as ollows (Ch is opoulos & Leon-Ledesman, 2010):
𝑦𝑦𝑡𝑡=𝛿𝛿0+𝛿𝛿1sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝛿𝛿2sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝑣𝑣𝑡𝑡 (3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T s a es
he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea de e minis ic
componen in he abo e model will be de e mined by choosing he op imal k among k alues
be ween 1 and 5, which will be he alue ha minimizes he esidual sum o squa es h ough
he leas squa es es ima o .
𝑣𝑣𝑡𝑡 =𝑦𝑦𝑡𝑡− [𝛿𝛿
0+𝛿𝛿
1𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)+𝛿𝛿
2𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)] (4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo . Th ee
sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
∆𝑣𝑣𝑡𝑡 = 𝛼𝛼1𝑣𝑣𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (5)
∆𝑣𝑣𝑡𝑡 = 𝜌𝜌𝑣𝑣𝑡𝑡−1(1−exp(−𝜃𝜃∆𝑣𝑣𝑡𝑡−𝑖𝑖
2))+∑𝛼𝛼𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (6)
∆𝑣𝑣𝑡𝑡 = 𝜆𝜆𝑣𝑣3𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (7)
He e θ>0 and 𝑢𝑢𝑡𝑡 ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o a uni
oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age is s a ed. I
he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y a ound he s uc u al
de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a iables
wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001). The ARDL
es equi es ha he dependen a iable I(1) and he independen a iables I(0) o I(1) be
s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL) me hod
in oduced by Sam e al. (2019) does no equi e he dependen a iable o be s a iona y a I(1).
To implemen he ARDL model in cases whe e he dependen a iable is s a iona y, McNown
e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged independen a iables.
The e o e, i can be u ilized especially when he dependen a iable is s a iona y a I(0) (Sam
e al., 2019). Th ee es s (F bounds es s a is ics, bounds es s a is ics, he Exogenous F-
Bound) a e employed o A-ARDL me hod. When hese h ee es s a e s a is ically signi ican
oge he , he e is a coin eg a ion ela ionship be ween he a iables. We es ablished he ARDL
model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡=𝛽𝛽0+∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝜋𝜋
𝑖𝑖=1 +∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽2𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽3𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +£1𝛥𝛥𝛥𝛥𝑡𝑡−1
+£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡 (8)
In Eq. (8), 𝑎𝑎0 deno es he cons an e m, while 𝛥𝛥 deno es he di e ence ope a o , and 𝑢𝑢𝑡𝑡
indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h. The null
and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
𝐻𝐻0: £1=£2=£3=£4= 0 (No coin eg a ion ela ionships exis ed be ween a iables)
(6)
o smoo h ansi ions as well as s uc u al changes. Fou ie ADF es s a is ics a e calcula ed
as ollows (Ch is opoulos & Leon-Ledesman, 2010):
𝑦𝑦𝑡𝑡=𝛿𝛿0+𝛿𝛿1sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝛿𝛿2sin(2𝜋𝜋𝜋𝜋𝑡𝑡
𝑇𝑇)+𝑣𝑣𝑡𝑡 (3)
He e, k deno es he equency numbe o he Fou ie unc ion, deno es he end, and T s a es
he numbe o obse a ions. The es s a is ic is calcula ed in h ee s eps.
In he i s s age, he app op ia e equency k* is de e mined. Thus, he nonlinea de e minis ic
componen in he abo e model will be de e mined by choosing he op imal k among k alues
be ween 1 and 5, which will be he alue ha minimizes he esidual sum o squa es h ough
he leas squa es es ima o .
𝑣𝑣𝑡𝑡 =𝑦𝑦𝑡𝑡− [𝛿𝛿
0+𝛿𝛿
1𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)+𝛿𝛿
2𝑠𝑠𝑠𝑠𝑠𝑠(2𝜋𝜋𝜋𝜋∗𝑡𝑡
𝑇𝑇)] (4)
In he second s age, he leas squa es esiduals in he i s s age a e es ed o uni oo . Th ee
sepa a e models wi h he ollowing linea and nonlinea s uc u e a e p oposed:
∆𝑣𝑣𝑡𝑡 = 𝛼𝛼1𝑣𝑣𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (5)
∆𝑣𝑣𝑡𝑡 = 𝜌𝜌𝑣𝑣𝑡𝑡−1(1−exp(−𝜃𝜃∆𝑣𝑣𝑡𝑡−𝑖𝑖
2))+∑𝛼𝛼𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (6)
∆𝑣𝑣𝑡𝑡 = 𝜆𝜆𝑣𝑣3𝑡𝑡−1 +∑𝛽𝛽𝑗𝑗∆
𝑝𝑝
𝑗𝑗=1 𝑣𝑣𝑡𝑡−𝑗𝑗 +𝑢𝑢𝑡𝑡 (7)
He e θ>0 and 𝑢𝑢𝑡𝑡 ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o a uni
oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age is s a ed. I
he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y a ound he s uc u al
de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a iables
wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001). The ARDL
es equi es ha he dependen a iable I(1) and he independen a iables I(0) o I(1) be
s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL) me hod
in oduced by Sam e al. (2019) does no equi e he dependen a iable o be s a iona y a I(1).
To implemen he ARDL model in cases whe e he dependen a iable is s a iona y, McNown
e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged independen a iables.
The e o e, i can be u ilized especially when he dependen a iable is s a iona y a I(0) (Sam
e al., 2019). Th ee es s (F bounds es s a is ics, bounds es s a is ics, he Exogenous F-
Bound) a e employed o A-ARDL me hod. When hese h ee es s a e s a is ically signi ican
oge he , he e is a coin eg a ion ela ionship be ween he a iables. We es ablished he ARDL
model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡=𝛽𝛽0+∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝜋𝜋
𝑖𝑖=1 +∑𝛽𝛽1𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽2𝛥𝛥𝛥𝛥𝛥𝛥𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽3𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +£1𝛥𝛥𝛥𝛥𝑡𝑡−1
+£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡 (8)
In Eq. (8), 𝑎𝑎0 deno es he cons an e m, while 𝛥𝛥 deno es he di e ence ope a o , and 𝑢𝑢𝑡𝑡
indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h. The null
and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
𝐻𝐻0: £1=£2=£3=£4= 0 (No coin eg a ion ela ionships exis ed be ween a iables)
(7)
He e θ>0 and u ep esen s he e o e m. I he null hypo hesis s a ing he exis ence o
a uni oo is ejec ed in he second s age, ha is, i he se ies is s a iona y, he hi d s age
is s a ed. I he null hypo hesis is ejec ed, i is concluded ha he se ies is s a iona y
a ound he s uc u al de e minis ic unc ion (Ch is opoulos & Leon-Ledesman, 2010).
3.2.2. Augmen ed ARDL me hod
ARDL coin eg a ion es , unlike adi ional coin eg a ion es s, p o ides analysis o a -
iables wi h di e en deg ees o s a iona i y (Pesa an & Shin, 1995; Pesa an e al., 2001).
The ARDL es equi es ha he dependen a iable I(1) and he independen a iables I(0)
o I(1) be s a iona y (Pesa an e al., 2001). Howe e , he augmen ed ARDL (A-ARDL)
me hod in oduced by Sam e al. (2019) does no equi e he dependen a iable o be
s a iona y a I(1). To implemen he ARDL model in cases whe e he dependen a iable is
s a iona y, McNown e al. (2018) and Sam e al. (2019) in oduced a new F- es o lagged
independen a iables. The e o e, i can be u ilized especially when he dependen a iable
is s a iona y a I(0) (Sam e al., 2019). Th ee es s (F bounds es s a is ics, bounds es
s a is ics, he Exogenous F-Bound) a e employed o A-ARDL me hod. When hese h ee
es s a e s a is ically signi ican oge he , he e is a coin eg a ion ela ionship be ween he
a iables. We es ablished he ARDL model as ollows.
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡
=𝛽𝛽
0
+∑𝛽𝛽
1
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑘𝑘
𝑖𝑖=1 +∑𝛽𝛽
1
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0 +∑𝛽𝛽
2
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0 +∑𝛽𝛽
3
𝛥𝛥𝐼𝐼𝛥𝛥
𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0
+£1𝛥𝛥𝛥𝛥𝑡𝑡−1 +£2𝛥𝛥𝛥𝛥𝑡𝑡−1+£3𝛥𝛥𝛥𝛥𝑡𝑡−1 +£4𝐼𝐼𝛥𝛥𝑡𝑡−1 +𝑢𝑢𝑡𝑡
(8)
In Eq. (8), a0 deno es he cons an e m, while Δ deno es he di e ence ope a o , and
u indica es he e o e m, espec i ely. k, l, m, and n e ms deno e op imal lag leng h.
The null and al e na i e hypo heses ega ding he ARDL bounds es a e p esen ed below.
H0: £1 = £2 = £3 = £4 = 0 (No coin eg a ion ela ionships exis ed be ween a iables)
H1: £1 = £2 = £3 = £4 ≠ 0 (Coin eg a ion ela ionships exis ed be ween a iables)
In Eq. (8), £1, £2, £3, £4 deno e he long- un pa ama e s. Howe e , in Eq. (9), he sho -
un and e o co ec ion model (ECM) is gi en below:
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
97
𝛥𝛥𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡
=𝛽𝛽
0
+∑𝛽𝛽
1
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑘𝑘
𝑖𝑖=1
+∑𝛽𝛽
2
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑙𝑙
𝑖𝑖=0
+∑𝛽𝛽
3
𝛥𝛥𝛥𝛥𝛥𝛥
𝑡𝑡−1
𝑚𝑚
𝑖𝑖=0
+∑𝛽𝛽4𝛥𝛥𝐼𝐼𝛥𝛥𝑡𝑡−1
𝑛𝑛
𝑖𝑖=0 +𝜚𝜚1𝛥𝛥𝛥𝛥𝐸𝐸𝑡𝑡−1 +𝑢𝑢𝑡𝑡
(9)
In Eq. (9), β1, β2, β3, and β4 s a e he sho - un pa ame e s, while ECM –1 deno es he
e o co ec ion e m wi h a lagged, and ϱ1 deno es he e o co ec ion pa ama e . The
ac ha he coe icien o ECM –1 is s a is ically signi ican and has a nega i e sign means
ha any long- un disequilib ium be ween he dependen a iable and a se o independen
a iables will con e ge back o he long- un equilib ium associa ion.
4. Empi ical esul s
Table 2 p o ides he s a is ical alues o he a iables used in he s udy. I appea s ha
he mean o GDP is 4.59%, he mean o TR is 13.43%, he mean o CE is 7.23%, he
mean o TE is 10.99% and he mean o IE is 1.78%. On he o he hand, he median o
GDP is 5.03%, he median o TR is 14.93%, he median o CE is 7.72%, he median o
TE is 11.50%, and he median o IE is 1.67%. In addi ion, he maximum alue o GDP
is 11.35%, while he minimum alue o GDP is -5.75%. The maximum alue o TR is
18.01%, while he minimum alue o TR is 7.85%. The maximum alue o CE is 8.83%,
while he minimum alue o CE is 4.33%. The maximum alue o TE is 22.80%, while he
minimum alue o TE is 3.79%. The maximum alue o IE is 2.81%, while he minimum
alue o TR is 0.85%. The JB no mali y dis ibu ion esul s o he a iables e eal ha
he a iables ha e a no mal dis ibu ion excep o he TR and CE a iables.
Table 2. Desc ip i e S a is ics o ha da a
GDP TR CE TE IE
Mean 4.59 13.43 7.23 10.99 1.78
Median 5.03 14.93 7.72 11.50 1.67
Maximum 11.35 18.01 8.83 22.80 2.81
Minimum -5.75 7.83 4.33 3.79 0.85
S d. De . 4.31 3.69 1.37 4.93 0.47
Skewness -0.77 -0.25 -0.87 0.42 0.10
Ku osis 2.96 1.37 2.47 2.58 1.99
Ja que-Be a 4.17 5.05 5.90 1.54 1.83
P obabili y 0.12 0.07 0.05 0.46 0.39
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
104
Acco ding o he esul s o he analyses, he posi i e sensi i i y o economic g ow h
in Tü kiye o public expendi u es in he sho and long un suppo s he Keynesian iscal
policy implemen a ion. Among he subcomponen s o public expendi u es, in es men
expendi u es a e he ac o ha a ec s economic g ow h he mos . This esul is consis en
wi h a p io i expec a ions and p e ious s udies showing ha public in es men expendi u es
lead o an inc ease in p oduc i e capaci y (A e is e al., 2021; A onso & Aubyn, 2019;
Sos illa-Ri e o & Rubio-Gue e o, 2022). In he long un, he e ec o ax inc eases on
economic g ow h is nega i e in line wi h he heo e ical expec a ion. This esul p o ides
impo an in o ma ion ha axes should no ha e a de e en e ec on economic, com-
me cial, in es men , and p oduc ion ac i i ies. In his sense, ou gene al esul s a e in line
wi h he esul s o Gio dano e al. (2007), Jawaid (2010), and Sel ane han e al. (2021) in
he li e a u e. These indings, which a e consis en wi h ou esea ch, indica e ha iscal
policy can be an e ec i e mac oeconomic policy ool and ha expendi u e-based iscal
adjus men s ha e a g ea e policy impac han ax-based iscal adjus men s. Mo eo e ,
ou esul s show ha he inc ease in ans e expendi u es hinde s economic g ow h. The
nega i e e ec o ans e expendi u es on economic g ow h can be aken as impo an
in o ma ion ha ans e expendi u es in Tü kiye a e no economic and a signi ican po ion
o ans e expendi u es consis s o du y losses and deb in e es paymen s. Howe e , he
indings o Du an (2022) on ans e expendi u es do no con i m his conclusion.
5. Conclusion and policy ecommenda ion
In his s udy, we empi ically in es iga e whe he iscal policy is e ec i e in s imula ing
economic g ow h in Tü kiye and, i so, wha is he size and du a ion o hese e ec s. In
doing so, we use he ax a iable and in pa icula he subcomponen s o public expendi u e
o analyse he mac oeconomic e ec s o iscal policy om 1980 o 2021. The es esul s
e eal ha while ax inc eases ha e a posi i e impac on economic g ow h in he sho
un and a nega i e impac in he long un, public expendi u e inc eases economic g ow h
in bo h he sho un and he long un. In his way, he s udy suppo s he implemen a ion
o Keynesian iscal policy while p o iding signi ican e idence o he e ec i eness o
expendi u e-based iscal policy. Mo eo e , he s udy shows ha o he Tu kish economy,
public in es men expendi u es ha e a signi ican and s ong e ec on economic g ow h,
while ans e expendi u es educe economic g ow h. The esul s o his s udy will be
pa icula ly use ul o p io i izing in es men spending and edesigning ans e spending.
5.1. Policy ecommenda ion
Ou esea ch indings p o ide a ious policy ecommenda ions ega ding he implemen-
a ion o iscal policy as a s ong mac oeconomic policy ool o economic g ow h in he
Tu kish economy. In o de o achie e sus ainable and high economic g ow h in Tü kiye,
i is necessa y o inc ease in es men expendi u es om he public budge as a ma e o
p io i y. In addi ion, emphasis should be placed on economically quali ied ans e ex-
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
105
pendi u es o p e en he dec ease in GDP due o ans e expendi u es and o p omo e he
inc ease in he sha e o ans e expendi u es in GDP. In his ega d, o inc ease economic
and social ans e expendi u es in Tü kiye, unp oduc i e ans e expendi u es ha do
no lead o an inc ease in he p oduc i e capaci y o he economy, such as du y losses
and deb in e es paymen s, should be educed i s . Fo his pu pose, p ima y su plus and
sound deb managemen s a egies should be pu sued o channel bo owed esou ces o
mo e p oduc i e a eas and educe he cos o bo owing. Gi en he long- e m nega i e
impac o axes on economic g ow h, i is impo an o es ablish a sound ax s uc u e ha
doesn’ discou age in es men and p oduc ion. In addi ion, since ex e nal shocks signi -
ican ly a ec he e ec i eness o iscal policy in Tü kiye, i is sugges ed ha a iables
ep esen ing ex e nal shocks should also be used in u u e s udies.
5.2. Limi a ion and u u e implemen a ion
Gi en he limi a ions o he s udy, we ecommend ha he scope o ou cu en esea ch
on he mac oeconomic impac o iscal policy be expanded o include addi ional a iables
such as income dis ibu ion, in la ion, and unemploymen . S udies in his di ec ion will
also be use ul in de e mining whe he economically g owing coun ies p oduce de elop-
men -o ien ed policies. Again using he case o Tü kiye, he scope o wo k can be ex ended
o include b oade ca ego ies o public expendi u e, axes, and non ax e enues, aking
in o accoun gene al go e nmen e enues and expendi u es ins ead o cen al go e nmen
e enues and expendi u es. Fu u e s udies could also include a la ge sample o coun ies
o assess he e ec i eness o he la ge iscal s imulus p o ided du ing COVID-19 and he
si ua ion in o he coun ies whe e iscal policy aces di e en challenges. I is expec ed
ha u he s udies on his aspec will p o ide a clea e pic u e o he mac oeconomic
e ec s o iscal policy.
Re e ences
A onso, A., & Aubyn, M. (2019). Economic g ow h, public, and p i a e in es men e u ns in 17 OECD econ-
omies. Po uguese Economic Jou nal, 18, 47-65. h ps://doi.o g/10.1007/s10258-018-0143-7
A onso, A., & Coelho, J. C. (2023). Public inances sol ency in he Eu o A ea. Economic Analysis and Policy,
77, 642-657. h ps://doi.o g/10.1016/j.eap.2022.12.027
Alesina, A., & A dagna, S. (2009). La ge changes in iscal policy: Taxes e sus spending. NBER Wo king
Pape , 15438.
Alesina, A., Fa e o, C., & Gia azzi, F. (2015). The ou pu e ec o iscal consolida ion plans. Jou nal o
In e na ional Economics, 96, 19-42. h ps://doi.o g/10.1016/ j.jin eco.2014.11.003
Al es, R. S., & Palma, A. A. (2023). The e ec i eness o iscal policy in B azil h ough he MIDAS Lens.
Jou nal o Policy Modeling, h ps://doi.o g/10.1016/j.jpolmod.2023.10.004
Ami i, K., Masba , R., Nazamuddin, B. S., & Aimon, H. (2023). Does ax e o mode a e he e ec o
go e nmen expendi u e on egional economic g ow h? A dynamic panel da a e idence om Indonesia.
Ekonomika, 102(2), 6-27. h ps://doi.o g/10.15388/Ekon.2023.102.2.1
A es is, P., Şen, H., & Kaya, A. (2021). Fiscal and mone a y policy e ec i eness in Tu key: A compa a i e
analysis. Panoeconomicus, 68(4), 415-439. h ps://doi.o g/10.2298/PAN190304019A
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
106
A in, K. P., B aun els, E., & Doppelho e , G. (2019). Re isi ing he g ow h e ec s o iscal policy: A
Bayesian model a e aging app oach. Jou nal o Mac oeconomics, 62, 1-16. h ps://doi.o g/10.1016/j.
jmac o.2019.103158
A izala, F., Gonzalez-Ga cia, J., Tsanga ides, C. G., & Yenice, M. (2021). The impac o iscal consolida ions on
g ow h in sub-Saha an A ica. Empi ical Economics, 61(1), 1-33. h ps://doi:10.1007/s00181-020-01863-x.
Blancha d, O., & Pe o i, R. (2002). An empi ical cha ac e iza ion o he dynamic e ec s o changes in go -
e nmen spending and axes on ou pu . Qua e ly Jou nal o Economics, 117(4), 1329-1368. h ps://www.
js o .o g/s able/4132480
Boug, P., B asch, T., Cappelen, A., Hamme sland, R., Hungnes, H., Kols ud, D., Sk e ing, J., S øm, B., &
Vig el, T. C. (2023). Fiscal policy, mac oeconomic pe o mance and indus y s uc u e in a small open
economy. Jou nal o Mac oeconomics, 76, 103524. h ps://doi.o g/10.1016/j.jmac o.2023.103524
Buchanan, J. M. (1983). The achie emen and he limi s o public choice in diagnosing go e nmen ailu e
and in o e ing bases o cons uc i e e o m in ana omy o go e nmen de iciencies. In H. Hanusch (Ed.),
Ana omy o Go e nmen De iciencies (pp.15-25). Sp inge -Ve lag.
Bunn, P., Le Roux, J., Reinold, K., & Su ico, P. (2018). The consump ion esponse o posi i e and nega i e
income shocks. Jou nal o Mone a y Economics, 96, 1-15. h ps://doi.o g/10.1016/j.jmoneco.2017.11.007
Calda a, F., & Kamps, C. (2017). The analy ics o SVARs: A uni ied amewo k o measu e iscal mul iplie s.
The Re iew o Economic S udies, 84, 1015-1040. h ps://doi.o g/10.1093/ es ud/ dx030
Cas o, D. F. (2006). The mac oeconomic e ec s o iscal policy in Spain. Applied Economics, 38, 913-924.
h ps://doi.o g/10.1080/00036840500369225
Ch is elis, D., Geo ga akos, D., Jappelli, T., Pis a e i, L., & Rooij, M. (2019). Asymme ic consump ion e ec s
o ansi o y income shocks. The Economic Jou nal, 129(622), 2322-2341. h ps://doi.o g/10.1093/ej/uez013
Ch is opoulos, D. K., & Leon-Ledesma, M.A. (2010). Smoo h B eaks and Non-linea Mean Re e sion:
Pos -B e on Woods Real Exchange Ra es. Jou nal o In e na ional Money and Finance, 29(6), 1076-1093.
h ps://doi.o g/10.1016/j.jimon in.2010.02.003
Chudik, A., Mohaddes, K., & Raissi, M. (2021). Co id-19 iscal suppo and i s e ec i eness. Economics
Le e s, 205, 109939. h ps://doi.o g/10.1016/j.econle .2021.109939
Donadelli, M., & G üning, P. (2021). Inno a ion dynamics and iscal policy: Implica ions o g ow h, as-
se p ices, and wel a e. The No h Ame ican Jou nal o Economics and Finance, 57, 1-31. h ps://doi.
o g/10.1016/j.naje .2021.101430
Dullien, S. (2012). Is new always be e han old? On he ea men o iscal policy in Keynesian models. Re iew
o Keynesian Economics, Inaugu al Issue, 5-23. h ps://doi.o g/10.4337/ oke.2012.01.01
Du an, F. (2022). Kamu ha camala ının ekonomik büyümeye e kisi: Tü kiye uygulaması. Akşehi Meslek
Yüksekokulu Sosyal Bilimle De gisi, 14, 25-42.
Fukuda, S. (2023). E alua ion o iscal policy using al e na i e GDP da a in Japan. Japan and he Wo ld
Economy, 67. h ps://doi.o g/10.1016/j.japwo .2023.101204
Gio dano, R., Momigliano, S., Ne i, S., & Pe o i, R. (2007). The e ec s o iscal policy in I aly: E idence
om a VAR model. Eu opean Jou nal o Poli ical Economy, 23, 707-733. h ps://doi.o g/10.1016/j.
ejpoleco.2006.10.005
Golpe, A., Sanchez-Fuen es, J., & Vides, J. C. (2023). Fiscal sus ainabili y, mone a y policy and economic
g ow h in he Eu o A ea: In sea ch o he ul ima e causal pa h. Economic Analysis and Policy, 78, 1026-
1045. h ps://doi.o g/10.1016/j.eap.2023.04.038
Goo jes B., & Haan J. (2022). P ocyclicali y o iscal policy in Eu opean Union coun ies. Jou nal o In e na-
ional Money and Finance, 120(1), 102276. h ps://doi.o g/10.1016/j.jimon in.2020.102276
Heimbe ge , P. (2023). The cyclical beha iou o iscal policy: A me a-analysis. Economic Modelling, 123,
106259. h ps://doi.o g/10.1016/j.econmod.2023.106259
Ahme Kös ekçi, Ali Celik. Modelling he Rela ionship Be ween Public Expendi u e, Tax Re enue and Economic G ow h...
107
Ho on, M., & El-Ganainy, A. (2019). Fiscal policy: Taking and gi ing away. IMF Finance and De elopmen .
h ps://www.im .o g/en/Publica ions/ andd/issues/Se ies/Back- o-Basics/Fiscal-Policy
Jawaid, S. T., Im iaz, A., & Naeemullah, S. M. (2010). Compa a i e analysis o mone a y and iscal policy: A
case s udy o Pakis an. Nice Resea ch Jou nal, 3(1), 58-67.
Ka agöz, K., & Keskin, R. (2016). Impac o iscal Policy on he mac oeconomic agg ega es in Tu key:
E idence om BVAR model. P ocedia Economics and Finance, 38, 408-420. h ps://doi.o g/10.1016/
S2212-5671(16)30212-X
Ka ahan, Ö., & Çolak, O. (2019). Examining he alidi y o Wagne ’s Law e sus Keynesian hypo hesis:
E idence om Tu key’s economy. Scien i ic Annals o Economics and Business, 66(1), 117-130. h ps://
doi.o g/10.2478/saeb-2019-0008
Ka aş, G., & Ka as, E. (2023). Tes ing con e gence o iscal policies in egions o Tu kiye. Ekonomika, 102(1),
26-40. h ps://doi.o g/10.15388/Ekon.2023.102.1.2
Keynes, J. M. (1936). The gene al heo y o employmen , ın e es , and Money. Ha cou , B ace and Company.
K uege , A. O. (1990). Go e nmen Failu es in De elopmen , Jou nal o Economic Pe spec i es, 4(3), 9-23.
h ps://doi.o g/10.1257/jep.4.3.9
Mawejje J., & Odhiambo, N. (2022). The de e minan s and cyclicali y o iscal policy: empi ical e idence om
Eas A ica. In e na ional Economics, 169(1) (2022), 55-70. h ps://doi.o g/10.1016/j.in eco.2021.12.001
McNown, R., Sam, C. Y., & Goh, S. K. (2018). Boo s apping he au o eg essi e dis ibu ed lag es o
coin eg a ion. Applied Economics, 50(13), 1509-1521. h ps://doi.o g/10.1080/00036846.2017.1366643
Moun o d, A., & Uhlig, H. (2009). Wha a e he e ec s o iscal policy shocks?. Jou nal o Applied Econo-
me ics, 24(6), 960-992. h ps://doi.o g/10.1002/jae.1079
Na ayan, P. K. (2005). The Sa ing and in es men nexus o China: E idence om coin eg a ion es s. Applied
economics, 37(17), 1979-1990. h ps://doi.o g/10.1080/00036840500278103
Na ayan, P. K., & Smy h, R. (2006). Wha de e mines mig a ion lows om low‐income o high‐income coun-
ies? An empi ical in es iga ion o Fiji–Us mig a ion 1972-2001. Con empo a y economic policy, 24(2),
332-342. h ps://doi.o g/10.1093/cep/byj019
Oni ade, S. T., Çe ik, S., E doğan, S., Asongu, S., & Bekun, F. V. (2020). An empi ical e ospec o he impac s
o go e nmen expendi u es on economic g ow h: New e idence om he Nige ian economy. Jou nal o
Economic S uc u es, 9(6), 1-13. h ps://doi.o g/10.1186/s40008-020-0186-7
Öze , M., & Ka agöl, V. (2018). Rela i e e ec i eness o mone a y and iscal policies on ou pu g ow h in
Tu key: An ARDL bounds es app oach, equilib ium. Qua e ly Jou nal o Economics and Economic
Policy, 13(3), 391-409. h p://dx.doi.o g/10.24136/eq.2018.019
Pa kyn, O., & Vehbi, T. (2014). The e ec s o iscal policy in New Zealand: E idence om a VAR model wi h
deb cons ain s. Economic Reco d, 90, 345-364. h ps://doi.o g/10.1111/1475-4932.12116
Pesa an, M. H., Shin, Y., & Smi h, R. J. (2001). Bounds es ing app oaches o he analysis o le el ela ionships.
Jou nal o Applied Econome ics, 16(3), 289-326. h ps://doi.o g/10.1002/jae.616
P esidency o he Republic o Tü kiye, Di ec o a e o S a egy and Budge (2023). h ps://www.sbb.go . /
ekonomik- e-sosyal-gos e gele /
Pula, L., & Elshani, A. (2018). The ela ionship be ween public expendi u e and economic g ow h in Koso o:
Findings om a Johansen co-in eg a ed es and a G ange causali y es . Ekonomika, 97(1), 47-62. h ps://
doi.o g/10.15388/Ekon.2018.1.11778
Ramey, V., & Zubai i, S. (2018). Go e nmen spending mul iplie s in good imes and in bad: e idence om
US his o ical da a. Jou nal o Poli ical Economics, 126, 850-901. h ps://www.jou nals.uchicago.edu/
doi/abs/10.1086/696277
Rompuy, P. V. (2021). Does subna ional ax au onomy p omo e egional con e gence? E idence om OECD
coun ies, 1995-2011. Regional S udies, 55(2), 234-244. h ps://doi.o g/10.1080/00343404.2020.1800623
ISSN 1392-1258 eISSN 2424-6166 Ekonomika. 2024, ol. 103(2)
108
Sam, C.Y., McNown, R., & Goh, S.K. (2019). An augmen ed au o eg essi e dis ibu ed lag bounds es o
coin eg a ion. Economic Modelling, 80, 130-141. h ps://doi.o g/10.1016/j.econmod.2018.11.001
Sel ana han, E. A., Sel ana han, S., & Jayasinghe, M. S. (2021). Re isi ing Wagne ’s and Keynesian’s p op-
osi ions and he ela ionship be ween sec o al go e nmen expendi u e and economic g ow h. Economic
Analysis and Policy, 71, 355-370. h ps://doi.o g/10.1016/j.eap.2021.05.005
Sos illa-Ri e oa, S., & Rubio-Gue e ob, J. J. (2022). The economic e ec s o iscal policy: Fu he e idence o
Spain. Qua e ly Re iew o Economics and Finance, 86, 305-313. h ps://doi.o g/10.1016/j.q e .2022.08.002
Te a, F. H. B., Filho, F. F., & Fonseca, P. C. D. (2021). Keynes on S a e and Economic De elopmen . Re iew
o Poli ical Economy, 33(1), 88-102. h ps://doi.o g/10.1080/09538259.2020.1823072
The Wo ld Bank (2023). Wo ld De elopmen Indica o s. h ps://da abank.wo ldbank.o g/sou ce/wo ld-de-
elopmen -indica o s
Weins ock, L. R. (2021). Fiscal policy: Economic e ec s. Cong essional Resea ch Se ice, h ps://c s epo s.
cong ess.go
Woldu, G. T., & Kano, I. S. (2023). Mac oeconomic e ec s o iscal consolida ion on economic ac i i y in
SSA coun ies. The Jou nal o Economic Asymme ies, 28, h ps://doi.o g/10.1016/j.jeca.2023.e00312.
Yang, W., Fid muc, J., & Ghosh, S. (2015). Mac oeconomic e ec s o iscal adjus men : A ale o wo app oaches.
Jou nal o In e na ional Money and Finance, 57, 31-60. h ps://doi.o g/10.1016/j.jimon in.2015.05.003
