Nga, Nguyen Thi Ngoc; Thach, Nguyen Ngoc
A icle
Reexamining he economic globaliza ion-wel a e s a e
nexus: a Bayesian mixed app oach o linea and non-linea
dynamics
Cogen Economics & Finance
P o ided in Coope a ion wi h:
Taylo & F ancis G oup
Sugges ed Ci a ion: Nga, Nguyen Thi Ngoc; Thach, Nguyen Ngoc (2024) : Reexamining he economic
globaliza ion-wel a e s a e nexus: a Bayesian mixed app oach o linea and non-linea dynamics,
Cogen Economics & Finance, ISSN 2332-2039, Taylo & F ancis, Abingdon, Vol. 12, Iss. 1, pp. 1-14,
h ps://doi.o g/10.1080/23322039.2024.2416987
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/321632
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/
Cogen Economics & Finance
ISSN: 2332-2039 (Online) Jou nal homepage: www. and online.com/jou nals/oae 20
Reexamining he economic globaliza ion-wel a e
s a e nexus: a Bayesian mixed app oach o linea
and non-linea dynamics
Nguyen Thi Ngoc Nga & Nguyen Ngoc Thach
To ci e his a icle: Nguyen Thi Ngoc Nga & Nguyen Ngoc Thach (2024) Reexamining
he economic globaliza ion-wel a e s a e nexus: a Bayesian mixed app oach o
linea and non-linea dynamics, Cogen Economics & Finance, 12:1, 2416987, DOI:
10.1080/23322039.2024.2416987
To link o his a icle: h ps://doi.o g/10.1080/23322039.2024.2416987
© 2024 The Au ho (s). Published by In o ma
UK Limi ed, ading as Taylo & F ancis
G oup
Published online: 17 Oc 2024.
Submi you a icle o his jou nal
A icle iews: 485
View ela ed a icles
View C ossma k da a
Ci ing a icles: 1 View ci ing a icles
Full Te ms & Condi ions o access and use can be ound a
h ps://www. and online.com/ac ion/jou nalIn o ma ion?jou nalCode=oae 20
GENERAL & APPLIED ECONOMICS | RESEARCH ARTICLE
Reexamining he economic globaliza ion-wel a e s a e nexus: a
Bayesian mixed app oach o linea and non-linea dynamics
Nguyen Thi Ngoc Nga and Nguyen Ngoc Thach
Depa men o Business Adminis a ion, Ho Chi Minh Uni e si y o Banking, Ho Chi Minh Ci y, Vie nam
ABSTRACT
P e ious esea ch on he globaliza ion-wel a e s a e nexus has ypically been con-
duc ed wi hin a linea mono onic amewo k, ocusing on wo p ima y hypo heses.
The e iciency hypo hesis p edic s ha globaliza ion sh inks he wel a e s a e’s size. In
con as , he compensa ion hypo hesis a gues ha globaliza ion inc eases demand o
social secu i y, leading o an expanded wel a e s a e. Empi ical e idence on his ela-
ionship is mixed, wi h mul icollinea i y sugges ed as one possible explana ion. This
s udy explo es he non-linea , non-mono onic aspec s o his nexus using a yea ly
panel o 10 ASEAN coun ies om 1970–2019, analyzed h ough a Bayesian mixed
eg ession. The s udy ocuses speci ically on he e ec s o economic globaliza ion. The
esul s demons a e a non-linea , non-mono onic ela ionship cha ac e ized by an
in e ed U-shape, whe e economic globaliza ion ini ially expands he wel a e s a e
bu , a e eaching a ce ain poin , leads o i s e enchmen . This ou come suppo s a
combina ion o bo h he compensa ion and e iciency hypo heses, aligning wi h he
p inciples o non-linea and complex sys em sciences. The s udy o e s a solid and eli-
able ounda ion o p edic ing long- e m globaliza ion p ocesses and o mula ing
comp ehensi e in eg a ion policies in ASEAN in he con ex o declining public expen-
di u es.
IMPACT STATEMENT
The s udy aims o explo e he mono onic and non-mono onic ela ionships be ween
economic globaliza ion and he wel a e s a e in ASEAN. Unlike simila esea ch, his
s udy employs a Bayesian hie a chical app oach o handle mul icollinea i y, e ealing
a non-linea and non-mono onic e ec o economic globaliza ion on he wel a e s a e,
consis en wi h he A mey cu e. The indings p o ide a obus ounda ion o o mu-
la ing long- e m globaliza ion policies o ASEAN coun ies.
ARTICLE HISTORY
Recei ed 11 July 2024
Re ised 9 Sep embe 2024
Accep ed 10 Oc obe 2024
KEYWORDS
Linea mono onic; non-
linea non-mono onic;
economic globaliza ion;
wel a e s a e; Bayesian
mixed eg ession; ASEAN
SUBJECTS
Economics; Finance;
In e na ional Rela ions;
In e na ional Poli ical
Economy
JEL CODES
F62
1. In oduc ion
The e a e wo p e ailing pe spec i es on how economic globaliza ion, o globaliza ion mo e b oadly,
impac s iscal au onomy, o en ep esen a i e o he wel a e s a e in heo e ical and empi ical analyses.
The i s pe spec i e, known as he e iciency hypo hesis, documen s ha ee ade and he mobili y o
p oduc ion ac o s inc ease compe i ion among na ional go e nmen s, p essu ing hem o educe ax
a es and iscal powe (U sp ung, 2008). On he con a y, he compensa ion hypo hesis posi s a posi i e
ela ionship be ween economic libe aliza ion and go e nmen expendi u e (Rod ik, 1998), a guing ha
globaliza ion heigh ens ex e nal isks, necessi a ing highe public spending as a o m o social insu ance.
Empi ical s udies in es iga ing he link be ween economic globaliza ion and he wel a e s a e ha e ypic-
ally elied on a linea , mono onic app oach, esul ing in indings consis en wi h one o he a o emen-
ioned hypo heses. Howe e , hese s udies o en p oduce mixed ou comes due o limi a ions such as
small sample sizes, a na ow ocus on speci ic dimensions o economic globaliza ion, o , mo e impo -
an ly, he biases inhe en in New onian classical sciences. Fi s , p ima acie e idence sugges ing a
CONTACT Nguyen Ngoc Thach [email p o ec ed] Ho Chi Minh Uni e si y o Banking, Ho Chi Minh Ci y, Vie nam
This a icle has been co ec ed wi h mino changes. These changes do no impac he academic con en o he a icle.
ß2024 The Au ho (s). Published by In o ma UK Limi ed, ading as Taylo & F ancis G oup
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 (h p://c ea i ecommons.o g/licenses/by/4.0/), 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 wo k is p ope ly ci ed. The e ms on which his a icle has been
published allow he pos ing o he Accep ed Manusc ip in a eposi o y by he au ho (s) o wi h hei consen .
COGENT ECONOMICS & FINANCE
2024, VOL. 12, NO. 1, 2416987
h ps://doi.o g/10.1080/23322039.2024.2416987
nega i e ela ionship be ween globaliza ion and go e nmen expendi u e is e iden om he apid
inc ease in in e na ional ade and capi al mobili y du ing he pe iod o ‘High globaliza ion’(Milano ic,
2016)o ‘New globaliza ion’(Baldwin, 2016) s a ing om 1990. This pe iod coincided wi h an agg ega e
decline in go e nmen consump ion as a sha e o GDP. These obse ed ends indica e a possible
e e sal in he ela ionship be ween globaliza ion and he wel a e s a e. Thus, by u ilizing a la ge da a
sample and emb acing he philosophy o nonlinea science and complex sys ems, we an icipa e a non-
linea , non-mono onic ela ionship be ween globaliza ion and he wel a e s a e. I is no ewo hy ha he
linea , mono onic assump ions o classical science ha e been undamen al o he social sciences since
New on’s publica ion o P incipia. Due o laws in models o o ecas ing, he unce ain ies o linea i y
p omp ed he de elopmen o he i s ma hema ical model o nonlinea phenomena in he 1960s
(Bu ke, 2009). To ou knowledge, aside om he ecen assessmen by Yong & Chi (2022), no s udies
ha e explo ed he non-linea , non-mono onic ela ionship be ween economic globaliza ion and he wel-
a e s a e. Addi ionally, no esea ch has speci ically ocused on ASEAN coun ies (Ande son & Obeng,
2021; Busemeye , 2009; Came on, 1978; D ehe e al., 2008; Ga e & Mi chell, 2001; Rod ik, 1997; Yong
& Chi, 2022, and many o he s). The ASEAN communi y is a ac ing g owing in e es o se e al easons,
including he in ol emen o many coun ies in he egion in global ade ag eemen s, he implemen a-
ion o he ASEAN Economic Communi y (AEC), high economic g ow h a es, and he egion’s p oximi y
o wo majo economic cen e s, China and India (Jamha i e al., 2021). Second, economic indica o s o
globaliza ion a e popula ly u ilized due o hei pe cei ed s ong impac on go e nmen size and he
wel a e s a e. I is wo h men ioning ha economic globaliza ion is a complex phenomenon ha canno
be su icien ly ep esen ed by a single indica o like inancial o ade libe aliza ion. This complexi y
p omp ed he c ea ion o he KOF globaliza ion indices by he Swiss Ins i u e o Economics (KOF s ands
o ‘Konjunk u o schungss elle’). Finally, mos impo an ly, ea lie equen is esea ch encoun e s issues
wi h mul icollinea i y, whe e some co a ia es a e s ongly co ela ed, leading o bias. To add ess mul i-
collinea i y, his esea ch employs a Bayesian mixed (o hie a chical) eg ession. The supe io i y o he
Bayesian amewo k o e equen is analysis in sol ing mul icollinea i y is highligh ed by Block e al.
(2011), Winship & Wes e n (2016), and ecen ly, Dan & Thach (2024). Fu he mo e, he mixed-e ec s
app oach allows o cap u ing a ia ions ac oss uni s (Nezlek, 2008).
F om he abo e discussion, we pose a subs an ial ques ion: Does a non-linea , non-mono onic o a
linea , mono onic ela ionship exis be ween economic globaliza ion and he wel a e s a e in an ASEAN
sample? The s udy aims o p o ide insigh s in o bo h he linea , mono onic and non-linea , non-mono-
onic e ec s o economic globaliza ion on he wel a e s a e using a panel o 10 ASEAN coun ies om
1970 o 2019. Compa ed o p e ious esea ch, we u ilize a longe ime ame o examine he ela ionship
beyond he adi ional mono onic amewo k. The indings e eal a non-linea , non-mono onic linkage
be ween economic globaliza ion and he wel a e s a e, suppo ing bo h he compensa ion and e iciency
hypo heses a di e en s ages o globaliza ion. This aligns wi h mode n philosophical pe spec i es ha
emphasize he non-linea i y and non-mono onici y in ela ions among phenomena and p ocesses in an
inc easingly globalized wo ld (Bu ke, 2009). The esul s ob ained in his s udy on non-linea , non-
mono onic pa e ns con ibu e signi ican ly o he in-dep h unde s anding o globaliza ion p ocesses’
impac s on na ional economies.
The es o he s udy is s uc u ed as ollows: Sec ion I discusses he mo i a ion, aim, and con ibu ion
o he s udy. Sec ion II e iews he analy ic amewo k and empi ical li e a u e. Sec ion III in oduces he
Bayesian mixed echnique, desc ibes he da a, and speci ies he baseline model. The main esul s a e
p esen ed and discussed in Sec ion IV, ollowed by he conclusion in Sec ion V.
2. Li e a u e e iew
Theo e ical and empi ical li e a u e on he globaliza ion-wel a e s a e ela ions is cha ac e ized by con-
lic ing, inconsis en and inconclusi e a gumen s and indings. Concep ually, he e a e wo con adic o y
hypo heses: compensa ion and e iciency. Empi ical analyses o he associa ion be ween globaliza ion
and he wel a e s a e (using go e nmen size as a p oxy o he wel a e s a e) p o ide mixed e idence.
The i s gene a ion o s udies, eme ging in he 1990s, ocused p ima ily on he ne e ec o globaliza-
ion on he wel a e s a e employing agg ega e da a, yielding ambiguous ou comes. In he mean ime,
2 N. NGUYEN THI NGOC AND T. NGUYEN NGOC
he second gene a ion o s udies applied mo e ad anced quan i a i e echniques and achie ed mo e
con incing esul s, u ilizing disagg ega ed da a and new a iables ep esen ing he scale o he wel a e
s a e and globaliza ion. Howe e , no inal consensus has been eached ega ding he ela ionship
be ween globaliza ion and he wel a e s a e. Re iewing 19 esea ch pape s published be ween 1995 and
2006, Gemmell e al. (2008) ound an equal numbe o s udies epo ing posi i e and nega i e co ela-
ions be ween ade o capi al low openness and go e nmen expendi u e. Mo e ecen ly, Ande son &
Obeng (2021), e iewing 13 empi ical wo ks published since 2006, e ealed simila ly mixed esul s.
Se e al ac o s con ibu e o hese con lic ing esul s, such as small sample sizes, a ious model speci ica-
ions (Ande son & Obeng, 2021), and di e en measu es o globaliza ion (Ash a e al., 2017; Kimako a,
2009; Meinha d & Po a ke, 2011), and di e se samples and obse a ions (Je e & Pa me e , 2015). Mos
c ucially, con en ional equen is echniques implemen ed in p e ious s udies su e om mul icollinea -
i y, whe e he key independen a iable, globaliza ion, and p ima y co a ia es a e s ongly co ela ed.
Mul icollinea i y can lead o biased es ima ion (Dan & Thach, 2024). Addi ionally, p e ious analyses
ocused on linea , mono onic connec ions, igno ing he possibili y o non-linea , non-mono onic ones. A
mono onic ela ionship is one whe e a iables mo e in a consis en di ec ion ela i e o each o he : as
one a iable inc eases o dec eases, he o he a iable also inc eases o dec eases, bu no necessa ily a
a cons an a e. In con as , a non-mono onic ela ionship allows he di ec ion o he ela ionship
be ween he a iables o change. This means ha as one a iable inc eases o dec eases, he o he a i-
able does no consis en ly inc ease o dec ease. A ela ionship, whe he linea o non-linea , can be
ei he mono onic o non-mono onic.
Nega i e linea , mono onic ela ions align wi h he e iciency hypo hesis. As emphasized by he clas-
sical app oach o poli ical economy, exempli ied by Adam Smi h and Da id Rica do, inc easing ade and
inancial libe aliza ion enhances compe i ion among na ional go e nmen s. The e iciency hypo hesis
posi s ha ade openness and ac o mobili y lead o a educ ion in go e nmen size, lowe ax e -
enue, and dec eased go e nmen expendi u e, ul ima ely e oding he wel a e s a e (Schulze & U sp ung,
1999). Du ing he globaliza ion p ocess, go e nmen s s a egically compe e in in e na ional ma ke s o
a ac o eign di ec in es men (FDI), skilled labo , and high- ech ad ancemen s. Consequen ly, global-
iza ion imposes cons ain s on na ional iscal au onomy, na ows axa ion bases, educes go e nmen
expendi u e, and al e s budge composi ion. By i s na u e, globaliza ion ende s all na ions ‘smalle .’
Small, open economies expe ience mo e in ense in e na ional compe i ion compa ed o la ge coun ies,
leading hem o depend less on capi al gains and co po a e axes. Consequen ly, as p edic ed by ax
compe i ion li e a u e, a global ace o he bo om in ax a es ensues (B e schge & He ich, 2002;
De e eux & G i i h 1998). F om his pe spec i e, a nega i e linea , mono onic connec ion exis s be ween
economic globaliza ion and he wel a e s a e. Empi ical esea ch suppo ing he e iciency poin o iew
e eals nega i e ela ionships be ween globaliza ion and he wel a e s a e. Fo example, Cusack &
Ga e (1992) iden i ied a nega i e ela ionship be ween in e na ional inancial in eg a ion and he
g ow h o go e nmen spending (excluding de ense spending) in a panel sample o 16 OECD coun ies
om 1955 o 1989. Rod ik (1997), u ilizing a yea ly panel da ase om OECD coun ies spanning 1966–
1991, explo ed a nega i e co ela ion be ween lagged ade libe aliza ion and bo h social expendi u e
and o al go e nmen expendi u e. His indings sugges ed ha in coun ies and pe iods wi h ully open
capi al accoun s, ade openness has a mo e signi ican ly nega i e co ela ion wi h social spending and
o al go e nmen expendi u e compa ed o coun ies wi h capi al es ic ions. Simila ly, Ga e &
Mi chell (2001), analyzing a sample o 18 OECD coun ies du ing 1961–1993, concluded ha ade libe -
aliza ion nega i ely a ec s go e nmen expendi u e, hough no go e nmen consump ion and social
secu i y ans e s. Using a sample o 17 OECD coun ies om 1961 o 1993, Ki el & Winne (2005) ound
ha FDI is nega i ely ela ed o he wel a e s a e, as measu ed by o al go e nmen expendi u e as a
sha e o GDP. Howe e , he e ec s o o he a iables, such as ade openness and impo s om low-
wage coun ies, a e non-signi ican . Mo e ecen ly, Busemeye (2009) also disco e ed a nega i e ela ion-
ship be ween ade libe aliza ion and go e nmen expendi u e in an OECD sample.
Con e sely, he compensa ion pe spec i e sugges s a posi i e linea , mono onic ela ionship, indica -
ing ha go e nmen scope is b oade in mo e open economies. This pe spec i e posi s ha inc eased
in e na ional in eg a ion heigh ens demand o social secu i y, p omp ing go e nmen s o expand social
wel a e expendi u es (Rod ik, 1998). G ea e in e na ional in eg a ion inc eases ex e nal isks, leading o
COGENT ECONOMICS & FINANCE 3
ins abili y and job luc ua ions, he eby aising he demand o social insu ance and compensa ion o
hose ad e sely a ec ed by globaliza ion (Kim, 2009; Rod ik, 1998). Poli icians espond o hese demands
by inc easing social wel a e spending o secu e e-elec ion (D ehe e al., 2008; U sp ung, 2008).
P o iding s ong suppo o his hypo hesis, Came on (1978) highligh ed ha inc eased ade depend-
ence signi ican ly con ibu ed o he g ow h o he public sec o in 18 indus ial coun ies be ween 1960
and 1975. Simila ly, Rod ik (1998) iden i ied a s ong posi i e in luence o ade libe aliza ion on go e n-
men consump ion in his analysis o 125 coun ies om 1985–1989 and 103 coun ies om 1990–1992.
Vaubel (2000,2005), Meinha & Po a ke (2011), and Fe ei a & Oli ei a (2019) also ound posi i e co e-
la ions be ween globaliza ion and social spending. Speci ically, Meinha & Po a ke (2011) e isi ed he
e idence by analyzing an annual panel da ase o 186 coun ies om 1970 o 2004, u ilizing da a om
he Penn Wo ld Tables on go e nmen sec o size and he KOF index o globaliza ion. Thei esul s indi-
ca e ha globaliza ion led o an inc ease in go e nmen sec o s wo ldwide.
In addi ion o he wo a o emen ioned di ec ions in he li e a u e, some s udies epo mixed o non-
signi ican indings. Kau man & Segu a-Ubie go, u ilizing da a om 14 La in Ame ican coun ies be ween
1973 and 1997, demons a ed ha ade openness and capi al libe aliza ion nega i ely impac ed social
secu i y spending. Howe e , hey ound ha capi al libe aliza ion posi i ely in luenced educa ion and
heal h spending. B e schge & He ich (2002) concluded ha while he e iciency hypo hesis explains ax
pa e ns, he compensa ion hypo hesis explains spending pa e ns in 14 OECD coun ies om 1967–
1996. In a ee alua ion o Ga e & Mi chell (2001) using ad anced me hodologies, Ki el & Winne
(2005) ound no suppo o ei he he e iciency o compensa ion hypo hesis. Ins ead, hei indings
indica e ha domes ic economic and demog aphic ac o s, such as unemploymen a es and depend-
ency a ios, ha e a mo e signi ican in luence on go e nmen expendi u es. D ehe e al. (2008) ound
ha globaliza ion does no impac budge composi ion, using da a om 60 coun ies (1971–2001) and
10 OECD coun ies (1991–2000) wi h ou di e en p oxies o globaliza ion. Recen indings showed
ha he e is no subs an ial e idence o a global ace o he bo om in ax a es (Pl€
umpe e al., 2009).
Ash a e al. (2017) ound ha g een ield FDI had a posi i e e ec on go e nmen consump ion, while
M&A FDI had no e ec , based on a sample o 130 de eloped and de eloping coun ies om 2003 o
2011. In a s udy by Kim e al., ade globaliza ion posi i ely in luenced o al go e nmen spending,
whe eas inancial globaliza ion had a nega i e impac , in an analysis o 53 OECD and non-OECD coun-
ies om 1980 o 2011. Ande son & Obeng (2021) e ealed ha de ju e ade globaliza ion gene ally
inc eased consump ion spending, aligning wi h he compensa ion hypo hesis. In con as , de ju e inan-
cial globaliza ion ended o dec ease consump ion spending, suppo ing he e iciency hypo hesis. Mos
ecen ly, Yong & Chi (2022) eassessed he impac o ade openness and inancial libe aliza ion on he
wel a e s a e, inding an in e ed U-shaped e ec o ade libe aliza ion on he wel a e s a e bu no la ge
in luence om inancial libe aliza ion. Tekin e al. (2023) ound suppo o bo h he compensa ion and
e iciency hypo heses in 11 ansi ion economies in he Cen al and Eas e n Eu opean (CEE) egion om
1996 o 2021.
F om he abo e analyses, h ee gaps a e iden i ied:
i. Ea lie s udies ha e p ima ily ocused on linea , mono onic ela ionships be ween o e all globaliza-
ion o i s sub-indices and he wel a e s a e, o e looking po en ial non-linea , non-mono onic in e -
ac ions. The di e gen and e en con adic o y ends o economic globaliza ion and go e nmen
spending—wi h inc easing economic openness and a sh inking sha e o go e nmen consump ion
in GDP o e ecen decades—sugges a non-linea , non-mono onic ela ionship be ween hese
p ocesses. Thus, we hypo hesize a non-linea ela ionship ollowing an in e ed U-shaped pa e n,
as mode n complex sciences sugges .
ii. The con en ional equen is echniques employed in a se ies o p io globaliza ion-spending s ud-
ies canno e ec i ely handle mul icollinea i y issues, leading o biased and non- obus ou comes.
‘Jus as exac collinea i y leads o non-iden i ica ion o pa ame e s, high collinea i y can be consid-
e ed as weak iden i ica ion o he pa ame e s’(Pesa an & Smi h, 2019).
iii. The e is a lack o empi ical analysis on he link be ween economic globaliza ion and he wel a e
s a e in he ASEAN con ex , making i challenging o design a comp ehensi e globaliza ion s a egy
o hese na ions.
4 N. NGUYEN THI NGOC AND T. NGUYEN NGOC
3. Bayesian mixed eg ession, model and da a
Al hough he implemen a ion o Bayesian echniques o da a analysis has e olu ionized many ields,
om gene ics o mac oeconomics since he 1990s, he Bayesian amewo k has been no ably absen in
empi ical esea ch on he nexus be ween globaliza ion and he wel a e s a e. In gene al, ou wo k in o-
duces esea che s in his a ea o Bayesian es ima ion o add ess model unce ain y a ising om s a is ical
complexi ies. Bayesian analysis o e s a mo e lexible and in ui i e amewo k o dealing wi h unce -
ain y and o e comes se e al d awbacks o he equen is app oach (Thach e al., 2019; B iggs, 2023).
Mos no ably, Bayesian in e ences p o ide mo e obus es ima es by combining p io knowledge wi h
a ailable da a and o e ing an en i e pos e io dis ibu ion o he pa ame e s. In Bayesian in e ences, he
da a x, which a ealiza ion o X ðxjHÞ, is in eg a ed wi h p io in o ma ion encapsula ed in a p io dis-
ibu ion wi h densi y pH
ðÞ
:This in eg a ion esul s in he pos e io dis ibu ion pHjx
ðÞ
, summa izing he
upda ed belie s. The pos e io dis ibu ion is de i ed om he join dis ibu ion ðxjHÞpH
ðÞ
using Bayes’
o mula:
pHjx
ðÞ
¼ xjH
ðÞ
pH
ðÞ
Ð ðxjHÞpH
ðÞ
dH(1)
Whe e mx
ðÞ
¼Ð ðxjHÞpH
ðÞ
dH is he ma ginal densi y o X(see Be ge , 1985; Be na do & Smi h, 1994;
Robe & Casella, 2004).
Speci ically, Bayesian me hods handle mul icollinea i y mo e e ec i ely han equen is me hods
(Block e al., 2011; Dan & Thach, 2024; Pesa an & Smi h, 2019; Winship & Wes e n, 2016). In equen is
eg ession, exac collinea i y be ween eg esso s makes hei indi idual coe icien s uniden i ied (Pesa an
& Smi h, 2019), leading o uns able coe icien es ima es and in la ed s anda d e o s. Wi h collinea i y,
small de ia ions om his assump ion can cause signi ican changes in es ima es (Winship & Wes e n,
2016). As emphasized by Pesa an & Smi h, (2019), gi en an in o ma i e p io , Bayesian pos e io means
a e well de ined. Fu he mo e, a Bayesian mixed (hie a chical) app oach wi h speci ic p io s e ec i ely
add esses non-s a iona i y in da a. Non-s a iona i y, which can esul om ends, seasonal e ec s, o
s uc u al b eaks, may lead o spu ious eg ession, p oducing misleading esul s wi h a i icially high R-
squa ed alues and signi ican -s a is ics ha do no e lec ue ela ionships. Bayesian mixed models
can manage non-s a iona i y by allowing pa ame e s o a y o e ime o ac oss di e en egimes.
Addi ionally, inco po a ing in o ma i e p io s in Bayesian in e ence helps egula ize es ima es and in e-
g a e p io knowledge, making i a aluable ool o handling non-s a iona i y (Spall, 1988).
The s udy adop s a Bayesian mixed eg ession o in es iga e he impac o economic globaliza ion on
he wel a e s a e. The baseline model is speci ied as ollows:
Expendi u ei ¼a0þa1Globaliza ionEci þa2Xi þRe ec s þei :(2)
To cap u e he non-linea and non-mono onic na u e in he globaliza ion-spending ela ionship, we
inco po a e he quad a ic e m o Globaliza ionEc in o he model (2).
Expendi u ei ¼b0þb1Globaliza ionEci þb2Globaliza ionEci 2þb3Xi þRe ec s þei :(3)
whe e Expendi u e is he wel a e s a e, Globaliza ionEc is he economic measu e o he KOF globaliza ion,
Xis a ec o o con ols, eis he andom e o , Re ec s is andom e ec s, and ,ia e yea and coun y,
espec i ely.
Ou analysis u ilizes he 2023 KOF globaliza ion da ase , co e ing he pe iod om 1970 o 2021, along
wi h he 2023 Wo ld De elopmen Indica o s (Wo ld Bank) and annual panel da a om he mos ecen
Penn Wo ld Tables, e sion 10.01, upda ed only o 2019. These da ase s include he dependen a iable,
wel a e s a e (Expendi u e), he key p edic o , economic globaliza ion (Globaliza ionEc), and a ious base-
line con ol a iables a ailable om 1950 o 2019 (da a on he KOF globaliza ion indices has been a ail-
able since 1970). Hence, ou panel da ase encompasses 10 ASEAN coun ies (B unei, Indonesia,
Cambodia, Laos, Myanma , Malaysia, he Philippines, Singapo e, Thailand, and Vie nam) om 1970 o
2019.
In Table 1, he measu e o he wel a e s a e is de ined as he go e nmen ’s sha e o eal GDP pe cap-
i a, as compiled by he Penn Wo ld Tables, excluding ans e paymen s. The key p edic o ,
COGENT ECONOMICS & FINANCE 5
Globaliza ionEc, is he economic dimension o he KOF globaliza ion index. This measu e is employed o
concen a e on he economic impac s o globaliza ion. No e ha he 2023 KOF index syn hesizes 24 a -
iables in o a comp ehensi e index and h ee sub-indices, each ep esen ing he economic, social, and
poli ical dimensions o globaliza ion. These a iables a e o ganized in o six g oups: (1) ac ual lows o
ade, in es men , and income paymen s o o eign na ionals, (2) es ic ions on in e na ional ade
and capi al accoun s, (3) da a on pe sonal con ac s wi h indi iduals in o eign coun ies, (4) da a on
in o ma ion lows, (5) da a on in e na ional cul u al in eg a ion, and (6) da a on in e na ional poli ical
in eg a ion. The economic globaliza ion sub-index comp ises he i s wo g oups, he social globaliza-
ion sub-index includes he nex h ee g oups, and he poli ical globaliza ion sub-index consis s o he
las g oup. Collec i ely, hese componen s o m he o e all index (Gygli e al., 2019).
D awing on Alesina & Waczia g (1998), D ehe e al. (2008), and Po a ke (2009), he s udy inco po-
a es eal pe capi a GDP (IncomePc), age dependency a io (Dependence), o al popula ion (Popula ion),
and u baniza ion a e (U baniza ion). Real pe capi a GDP p o ides insigh in o he o e all socio-eco-
nomic condi ion. Acco ding o Wagne ’s law, inc eases in he le els o GDP pe capi a lead o inc eases
in go e nmen expendi u e as a sha e o GDP. As a coun y becomes weal hie , he demand o public
se ices such as educa ion, heal hca e, in as uc u e, and social wel a e ends o inc ease. Highe income
le els enable ci izens o expec and demand be e and mo e comp ehensi e public se ices. Besides,
economic g ow h o en b ings abou g ea e complexi y in economic ac i i ies and social s uc u es,
necessi a ing mo e go e nmen in e en ion in e ms o egula ion, edis ibu ion, and p o ision o pub-
lic goods. The age dependency a io adjus s o demog aphic s uc u e and de elopmen . As he age
dependency a io inc eases, meaning he e a e mo e dependen s (young and elde ly) ela i e o he
wo king-age popula ion, i leads o inc eased go e nmen spending o se e al easons. Olde popula-
ions ypically equi e mo e heal hca e and social se ices, leading o highe public expendi u e on hese
se ices. Fu he mo e, a highe p opo ion o young dependen s necessi a es mo e in es men in educa-
ion and ela ed se ices. Las ly, inc eased numbe s o elde ly dependen s equi e mo e spending on
pensions and social secu i y p og ams. O e all, mo e dependen s gene ally inc ease he need o a ious
wel a e p og ams o suppo non-wo king popula ions. We measu e u baniza ion by he p opo ion o
he popula ion li ing in u ban a eas. As expec ed, go e nmen expendi u e will inc ease wi h ising
u baniza ion. The a ionale is s aigh o wa d: highe u baniza ion can lead o conges ion, po en ially
dec easing ci izen wel a e, which may necessi a e inc eased go e nmen expendi u e o o se he wel-
a e loss. Howe e , spending on non- i al public goods like oads and s ee ligh ing may dec ease wi h
g owing u baniza ion due o economies o scale (Ande son & Obeng, 2021). To al popula ion se es as a
p oxy o coun y size, e lec ing ha la ge popula ions end o co ela e wi h a smalle go e nmen
sha e o GDP (Alesina & Waczia g, 1998). Rega ding o al popula ion, Alesina & Waczia g (1998) sugges
ha coun ies wi h la ge popula ions end o ha e lowe go e nmen consump ion spending as a sha e
o GDP o se e al easons. Fi s ly, he pe capi a cos o p o iding non- i al goods is educed in la ge
popula ions due o economies o scale (see Rod ik, 1998; Je e & Pa me e , 2015). Addi ionally, la ge
Table 1. In e p e ing a iables in he Bayesian mixed model.
Va iables P oxies Pe iod No a ion Da a sou ce
Dependen Wel a e s a e Na u al loga i hm o sha e o
go e nmen consump ion in
eal GDP
1970–2019 Expendi u e PWT 10.01 da ase
Independen Economic globaliza ion Na u al loga i hm o he economic
sub-index o he KOF globaliza ion
index
1970–2019 Globaliza ionEc Swiss Ins i u e o
Economics
Real pe capi a income Na u al loga i hm o eal pe
capi a GDP
1970–2019 IncomePc PWT 10.01 da ase
Age dependence a e Na u al loga i hm o he sum o he
popula ion in he ages 0–14 yea s
and he popula ion in he ages
65þas a pe cen age o he o al
popula ion
1970–2019 Dependence 2023WDI da ase
Coun y size Na u al loga i hm o o al popula ion 1970–2019 Popula ion PWT 10.01 da ase
U ban popula ion Na u al loga i hm o u baniza ion
a e
1970–2019 U baniza ion 2023 WDI da ase
Sou ce: Au ho s.
6 N. NGUYEN THI NGOC AND T. NGUYEN NGOC
popula ions exhibi mo e he e ogeneous p e e ences o public goods p o ision. The o e all e ec
depends on balancing he cos s o inc eased he e ogenei y in p e e ences agains he bene i s o lowe
pe capi a cos s in public goods p o ision, wi h he la e po en ially ou weighing he o me .
Consequen ly, he impac o o al popula ion on go e nmen spending can be ei he posi i e o nega-
i e (Alesina & Waczia g, 1998; Rod ik, 1998). Log ans o ma ions a e applied o all a iables o no mal-
ize skewed da a, educe he impac o ou lie s, and s abilize a iance (Table 1). Mo eo e , aking he
loga i hm can s abilize he a iance.
Fo se ing p io s in he model, s anda d no mal p io s a e used. Gi en ha all pa ame e s can ake
ei he posi i e o nega i e alues, we assume a no mal dis ibu ion wi h a mean o 0 and a mildly
in o ma i e a iance o 1 o he Globaliza ionEc eg esso , simila o he emaining co a ia es (Block
e al., 2011). A non-in o ma i e in e se gamma (0.01, 0.01) is assigned o he a iance componen s,
including he a iance o andom e ec s and he o e all a iance, as sugges ed by Van de Schoo e al.
(2015).
P io dis ibu ions o he model pa ame e s a e as ollows:
aNo mal 0, 1
ðÞ
,
02Igammað0:01, 0:01Þ,
12Igamma 0:01, 0:01
ðÞ
,
whe e ais a ec o o s uc u al pa ame e s and 02, 12a e he a iance o andom e ec s and he
o e all a iance, espec i ely.
4. Simula ion ou comes
4.1. Desc ip i e s a is ics and pai wise co ela ions
The desc ip i e s a is ics able (Table 2) p o ides a comp ehensi e o e iew o he key a iables, high-
ligh ing hei a e age alues, a iabili y, and ange, which is c ucial o unde s anding he dis ibu ion
and cha ac e is ics o he da a, as well as o in e p e ing subsequen analy ical esul s. Expendi u e, ep-
esen ing go e nmen consump ion spending as a sha e o GDP, has an a e age alue o app oxima ely
15.77%. The s anda d de ia ion o 5.63% indica es mode a e a iabili y a ound he mean. The minimum
alue is 4.82%, and he maximum is 31.88%, showing a b oad ange o go e nmen expendi u e sha es
ac oss di e en coun ies. The Economic Globaliza ion Index (Globaliza ionEc) has a mean sco e o 50.93,
sugges ing a mode a e le el o economic globaliza ion on a e age among he sampled coun ies. The
s anda d de ia ion o 19.70 shows subs an ial a ia ion. The index anges om 17.68 o 94.58, indica ing
ha some coun ies a e highly globalized while o he s a e much less so. Pe capi a income (IncomePc)
has a mean o $14,058.82, bu he high s anda d de ia ion o $22,210.67 indica es a signi ican dispa i y
in income le els ac oss coun ies. The minimum alue is $533.03, and he maximum is $99,005.20, high-
ligh ing he wide income ange om e y low-income o e y high-income coun ies. The dependence
a io (Dependence) has a mean o 64.23, sugges ing ha , on a e age, he e a e 64 dependen s o e e y
100 wo king-age indi iduals. The s anda d de ia ion o 17.77 indica es mode a e a iabili y. The depend-
ence a io anges om 27.31 o 98.21, e lec ing a subs an ial ange o dependency bu dens ac oss
coun ies. Popula ion size (in millions) has an a e age o app oxima ely 48.94 million people, wi h a high
s anda d de ia ion o 58.46 million, indica ing conside able a ia ion in coun y sizes. The smalles popu-
la ion is a ound 0.22 million, and he la ges is abou 270.63 million, demons a ing a as ange o
popula ion sizes among he coun ies s udied. U baniza ion, exp essed as he pe cen age o he
Table 2. Desc ip i e s a is ics.
Va iable Obs Mean S d. de Min Max
Expendi u e 486 0.1576643 0.0563471 0.048244 0.318799
Globaliza ionEcI 486 50.92878 19.69739 17.6842 94.58176
IncomePc 486 14058.82 22210.67 533.028 99005.2
Dependence 486 64.2271 17.76848 27.31119 98.2091
Popula ion 486 48.94011 58.46113 0.218175 270.6256
U baniza ion 486 41.869 25.94483 4.477 100
Sou ce: Au ho s.
COGENT ECONOMICS & FINANCE 7
Nezlek, J. B. (2008). An in oduc ion o mul ile el modeling o social and pe sonali y psychology. Social and
Pe sonali y Psychology Compass,2(2), 842–860. h ps://doi.o g/10.1111/j.1751-9004.2007.00059.x
Pesa an, M. H., & Smi h, R. P. (2019). A Bayesian analysis o linea eg ession models wi h highly collinea eg esso s.
Econome ics and S a is ics,11,1–21. h ps://doi.o g/10.1016/j.ecos a.2018.10.001
Pl€
umpe , T., T oege , V. E., & Winne , H. (2009). Why is he e no ace o he bo om in capi al axa ion? Tax compe i-
ion among coun ies o unequal size, di e en le els o budge igidi ies and he e ogeneous ai ness no ms.
In e na ional S udies Qua e ly,53(3), 761–786. h ps://doi.o g/10.1111/j.1468-2478.2009.00555.x
Po a ke, N. (2009). Did globaliza ion es ic pa isan poli ics? An empi ical e alua ion o social expendi u es in a
panel o OECD coun ies. Public Choice,140(1–2), 105–124. h ps://doi.o g/10.1007/s11127-009-9414-2
Robe , C. P., & Casella, G. (2004). Mon e Ca lo s a is ical me hods. Sp inge Science þBusiness Media, LLC. second
edi ion.
Rod ik, D. (1997). Has globalisa ion gone oo a ?. Ins i u e o In e na ional Economics.
Rod ik, D. (1998). Why do mo e open Economies ha e bigge go e nmen s? Jou nal o Poli ical Economy,106(5),
997–1032. h ps://doi.o g/10.1086/250038
Rud a, N. (2002). Globaliza ion and he decline o he wel a e s a e in less-de eloped Coun ies. In e na ional
O ganiza ion,56(2), 411–445. h ps://doi.o g/10.1162/002081802320005522
Shalizi, C. R. (2019). The T u h abou Linea Reg ession. Re ie ed om he in e ne : h ps://www.s a .cmu.edu/
cshalizi/TALR/
Schulze, G. G., & U sp ung, H. W. (1999). Globaliza ion o he economy and he na ion-s a e. The Wo ld Economy,
22(3), 295–352. h ps://doi.o g/10.1111/1467-9701.00205
Shalizi C.R. (2019). The T u h abou Linea Reg ession. Accessed 18 Ap il, 2014. h ps://www.s a .cmu.edu/cshalizi/
TALR/TALR.pd
Spall, J. (1988). Bayesian analysis o ime se ies and dynamic models. 1s edi ion. CRC P ess.
Tekin, A., C¸ına ,
_
I. T., Sa
gdıc¸, E. N., & Yıldız, F. (2023). T ade openness and sus ainable go e nmen size: E idence
om Cen al and Eas e n Eu opean Coun ies. Sus ainabili y,15(15), 11836. h ps://doi.o g/10.3390/su151511836
Thach, N. N., Anh, L. H., & An, P. T. H. (2019). The e ec s o public expendi u e on economic g ow h in Asia
Coun ies: A Bayesian model a e aging app oach. Asian Jou nal o Economics and Banking,3(1), 126–149.
U sp ung, H. W. (2008). Globaliza ion and he wel a e s a e. In Du lau , S. N., and Blume, L.E. (Eds.) The New Palg a e
Dic iona y o Economics. Second Edi ion. Palg a e Macmillan (K€
oln).
Van de Schoo , R., B oe e, J. J., Pe yck, K. H., Zonde an-Zwijnenbu g, M., & Van Loey, N. E. (2015). Analyzing small
da a se s using Bayesian es ima ion: The case o pos auma ic s ess symp oms ollowing mechanical en ila ion
in bu n su i o s. Eu opean Jou nal o Psycho auma ology,6.h ps://doi.o g/10.3402/ejp . 6.25216
Vaubel, R. (2000). In e na ionale poli ische We bewe b: Eine eu op€
aische We bewe bsau sich €
u Regie ungen
und die empi ische E idenz. In Schenk, K.-E., Schmid chen, D., S ei , M. E. and Vanbe g, V. (Eds.) Jah buch €
u
Neue Poli ische
€
Okonomie. (pp.280–309). Moh -Siebeck.
Vaubel, R. (2005). Sozialpoli ische Konsequenzen de Globalisie ung-Theo ie und Empi ie. In F ey ag, A. (Ed.)
Wi scha liche S uk u wandel, na ionale Wi scha spoli ik und poli ische Ra ionali €
a . (pp. 143–158). K€
olne
Uni e si €
a s e lag GmbH.
Ve beke, G., & Molenbe ghs, G. (2000). Linea Mixed Models o Longi udinal Da a. Sp inge .
VNTR. (2022). ASEAN - A success ul economic egion model in globaliza ion. Re ie ed om h ps:// n .moi .go . n/
news/asean-a-success ul-economic- egion-model-in-globaliza ion
Wagne , A. (1893). G undlegungde poli schen okonomie. C. F. Win e .
Winship, C., & Wes e n, B. (2016). Mul icollinea i y and model misspeci ica ion. Sociological Science,3, 627–649.
h ps://doi.o g/10.15195/ 3.a27
Yong, M., & Chi, Y. (2022). Openness and go e nmen size: Re isi ing he ela ionship using a la ge c oss-coun y
panel. In e na ional Re iew o Economics & Finance,79, 448–465. h ps://doi.o g/10.1016/j.i e .2022.02.056
14 N. NGUYEN THI NGOC AND T. NGUYEN NGOC
