Too much or too little? Assessing the optimal size of government in inclusive development

Uni e si y o Minho
School o Economics and Managemen
F ancisco Tomaz Cabelei a
Too much o oo li le? Assessing he op imal
size o go e nmen in inclusi e de elopmen
ma ch 2025
Uni e si y o Minho
School o Economics and Managemen
F ancisco Tomaz Cabelei a
Too much o oo li le? Assessing he op imal
size o go e nmen in inclusi e de elopmen
Mas e ’s Disse a ion in Economics
Disse a ion supe ised by
P o esso F ancisco José Al es Coelho Veiga
ma ch 2025
Copy igh and Te ms o Use o Thi d Pa y Wo k
This disse a ion epo s on academic wo k ha can be used by hi d pa ies as long as he in e na ionally
accep ed s anda ds and good p ac ices a e espec ed conce ning copy igh and ela ed igh s.
This wo k can he ea e be used unde he e ms es ablished in he license below.
Reade s needing au ho iza ion condi ions no p o ided o in he indica ed licensing should con ac he
au ho h ough he Reposi ó iUM o he Uni e si y o Minho.
License g an ed o use s o his wo k:
CC BY-NC-ND
h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
i
Acknowledgemen s
I he eby exp ess my g a i ude o all who, di ec ly o indi ec ly, con ibu ed o he elabo a ion o his
disse a ion, whe he h ough supe ision o c ucial insigh s.
Special ecogni ion goes o my supe iso , P o esso F ancisco Veiga, o his in aluable guidance
and ime, which we e indispensable o he ealiza ion and subsequen imp o emen o his wo k. His
men o ship was no only commendable and suppo i e bu also excelled in bo h academic expe ise and
app oachabili y.
I also acknowledge he ema kable impo ance o ex e nal judgmen and he exchange o in o ma ion
and opinions wi h o he s. The e o e, I would like o ex end my g a i ude o my amily, as well as o my
iends, pa icula ly my colleagues Ped o Machado and F ancisco Ribei o, who, as ellow wo k and aca-
demic pee s, played a c ucial ole in discussing opics o bo h academic and non-academic na u e. Thei
conce n and suppo a e also g ea ly che ished.
ii

S a emen o In eg i y
I he eby decla e ha ing conduc ed his academic wo k wi h in eg i y.
I con i m ha I ha e no used plagia ism o any o m o undue use o in o ma ion o alsi ica ion o esul s
along he p ocess leading o i s elabo a ion.
I u he decla e ha I ha e ully acknowledged he Code o E hical Conduc o he Uni e si y o Minho.
Uni e si y o Minho, B aga, ma ch 2025
F ancisco Tomaz Cabelei a
iii
Abs ac
This a icle aims o con ibu e o he g owing body o esea ch on he op imal size o go e nmen
when examining non- adi ional economic de elopmen me ics, speci ically by being he i s o depic
e idence o he A mey Cu e o he Inequali y-adjus ed Human De elopmen Index (IHDI). Fu he mo e,
he e idence, o he a ia ion o op imal go e nmen size acco ding o me ics o de elopmen , is sca ce,
and as such his s udy a emp s o es an unde mining o he ole o go e nmen p esen in he li e a u e,
a di ec consequence o w ongly p oxying wel a e by no ma i e economic indica o s.
The s udy also a emp s o a oid he inco ec ly igno ed e ec s o go e nance by composing a a iable
o ins i u ional quali y h ough he use o wo ld go e nance indica o s. The esul s p o ide e idence o an
op imal size o go e nmen anging om 27.22% o 32.54% o G oss Domes ic P oduc (GDP), 33.34% o
42.20% o Human De elopmen Index (HDI), and las ly, 31.23% o 40.69% o IHDI. The esul s imply ha
he dependen a iable used makes a conside able di e ence in he op imal size o go e nmen compu ed,
none heless, i exhibi s ha while wo ldwide go e nmen size is no o e g own, he b oad sen imen o
o e bea ing go e nmen p esence is ac ual, a leas o he Wes e n Wo ld.
The esul s also imply a s onge connec ion be ween go e nmen size and economic p ospe i y
o de eloping coun ies, which con adic s he cu en li e a u e. Addi ionally, indings sugges ha
i is possible ha he mo e disad an aged a coun y is in e ms o de elopmen , he mo e easily
i can de i e u ili y om de elopmen policies ela ed o in e en ions in go e nmen size. A s udy
o expendi u e componen s was also conduc ed, in which public consump ion was in e ed o e-
po highe e idence o he A mey Cu e han public in es men . Resul s also exhibi ed ha o he
Classi ica ion o he Func ion o Go e nmen (COFOG) componen s, expendi u e in public o de and sa e y
eme ged as he mos undamen al elemen in he p omo ion o economic g ow h (GDP), de elopmen
(HDI), and e en inclusi e de elopmen (IHDI).
Keywo ds Op imal size o go e nmen , Inclusi e economic de elopmen , Economic de elopmen , Eco-
nomic g ow h, Composi ion o public expendi u e
i
Resumo
Es e a igo p e ende con ibui pa a o c escen e co po de in es igação sob e a dimensão ó ima do
es ado quando se examinam mé icas de desen ol imen o económico não adicionais, especi icamen e
po se o p imei o a ap esen a p o as da Cu a de A mey pa a o IHDI. Além disso, a e idência, ou a
a iação da dimensão ó ima do es ado de aco do com as mé icas de desen ol imen o, são escassas
e, como al, es e a igo en a es a a sub alo ização do papel do es ado p esen e na li e a u a, uma
consequência di e a de uma ep esen ação alaciosa do bem-es a a a és de indicado es económicos
no ma i os.
O es udo ambém en a e i a os e ei os e adamen e igno ados da go e nança, compondo uma a -
iá el de qualidade ins i ucional a a és da u ilização dos indicado es de go e nança global. Os esul ados
o necem p o as de uma dimensão ó ima do es ado que a ia en e 27,22% e 32,54% pa a o GDP, 33,34%
e 42,20% pa a o HDI e, po úl imo, 31,23% e 40,69% pa a o IHDI. Os esul ados implicam que a a iá el
dependen e u ilizada az uma di e ença conside á el na dimensão ó ima do es ado calculada, no en an o,
mos a que, embo a a dimensão do es ado a ní el mundial não seja excessi a, o sen imen o ge al de uma
p esença exage ada do es ado é ac ual, pelo menos pa a o Mundo Ociden al.
Os esul ados ambém suge em uma ligação mais o e en e o peso do es ado e a p ospe idade
económica nos países em desen ol imen o, o que con a ia a li e a u a a ual. Além disso, os esul ados
suge em que é possí el que quan o mais des a o ecido o um país em e mos de desen ol imen o,
maio acilidade e á em ex ai u ilidade de polí icas de desen ol imen o elacionadas com al e ações no
amanho do es ado. Foi ambém e e uado um es udo das componen es da despesa, em que se in e iu
que o consumo público ap esen a maio e idência da Cu a de A mey do que o in es imen o público. Os
esul ados ambém mos am que, pa a os componen es COFOG, as despesas com o dem e segu ança
pública eme gem como o elemen o mais undamen al na p omoção do c escimen o económico (GDP),
do desen ol imen o (HDI) e mesmo do desen ol imen o inclusi o (IHDI).
Pala as-cha e Dimensão óp ima do es ado, Desen ol imen o económico inclusi o, Desen ol imen o
económico, C escimen o económico, Composição da despesa pública
Con en s
1 In oduc ion 1
2 Li e a u e e iew 4
2.1 The measu emen and his o ical dynamics o go e nmen size ............. 4
2.1.1 How should we measu e he size o go e nmen ? ............... 4
2.1.2 The his o y o go e nmen in e en ion .................... 5
2.2 The ole o go e nmen s: A e iew o he li e a u e ................... 8
2.2.1 The economic and social ole o go e nmen ................. 8
2.2.2 De elopmen and economic g ow h ...................... 10
2.2.3 P io schola ship ............................... 13
2.3 The in luence o go e nance and he impo ance o expendi u e alloca ion ....... 16
2.3.1 Go e nmen quali y .............................. 16
2.3.2 Expendi u e componen s. Do hey all ma e ? ................. 18
3 Me hodology and da a 20
3.1 Da a sou ce and desc ip i e s a is ics ......................... 20
3.1.1 Da ase cons uc ion ............................. 20
3.1.2 Desc ip i e s a is ics and p elimina y esul s ................. 25
3.2 Econome ic me hodology ............................... 32
3.2.1 Econome ic app oach o economic g ow h indica o s ............. 32
3.2.2 Econome ic app oach o economic de elopmen indica o s .......... 36
4 Resul s 39
4.1 Main analysis ..................................... 39
4.1.1 Benchma k esul s .............................. 39
4.1.2 Componen s o go e nance .......................... 50
i
VIF Va iance In la ion Fac o . 23
WB Wo ld Bank. 16,21,22,24,28,30,31,88,89,90,91
xiii

1 In oduc ion
Many may ask why he deba e abou go e nmen in e en ion is s ill going, e en a e ex ensi e e-
sea ch ega ding i s op imal size. The eason pa ly elies on he subs an ial di e gence in he me hod-
ologies and me ics used (Be gh and Hen ekson,2011)1. Fo a ious decades, mos academics ha e
ocused on employing he GDP o ep esen he p ospe i y o na ions wo ldwide, which has p o en use ul
in un eiling c ucial insigh s abou he impo ance o a cen al go e nmen in es ablishing a solid ounda ion
o economic p ospe i y (Mahle 1992;Be y and Lowe y 1984).
The size o go e nmen mac oeconomically suppo s i sel h ough he implica ions ha a ious models
sugges . Many de elopmen and g ow h models, such as he Solow-Swan (Solow,1956), es ablish ha
gi en diminishing e u ns o p oduc ion ac o s, coun ies would ha e a ha de ime con inuing o g ow
a e some benchma k. Go e nmen s would need some le el o p esence in economic ac i i y o escape
his cu se o dec easing ma ginal e u ns o g ow h. Acco dingly, o he models, such as Rome (1986),
depic ha a he agg ega e le el, diminishing ma ginal e u ns can be a oided, allowing g ow h o be
sel -sus aining h ough he p oli e a ion o posi i e ex e nali ies ha he exchange o know-how (“lea ning-
by-doing”) may pe mi .
In his con ex , he go e nmen appea s essen ial, gi en ha i ep esen s he en i y esponsible o
p o iding he se ices, such as basic in as uc u e and educa ion, ha allow o he ans e and p oli -
e a ion o ha know-how. None heless, o he models, such as Ba o (1990) and Ba o and Sala-i-Ma in
(1995), e e o he ac ha he e a e op imal axa ion le els, and by associa ion, op imal go e nmen
p esence, in which he posi i e ex e nali ies a e augmen ed ela i e o he economic dis o ions ha he
equi ed axes impose. This ela ionship is simila o ha o he A mey Cu e, which depic s ha he e
is a poin a which he size o go e nmen is op imal o p omo e economic g ow h, gi en ha a e a ce -
ain poin , augmen ed go e nmen al p esence would only ha e de imen al e ec s on g ow h (Vedde and
Gallaway,1998).
This esea ch aims o help sol e ha dilemma, no in he con ex o op imal axa ion, bu o a p oxy,
1The au ho also depic s ha li e a u e is ull o con adic o y indings ha a e no explained jus because o he a ia ion in he de ini ions bu also due o he
he e ogenei y in he coun ies s udied.
1
which is he size o go e nmen . Mos li e a u e on his opic poin s o G oss Domes ic P oduc (GDP) as
he de e minan ac o be maximised, bu gi en he amoun o c i icism ha i ecei ed om many Nobel
lau ea es, such as Kuzne s (1941), Hicks (1948), and No dhaus and Tobin (1972)2, as a lawed measu e
o s anda d o li ing, I a gue ha a mo e use ul goal would be o a emp o assess he go e nmen size
ha maximises a mo e inclusi e measu e o economic de elopmen .
A me ic ha add esses aspec s beyond he mone a y would p esen i sel as mo e adequa e o o e -
come he possible bias ha may a ise when assessing he op imal size o go e nmen wi h no ma i e
me ics such as he GDP, since i g ea ly o e looks noncon en ional in e en ions ha he go e nmen
enac s, namely in ea ing social issues and dealing wi h inequali y ( an den Be gh,2009).
The Human De elopmen Index (HDI) could be an al e na i e o sol e his issue, se ing as an im-
p o emen om he p e ious me hodologies, wi h some wo ks such as Ma ins and Veiga (2014) and
Da ies (2009) al eady using i as a benchma k o mo e adequa ely imp o e he compu a ions o he op-
imal size o go e nmen . None heless, as sugges ed by an den Be gh (2009) and Saga and Najam
(1998), gi en a ious measu abili y p oblems, i s ill ails o include some ele an dynamics, wi h g ea e
emphasis on inequali y, a g owing conce n in mode n imes.
Du ing a ious yea s he issue emained un esol ed, howe e , Alki e and Fos e (2010) add essed
his sho coming by a emp ing o ein o ce he me ic o HDI by discoun ing inequali y in all i s sub-
componen s (heal h, educa ion, and s anda d o li ing). The esul ing new me ic is designa ed as he
Inequali y-adjus ed Human De elopmen Index (IHDI). This is he main me ic his s udy a emp s o use
o enhance he lack o e idence on op imal go e nmen size o measu es o economic de elopmen .
Addi ionally, gi en he b oad e idence ega ding he ele ance o he quali y o go e nmen in p omo ing
economic g ow h and de elopmen , such as Acemoglu e al. (2001) and Ni ola and Sahu (2019), his
disse a ion also s i es o disen angle he possible mixed-up e ec s ha p e ious academia may ha e
o e looked h ough he inco po a ion and analysis o go e nance.
A e his in oduc ion, I p esen Pa 2, which e lec s he li e a u e e iew, whe e he academic oun-
da ions ela ed o he measu emen o go e nmen size a e depic ed, alongside any heo y necessa y o a
be e unde s anding o he ela ionship be ween go e nmen size and economic p ospe i y, coupled wi h
p e ious empi ical e idence exhibi ed by o he au ho s.
Then, Pa 3, he me hodology and da a, displays how he da abases we e cons uc ed alongside
hei da a sou ces. Addi ionally, his sec ion also desc ibes any da abase dynamics while ying o assess
p elimina y esul s. I also epo s he econome ic models used o ex apola e he esul s, coupled wi h
2The economic s udy has been blindly obedien o agg ega e ma e ial p og ess, neglec ing he dis o ion o na ional p io i ies, he wo sening o he income
dis ibu ion, and he en i onmen al damages ha i poses.
2
he a ionale o he a iables used.
Following, Pa 4, he esul s, a e p esen ed, whe e any ele an inding is epo ed, while also ying o
compa e how i ela es o he hypo hesis made and o he p e ious e idence. He e a obus ness analysis is
also ca ied ou o gua an ee he s eng h o he conclusions, while also using he esul s o in e adequa e
policymaking decisions and ad oca ing o u he academic esea ch on app op ia e opics. Finally, Pa
5, he conclusion, w aps up he signi ican and no el in e ences, depic ing how hey se e o enhance he
p esen schola ship.
3
2 Li e a u e e iew
2.1 The measu emen and his o ical dynamics o go e nmen
size
2.1.1 How should we measu e he size o go e nmen ?
Economic unde pe o mance and poli ical ins abili y, namely in he
”old wo ld”
, ha e c ea ed an eme -
gence in he discussion o go e nmen al p esence in he economy, wi h di e en elec o a es ad oca ing
o dis inc go e nmen app oaches as well as pushing na a i es ha a e be e aligned wi h he in e es s
o hei class-de ined co e poli ical cons i uencies (Hibbs,1977). Dis ega ding ideological biases, hose
poli ical na a i es s ill p esen some logical easoning o hei s a emen s, gi en ha he go e nmen
no only has he pu pose o sus aining social cohesion bu also, a he same ime, gua an eeing s able
economic g ow h in he long e m (A anson and O deshook 1981;Tanzi and Schuknech 1997).1
Consequen ly, I a gue ha he go e nmen ’s eal objec i e should be de e mining he op imal equi-
lib ium o go e nmen al p esence ha maximises bo h social and economic condi ions. None heless,
al hough he b oad concep o go e nmen size is common, he e is no s anda d amewo k o compu e i
since he e a e nume ous indica o s and me hods o do so (Di Ma eo 2013;Vedde and Gallaway 1998).
Mo eo e , he eason why schola ship ela ed o go e nmen size is qui e he e ogeneous, equen ly p oduc-
ing con lic ing esul s, de i es subs an ially om he dis inc measu emen s and de ini ions o go e nmen
size employed by he di e en au ho s (Be gh and Hen ekson,2011).
The e is no de ini i e answe o which measu emen o go e nmen size is he mos accu a e one, wi h
many au ho s using, o ins ance, he o al numbe o go e nmen employees, o he nominal expendi u es,
and e en e enues as a pe cen age o GDP (Labon e,2010). None heless, Labon e (2010) s a es ha
he e a e wo main ways by which we can measu e go e nmen ha a e commonly accep ed o be mo e
1A anson and O deshook (1981) s a e ha ew can deny he bene icial e ec s o go e nmen in e en ion in domains such as na ional de ence, police p o ec ion,
and common-law cou s.
4
p ominen . Mo e speci ically, measu ing go e nmen size acco ding o expendi u es (ou lays) o e enues
( eceip s). Bo h ways a e me hodologically alid, al hough some au ho s sugges ha he di e ences
exhibi ed by he wo app oaches can s ill be subs an ial, and hus i is c ucial o only op o one (Labon e,
2010).
We can exhibi wo main mo i es as o why exp essing go e nmen size as expendi u es is mo e
app op ia e han by e enues. The i s ela es o he highe ola ili y o e enue le els, which can only be
indi ec ly in luenced by legisla o s. Re enues depend a lo on economic condi ions and a e e y p one o
changes in he sho e m due o unexplained o uncon ollable ac o s. Consequen ly, using i as a p oxy
o he size o go e nmen would, by associa ion, e lec a alse pe cep ion o a he e ogeneous and ola ile
na u e o go e nmen size (Labon e,2010).
The second eason conce ns he possible di e gence ha may a ise in e ms o go e nmen size
when measu ing i by e enues compa ed o ou lays, pa icula ly in he con ex o budge de ici s. In his
amewo k, gi en ha budge de ici s canno las o e e and need o each a neu al s a e soone o la e , a
educ ion in axes ha is no acco dingly ma ched by a cu in spending does no lead o any educ ion o he
size o go e nmen . Consequen ly, wha akes place is a empo al manipula ion o he ax incidence, and
he measu emen by e enues ends up p o iding a misleading e e ence ha he size o he go e nmen
is smalle han i is (Labon e,2010).
Acco dingly, his wo k makes use o go e nmen spending as i s p oxy o go e nmen size, gi en
ha i exhibi s an appa en edge when compa ed wi h o he me ics, especially when i is measu ed as a
pe cen age o GDP, as depic ed by Vedde and Gallaway (1998) and Labon e (2010)2.
2.1.2 The his o y o go e nmen in e en ion
The size o go e nmen is a he ola ile, gi en ha i changes acco ding o socie y’s needs in each pe-
iod. Be o e Wo ld Wa 1, mos economis s ad oca ed o a minimal go e nmen
(”Laissez- ai e”)
in which
i s only unc ions would be he main enance o law en o cemen , adminis a ion, and na ional de ence,
gi en he assumed abili y o he ma ke o sel - egula e.3As depic ed in Figu e 1, a ha ime, he a e age
sha e o public expendi u e wo ldwide was no mo e han 10% o GDP. None heless, he impac o he 1s
Wo ld Wa , in 1914, p oposed inc emen al go e nmen in e en ion bo h du ing, due o wa e o s, bu
also a e he con lic , gi en he equi emen o public measu es and policies o sol e po e y, ins abili y,
2Labon e (2010) also exhibi ed ha go e nmen expendi u es ha e dis inc impac s on he economy, mo e speci ically, in he case o go e nmen ans e s and
go e nmen pu chases o goods and se ices.
3The ideology o sel - egula ing ma ke s, o , in o he wo ds, he ma ke economy, was an economic hough ha o igina ed in England, which depic ed ha
p ices mus be allowed o egula e hemsel es in a sys em o minimal ou side in e e ence.
5

and unemploymen (Muelle ,2003).
Consequen ly, a e he con lic , e en hough downwa d p essu es o go e nmen al p esence we e
epo ed, go e nmen expendi u es did no all o p e-wa le els and consis en ly con inued o g ow. The
G ea Dep ession, om 1929 un il 1939, also p esen ed an impo an ma k in he ele ance o go e nmen
p esence in he economy, gi en he equi ed public incen i es ha he c isis made impe a i e (Be nanke,
2000).4
0%
5%
10%
15%
20%
25%
30%
35%
40%
1890
1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
2010
2020
Figu e 1: Wo ld a e age go e nmen size (% o GDP).
Sou ce:
In e na ional Mone a y Fund (IMF)/Own Compu a ions.
The 2nd Wo ld Wa p oposed a simila phenomenon o ha o he i s one, in which wa e o s
inc eased he size o go e nmen , and by i s end in 1945, he global a e age weigh o go e nmen had
al eady su passed 20% o GDP, which, al hough su e ing a sho - e m down u n in he ollowing yea s,
kep expanding un il he nine ies. One ounda ional ac o o ha g ow h appea s in he o m o he pa ial
eme gence o social insu ance p og ams, which D yzek and Goodin (1986) ound o be linked o he e en s
o he 2nd Wo ld Wa , a leas o G ea B i ain. Ano he eason ha pa ially explains ha upsu ge de i es
om he inc emen al dis us o he idea o a sel - egula ing ma ke p oposed by Adam Smi h (Smi h,
1776), which was losing ele ance o he ideas o o he es eemed schola s, mo e pa icula ly, o hose o
John Mayna d Keynes (Muelle 2003;Be nanke 2000).
As Keynes commonly used o s a e, gi en he un easonable long pe iod ma ke s equi e o sel - egula e,
go e nmen al in e en ion is essen ial o sol e issues such as low g ow h and high unemploymen , gi en
ha
”in he long un, we a e all dead”
(Keynes,1923).5Fu he mo e, gi en he g owing ubiqui y o social
4The augmen ed go e nmen in e en ion du ing his ime was an e iden signal ha he public was losing con idence in he sel -co ec ing mechanisms o a
ma ke economy.
5Keynes sugges ed ha bo h wages and p ices we e
”s icky”
, in he sense ha hey we e no capable o e icien ly esponding o ma ke demand, a leas in a
6
secu i y sys ems and non-mone a y p og ams (Muelle ,2003), he a e age size o public expendi u e
peaked in 1992, when he wo ld a e age go e nmen expendi u e was mo e han 35% o GDP.
A he ime, he e we e s ong belie s ha mos de eloped na ions, pa icula ly in Eu ope and No h
Ame ica, we e in unsus ainable ax b acke s and had go e nmen p esence ha endange ed economic
lou ishing. Addi ionally, schola ship c i icising he ad e se e ec s o o e sized go e nmen s became com-
mon, wi h many academics sugges ing ha he nega i e ex e nali ies exhibi ed by he public sec o could
be as bad as hose o he p i a e (Gwa ney e al. 1998;Muelle 2003).
Some o he c i iques de i e om he economic dis o ions ha he associa ed income ax bu dens
o igina e, namely in he con ex o loss o e iciency and he blu ing o p ice signals (Muelle ,2003). This
e ec al e s he choice be ween wo k and leisu e, and as exhibi ed by Alesina and Pe o i (1997), he
inc emen o ax b acke s in labou can cause an a i icial inc ease in labou cos s, which, by associa ion,
causes highe unemploymen and an o e all loss o compe i i eness.
O e axing can also ha e co up i e e ec s on he economy by p omo ing economic ac i i y in he
shadow economy and unde epo ing o e enues. Ins ead o being in luenced o wo k less, axes can
induce indi iduals o simply epo ha hey gain less by a emp ing o ecei e hei income h ough he
unde g ound economy (Muelle ,2003). Fo ins ance, And eoni e al. (1998) sugges ha his p oblem is
subs an ial, a leas in he Uni ed S a es, whe e oughly 17% o all po en ial ax e enue om income ax
( ede al) is los due o ax e asion.
A la ge go e nmen al p esence can also ins iga e co up ion, gi en ha he p o ision o public goods
and se ices can ha e bo h di ec and indi ec bene i s o indi idual en i ies. In his sense, co up ion
eme ges due o he emp a ion o he p i a e sec o o in e e e in public decision-making by b ibing bu-
eauc a s, which will sac i ice he public in e es s o hose o hei own and hei co up o s (Muelle ,
2003).
None heless, co up ion can be subs an ially deal wi h by endowing highe compensa ions, as p o-
posed by Goel and Nelson (1998), who ound co up ion o be in e sely ela ed o he wages paid o he
employees o he public sec o . Howe e , he a ilia ed cos s o inc easing wages o elimina e all co up ion
a e oo high and no jus i iable o he gains ha esul om educed co up ion, and in his sense, he e
will always be a ade-o be ween bu eauc a s’ wages and co up ion s anda ds.
Subsequen ly, all hese c i iques led mos na ions o unde ake sligh dec eases in he sizes o go -
e nmen in he la e nine ies, d opping o alues simila o hose o he ea ly eigh ies. None heless, he
eme gence o he global inancial c isis made i un eliable o go e nmen s o become less p esen in he
easonable amoun o ime.
7
economy (Lae en e al.,2010), e en when knowing o he dange s o o e -in e en ionism. As o he sec-
ond decade o he 21s cen u y, al hough he idea o educing go e nmen s ill p e ails, mos na ions ha e
been incapable o subs an ially doing so, e en mo e so wi h he mani es a ion o he COVID-19 pandemic
(Wo ld Bank,2022).6
2.2 The ole o go e nmen s: A e iew o he li e a u e
2.2.1 The economic and social ole o go e nmen
E en hough he ole o go e nmen in dealing wi h shocks, such as c isis and ins abili y, was p e iously
exhibi ed, i is impo an o ocus on i s ele ance om a mac oeconomic s andpoin . To ully explain he
mac oeconomic ole o go e nmen pe se, we need o i s unde s and he de e minan s o g ow h (GDP).
The Solow (1956) model is one o he mos undamen al amewo ks in his con ex , and i desc ibes ha
GDP le els a e de e mined by h ee main ac o s: capi al s ock (equipmen and buildings a ailable o
p oduc ion), labou (hou s wo ked), and echnology.
The sha e o GDP g ow h ha is no explained by labou o capi al is wha we e e o as he Solow
Residual o To al Fac o P oduc i i y (TFP), which ela es o he e ec o echnology.7One assump ion ha
his model p oposes is economic con e gence, gi en ha hose p oduc ion ac o s, pa icula ly capi al and
labou , p esen diminishing e u ns, hus, wi hou any exogenous change in TFP, all coun ies condi ionally
con e ge ( hey con e ge o he same poin i hey possess he same popula ional g ow h a e, sa ings a e,
and capi al dep ecia ion a e) o he same GDP pe capi a in he long un.
The p oposi ion o con e gence, pa icula ly o absolu e con e gence, is no homogeneous as o
he la e 20 h and ea ly 21s cen u ies, gi en ha he e idence wo ldwide is no b oadly suppo i e and
commonly p esen s opposi e implica ions. The schola ship is subs an ially di ided, wi h many au ho s
sugges ing ha con e gence holds, e en absolu ely (Naha and Inde ,2002), and o he s sugges ing he
opposi e ela ionship (Quah,1993,1997), and he e a e e en some cases whe e absolu e con e gence is
dependen on he ime ame and se o coun ies ha a e used in he analysis (Ma hu ,2005).
The na u e o absolu e con e gence is undoub edly ambiguous, none heless, ha ela ion does no
seem o anspose o condi ional con e gence. Al hough some au ho s, such as Nell (2020), exposed no
condi ional con e gence o he pos -1989 globalisa ion e a, and Bold in and Cano a (2001) and Cano a
6Du ing he pandemic pe iod, he e was unp eceden ed go e nmen in e en ion and egula o y o bea ance, wi h g ea emphasis on he inancial sec o , whe e
hose policies aided banks in main aining capi al le els and liquidi y.
7Inc eases in TFP a e commonly a ibu ed in hei en i e y o echnological p og ess, none heless, i can also be in luenced by educa ion and go e nmen
egula ion.
8
and Ma ce (1995) p o ide e idence sugges ing ha i ei he is non-exis en o simply no subs an ial
enough o s ongly a i m i s widesp ead p esence, hese iews a e ela i ely isola ed. Mos o he g ow h
li e a u e, in con as , inds empi ical e idence o condi ional income con e gence.
The wide bulk o empi ical e idence sugges s ha economies ha a e placed a lowe ini ial income
le els end o g ow a as e a es han hose es ablished wi h highe ini ial incomes, a e con olling o
sa ing and popula ion g ow h a es, co obo a ing he commonness o condi ional con e gence (Ba o
e al. 1991,1992;Cho and G aham 1996;Kai ila 2004;Nonneman and Vanhoud 1996;Mu hy and
Chien 1997).
Conside ing his con lic ing eali y, one migh ask onesel which o he ac o s can help explain g ow h,
as Rome (1986) lea ning-by-doing model exhibi s. His wo k depic s ha e en hough a he indi idual le el
diminishing e u ns s ill hold, a he agg ega e le el, ha issue can be deal wi h h ough he p oli e a ion
o posi i e ex e nali ies. Those a ise due o he exchange o know-how be ween en e p ises, which makes
i possible o o e come he p oblem o diminishing e u ns, enabling g ow h o be sel -sus aining. This
dynamic p oposes he ele ance o go e nmen al economic p esence since he public sec o se es as he
en i y esponsible o p o iding basic esea ch, in as uc u e, and educa ion ha allows o he p oli e a ion
o ha know-how and associa ed posi i e ex e nali ies8.
Despi e ha , hose se ices mus be unded by axes, which on hei own can also c ea e economic
dis o ions. Following his a ionale, Ba o (1990) and Ba o and Sala-i-Ma in (1995) p opose ha he e is
an op imal axa ion le el o in es ing in hose se ices, gi en ha he e is a poin a which he ne bene i
in e ms o economic g ow h ha de i es om p o iding hose public se ices is maximised.9By p oxy,
he exis ence o op imal axa ion le els o maximise g ow h implies ha go e nmen al p esence is highly
bene icial o augmen ing economic pe o mance, none heless, only un il a h eshold.
The a ionale o go e nmen al in e e ence is also sha ed wi h Vedde and Gallaway (1998), who
s a e ha he e ec o go e nmen size on economic g ow h beha es as he A mey Cu e p oposes, in a
quad a ic o m (in e ed U-shape), whe e oo li le go e nmen p e en s he c ea ion o an op imal layou
o g ow h, while oo much go e nmen punishes he p i a e sec o . “Why?” is wha one migh ask.
Vedde and Gallaway (1998) s a e ha i pa ially ela es o he agili y ha a lacking go e nmen would
c ea e, namely in lack o p ope y igh s ( ule o law), which would pu o wa d a disincen i e o bo h sa e
and in es , gi en ha exp op ia ion would be eal and common. The A mey Cu e, depic ed in Figu e 2,
8In he case o ne wo k ex e nali ies, he e is a lack o p i a e incen i e o de elop hem, and in his sense, public policy plays a c ucial ole in bo h he
implemen a ion and dis ibu ion o he associa ed in as uc u e and
”know-how”
.
9All and any public expendi u e mus be inanced by ax in p esen o u u e p i a e weal h, which educes he ne e u n on in es men while also slowing down
p i a e capi al accumula ion.
9
go e nmen size in inclusi e de elopmen (IHDI) as a p ima y esul while also s i ing o de e mine i he
esul s showcased by he al e na i e me ics could p opose a possible unde - ep esen a ion o he ole o
go e nmen in he p esen li e a u e.
2.3 The in luence o go e nance and he impo ance o expen-
di u e alloca ion
2.3.1 Go e nmen quali y
As p e iously es ablished, go e nmen al in ol emen is ele an o suppo and acili a e bo h eco-
nomic g ow h and de elopmen . E en so, he exis ence o ex emely success ul na ions ha ha e no only
commonly epo ed highe g ow h a es han he wo ld a e age bu also depic ed much highe GDP pe
capi a le els, such as I eland, Sou h Ko ea, Singapo e, Qa a , and e en Swi ze land, ha ha e ela i ely
low public sec o s (21.23%, 28.68%, 15.39%, 24.25%, and 31.53% as o 2022), migh d i e us o sc u inise
wha o he ac o s may be a play.
One con ibu ing ac o is he quali y o go e nmen , which we can de ine as go e nance, a concep
ha depic s how well a na ion manages i s esou ces and how anspa en and accoun able i s ins i u ions
a e (Baland e al.,2010). Go e nance is o di e impo ance o p omo e economic p ospe i y, wi h Wol owi z
(2006), o me P esiden o he Wo ld Bank (WB), e en sugges ing ha dys unc ional go e nance is he
la ges h ea o de elopmen in mos o he unde de eloped wo ld.
The ele ance o go e nance is no necessa ily ecen , wi h Acemoglu e al. (2001)19 and No h and
Thomas (1973) depic ing ha he e is a causal e ec o economic ins i u ions and go e nance on de elop-
men , wi h he he e ogenei y in hei quali y being a majo explana ion o he con empo a y disc epancy
be ween ich and poo coun ies. Recen ly, howe e , economis s s a ed o join ly ackle he dynamics
and e ec s o size and quali y o go e nmen on g ow h, wi h Ni ola and Sahu (2019) sugges ing ha o
enhance bo h s a e-le el and na ional-le el economic g ow h, na ions should accen ua e hei e o s on
p omo ing he quali y o ins i u ions, al e na i ely o blindly employing oo much a en ion o he size o
go e nmen s.
The e is also empi ical e idence o a seemingly in e ac i e ela ionship be ween go e nmen size and
quali y, gi en ha in some cases, inc emen ing bo h can ha e posi i e e ec s on g ow h ou comes, despi e
ha , coun ies wi h highe go e nance make mo e e ec i e use o inc emen al go e nmen size han hose
19 Acemoglu e al. (2001) showcase he example o B i ish colonies compa ed o hose o he F ench, Spanish, and Po uguese, which, due o good economic
and poli ical ins i u ions ha hey inhe i ed om B i ain, exhibi much supe io p ospe i y han he la e .
16

wi h lowe go e nance (Coo ay,2009). A onso and Jalles (2016) also po ay ha he s udy o he op imal
le el o go e nmen alls unde he in e ac ion o size and go e nance, depic ing ha ins i u ional quali y
has a posi i e and signi ican impac on he le el o eal GDP pe capi a and ha he supposedly nega i e
impac o go e nmen size on g ow h is mo e signi ican o na ions in a se ing o low quali y o go e nance.
Kim e al. (2018) also exhibi simila ela ionships, s a ing ha bo h he size and quali y o go e nmen s
se e as undamen al ools o p omo e p oduc i i y and, hence, economic g ow h. Mo e p ecisely, hei
esul s sugges ha be e go e nance acili a es coun ies in bene i ing om enla ged go e nmen size,
gi en ha hey ac as complemen s in p omo ing e iciency and ou pu g ow h20.
In his con ex , my las esea ch opic a emp s o disen angle he possible mixed-up e ec s o go -
e nmen size and quali y on g ow h and de elopmen , o sugges policy ecommenda ions in e ms o
go e nmen al esou ce alloca ion, and o gua an ee ha any bias is a oided in he compu a ion o he
op imal size o go e nmen . Following his p emise, coupled wi h Goh and Mohd Aznan (2023) ecom-
menda ions21, his s udy inco po a es go e nance indica o s as a iables o con ol.
The e is a possibili y ha p e ious li e a u e may ha e w ongly es ima ed he size o he go e nmen ,
po en ially due o he co ela ion o he go e nmen size wi h an unobse ed a iable, go e nance. This
p oblem is wha we e e o as endogenei y, and he omission o go e nmen quali y could pa ially explain
why go e nmen size ends o be so highly ele an in mos empi ical s udies, e en hough, punc ually,
some empi ical wo ks s ill ind e idence o inexis en e ec s o go e nmen size on bo h g ow h and de el-
opmen .
The g ea e bulk o academic s udies imply he ounda ional ole ha good go e nance has in endo sing
g ow h, none heless, he e a e s ill au ho s ha con adic he leading wa e by sugges ing o he dynamics.
Fo ins ance, Sachs e al. (2004) exhibi ha poo coun ies a e no poo due o bad go e nance, bu
a he because hey lack he economic esou ces o a o d good go e nance in he i s place, and ha
su p isingly, many unde de eloped na ions epo ela i ely a ou able le els o go e nance. Fu he mo e,
Hausmann e al. (2008) a gue ha go e nance is indeed a ele an issue, ne e heless, i ep esen s only
one o many o he ac o s ha may sabo age de elopmen .
20 Kim e al. (2018), by di iding hei sample in o high- and low- esou ce coun ies, iden i y ha his ela ionship is ela i ely mo e p onounced o coun ies wi h
an abundance o na u al esou ces.
21 Goh and Mohd Aznan (2023) s ess he pe inence o inco po a ing go e nance in he me hodology while also exhibi ing ha di e en egions may un eil
dis inc esul s.
17
2.3.2 Expendi u e componen s. Do hey all ma e ?
When assessing he op imal size o go e nmen , a subs an ial numbe o academics ely on he as-
sump ion ha all go e nmen expendi u es beha e simila ly, igno ing any possible he e ogeneous e ec s
ha dis inc componen s o expendi u e could ha e on g ow h (Vedde and Gallaway 1998;Asimakopoulos
and Ka a ias 2016;Al unc and Aydın 2013;Makin e al. 2019;Aly and S azicich 2000). None heless,
one migh ques ion he alidi y o ha assump ion, gi en ha , o ins ance, in es men and consump ion
expendi u es could lead o dis inc esul s, bo h in absolu e e ms and in e ms o empo al incidence.
O he schola s ha e a emp ed o add ess his dynamic, namely by decomposing go e nmen size in o
adequa e componen s and compu ing he op imal size o go e nmen acco dingly. Hajamini and Falahi
(2018)22 o ins ance, use dis inc me ics and componen s o make hei compu a ions, mo e pa icula ly
by decomposing go e nmen size in o inal consump ion expendi u e and cu en expendi u e o he han
inal consump ion.
Simila ly, Chen and Lee (2005) also decompose he o e all go e nmen size, exp essed as go e n-
men expendi u e (as a % o GDP), in o wo ca ego ies, speci ically go e nmen consump ion expendi u e
and go e nmen in es men expendi u e. His idea de i ed om Lin (1994) hypo hesis ha while bo h
componen s could ha e bene icial epe cussions on g ow h, he impac o go e nmen consump ion will
be less signi ican han he con ibu ion o go e nmen in es men , gi en ha he la e has an upli ing
e ec on p i a e in es men . Likewise, Ba o (1991) ound ha public in es men ends o be posi i ely co -
ela ed wi h bo h g ow h and p i a e in es men , none heless, con a y o he p e ious case, he exhibi ed
ha public consump ion spending has ad e se e ec s on g ow h and p i a e in es men .
Da ies (2009), in his a emp o use HDI o compu e he op imal size o go e nmen , ollows a simila
me hodology, namely by assessing he op imal size o go e nmen sepa a ely o go e nmen consump ion
and go e nmen in es men expendi u es and epo ing he op imal size o go e nmen as he summa-
ion o he p eceden op imal sizes. Acco dingly, his esul s add ess no only he dis inc e ec s o he
expendi u es’ componen s bu also hei op imal alloca ion. Simila ly, A onso and Fu ce i (2010) also de-
compose go e nmen sizes acco ding o expendi u es, howe e , con a y o Da ies (2009), hey ind ha
go e nmen in es men has s a is ically signi ican nega i e e ec s on g ow h.
Ma ins and Veiga (2014) decompose he size o go e nmen o an e en u he le el, speci ically
acco ding o de ence, heal h, educa ion, social p o ec ion, and a clus e ed g oup o he emaining ex-
pendi u e subcomponen s. This disagg ega ion made i possible o each indings sligh ly di e en om
p e ious academia, explici ly in he con ex ha al hough non-linea ela ionships we e ound o HDI g ow h
22 Hajamini and Falahi (2018), in con as o o he s udies, use go e nmen g oss ixed capi al o ma ion as hei p oxy o go e nmen size.
18
in all subcomponen s, he impac was he e ogeneous acco ding o he componen analysed. De ence, ed-
uca ion, and social p o ec ion expendi u e had ela ionships simila o ha o he A mey Cu e (in e ed
U-shape), whe e inc emen s would lead o augmen ed g ow h p ospec s un il a benchma k. Heal h and
he es o he subcomponen s display opposi e dynamics, gi en ha hey beha ed opposi ely o he A mey
Cu e, o in a U-shaped ela ionship, whe e inc eases in expendi u e always lead o de imen al e ec s,
none heless, a e a baseline, all and any inc emen al expendi u e would ha e bene icial e ec s on g ow h
ou looks.
De a ajan e al. (1996) e en sugges ha an op imal mix o public spending could lead o highe s eady
s a es o economic g ow h, gi en ha seemingly p oduc i e expendi u es could become poin less when
used in excess. Thei indings pu o wa d ha , o i s se o de eloping coun ies, inc emen s in capi al
expendi u e ha e nega i e e ec s on g ow h p ospec s, and con a y o he s anda d sen imen 23, de el-
oping na ions ha e been misalloca ing public esou ces in a ou o capi al expendi u es a he expense o
cu en expendi u es.
Ne e heless, Bose e al. (2007), in a simila analysis o de eloping coun ies, ound con lic ing esul s,
epo ing posi i e e ec s on g ow h o capi al expendi u e and insigni ican e ec s o cu en expendi u e.
Addi ionally, a a disagg ega ed le el, Bose e al. (2007) unco e ed ha he only componen s o ha e a
cons uc i e impac on g ow h a e he in es men in educa ion and o al expendi u e in educa ion24.
Dis inguishing he di e en componen s o expendi u e is highly ele an o acqui ing obus esul s,
gi en ha he absence o decomposi ion may explain why he he e ogenei y in empi ical e idence is s ill
s ong. The adi ional assump ion ha all he e ec s o go e nmen expendi u e a e he same migh be
c ea ing noise ha dis u bs he accu a e e ec o go e nmen size, obs uc ing he op imal compu a ion
o he size o go e nmen , and consequen ly, damaging he discussion o he ole o go e nmen .
23 A he ime, schola s belie ed ha o e spending in cu en expenses was damaging g ow h in de eloping coun ies, none heless, De a ajan e al. (1996)
exhibi ed ha he bias was in he opposi e di ec ion.
24 Bose e al. (2007) conclusions we e pa adoxical o hose o De a ajan e al. (1996) no only on he ele ance o cu en expendi u e bu also on he e ec s o
educa ion, which he la e depic ed o be nega i e.
19
3 Me hodology and da a
3.1 Da a sou ce and desc ip i e s a is ics
3.1.1 Da ase cons uc ion
Re e encing subsec ion 2.1.1, my p imo dial indica o e lec s he size o he go e nmen , which was
a gued o be mo e p ecisely ep esen ed by go e nmen expendi u e as a pe cen age o GDP (Vedde and
Gallaway,1998;Labon e,2010). Such an indica o was employed acco ding o da a made a ailable by
he IMF, which, al hough no he only sou ce o in o ma ion o go e nmen sizes, is he one ha depic s
mo e obus and long-las ing da a, which can go back as ea ly as wo cen u ies.
Commonly, he da a o go e nmen size ends o ange om ze o o 100 pe cen o GDP, none heless,
issues such as o eign aid can b ing o wa d anomalies in e ms o ha size, wi h punc ual cases whe e
he ange is b eached. The ew ou lie s ha all ou side ha scope we e excluded om he da ase o a oid
bias and miscompu a ions.
Addi ionally, conside ing he hypo hesis o he he e ogeneous na u e and ele ance o dis inc expendi-
u e componen s made in subsec ion 2.3.2, da a om he Go e nance Finance S a is ics o he IMF is also
inco po a ed o depic he alloca ion o expendi u e componen s, mo e pa icula ly o depic go e nmen
in e en ion acco ding o Classi ica ion o he Func ion o Go e nmen (COFOG) componen s. Con a y o
many au ho s’ p ocedu es, such as Hajamini and Falahi (2018) and Chen and Lee (2005), which only
agg ega e go e nmen expendi u e in a conside able ew clus e s, his in o ma ion allows o a mo e pa -
icula analysis o he impac o go e nmen expendi u e by decomposing i a a mo e in-dep h le el.
Mo e speci ically, such da a po ays, as a pe cen age o GDP, go e nmen expendi u e in gene al
public se ices; de ence; public o de and sa e y; economic a ai s; en i onmen al p o ec ion; housing
and communi y ameni ies; heal h; ec ea ion, cul u e, and eligion; educa ion; and las ly, social p o ec ion.
This disagg ega ion should be su icien o add ess he he e ogenei y o he componen le el and, simila ly
o Lin (1994), enable he dis inguishmen be ween p oduc i e and unp oduc i e go e nmen in e en ion.
20
The ex eme disagg ega ion, al hough bene icial in some espec s, can limi he obus ness o he
esul s, gi ing hei highe dependence on andomness, con a y o b oad go e nmen expendi u e, which
is be e a p o iding b oad s uc u al changes in he go e nmen ’s app oaches o economic in e en ion.
As such, and conside ing he di e en impac s o go e nmen consump ion and in es men depic ed by Lin
(1994), a b oade agg ega ion o go e nmen componen s is also used, mo e pa icula ly o go e nmen
expendi u e in consump ion and go e nmen expendi u e in in es men .
Gi en calcula ion misadjus men s om di e en da a sou ces, he go e nmen in es men as a pe -
cen age o GDP was e ie ed om he IMF a he simila i y o he o e all go e nmen expendi u e, and
he go e nmen consump ion was p oxied by sub ac ing he in es men om he o e all go e nmen ex-
pendi u e (in % o GDP).
The s udy o op imal alues o go e nmen size ends o add ess he e ogenei y a he le el o bo h
de elopmen and income le els, and as such, da a was e ie ed o bo h indica o s. A dummy a iable
was cons uc ed (=1 i de eloped, =0 i de eloping) ep esen ing he coun y’s de elopmen s a us in he
yea 2023 acco ding o he IMF. The ime a ia ion o de elopmen s a us was igno ed due o wo basic
easons, he i s being a lack o da a o accompany he de elopmen dynamics o e ime, and secondly,
wi h he ela i ely small ime dimension, is plausible o s a e ha he change in de elopmen s a us is oo
small o e ec i ely damage he consequen ial esul s ha de i e om his indica o .
Fo income le els, a ca ego ical a iable was cons uc ed o he clus e o income le els acco ding o
da a om he WB o 2022 (=0 i low income, =1 i lowe medium income, =2 i uppe medium income,
=3 i high income). Simila ly o he p e ious case and, o he same easons, he ime dynamics o he
a iable a e also igno ed.
To assess he impac o he size o go e nmen expendi u e on economic g ow h, in o ma ion om
he WB o GDP in cons an 2015 US$ was e ie ed. Addi ionally, om he same sou ce, he popula ion
depic ed by he o al numbe o indi iduals was also inco po a ed in o he da ase , and coupled wi h he
p e ious indica o , i was possible o cons uc an indica o o eal GDP pe capi a in cons an US$.
This indica o is la e inco po a ed as one o he a iables ha I a emp o explain, gi en ha i is in
he s udy o he op imal size o go e nmen , conside ed by many o be he mos used and app op ia e
me ic o measu e he inc emen in wel a e ha esul s om in insic changes in go e nmen s uc u es.
The Uni ed Na ions De elopmen P og am (UNDP) makes a ailable mos da a ela ed o de elopmen
me ics, mo e pa icula ly HDI and IHDI, which a e bo h used in his wo k. As ep esen ed in sec ion 2.2,
he HDI is compu ed by he geome ic mean o no malised indices o h ee key dimensions o human
de elopmen : a long and heal hy li e (heal h index), access o knowledge (educa ion index), and decen
21

s anda ds o li ing (income index). The heal h index is calcula ed using in o ma ion abou he li e ex-
pec ancy a bi h and educa ion by using a mix o da a o expec ed yea s o schooling and mean yea s o
schooling, and he la e is es ima ed using da a o he G oss Na ional Income (GNI) pe capi a in cons an
2017 Pu chasing Powe Pa i y (PPP).
Table 1: Dimensions o HDI and IHDI.
Dimension Indica o
Heal h Li e expec ancy a bi h (yea s)
Educa ion Expec ed yea s o schooling (yea s)
Mean yea s o schooling (yea s)
S anda d o li ing GNI pe capi a (2017 PPP$)
Dimension index
=
Ac ual alue
−
Minimum alue
Maximum alue
−
Minimum alue
HDI
=3
√
Income index
·
Educa ion index
·
Heal h index
The IHDI, simila ly o i s p edecesso , can only a y be ween ze o and one, wi h highe alues co e-
sponding o highe le els o de elopmen , none heless, he IHDI ends o p esen lowe alues gi en ha i
ollows he same compu a ions as he HDI bu is la e discoun ed acco ding o he wi hin-coun y inequali y
(A) a each o he key componen s o de elopmen .
IHDI
=3
√
Income index*
·
Educa ion index*
·
Heal h index*
=3
√(1 −A
Heal h
)·(1 −A
Educa ion
)·(1 −A
Income
)
The inequali y unc ion (A) ep esen s he inequali y o discoun ac o o each o he unde lying
de elopmen ac o s, which will lead o highe di e gence be ween he HDI and he IHDI he highe i s
alue. Fo in e p e a ion pu poses, he alues o bo h HDI and IHDI we e escaled o ange om ze o o
100.
Ax= 1 −
n
√X1···Xn
¯
X
The cu en da abase, ollowing Goh and Mohd Aznan (2023) sugges ions, also inco po a ed WB
wo ld go e nance indica o s o app aise he de elopmen and g ow h dynamics mo e accu a ely. Such
22
indica o s a e essen ial o p oxy o managemen and ins i u ional compe encies, gi en ha hey desc ibe
b oad pa e ns in he pe cep ions o he quali y o go e nance o e ime.
Mo e pa icula ly, hey a e comp ised o six me ics ha a emp o depic ce ain c ucial ac o s o
go e nance, such as oice and accoun abili y, an indica o ha cap u es he pe cei ed ex en o which
ci izens can selec hei go e nmen as well as eedom o speech; poli ical s abili y and absence o i-
olence/ e o ism, which measu es he pe cei ed likelihood o poli ical ins abili y o associa ed iolence;
and las ly, go e nmen e ec i eness, which cap u es pe cep ions ela ed o he quali y o public se ices
along wi h i s independence om poli ical in luence.
The o he h ee indica o s co espond o egula o y quali y, he pe cep ion o he abili y o go e nmen s
o o mula e and implemen policies and egula ions o p omo e and egula e he p i a e sec o ; ule o
law, which depic s he pe cei ed con idence in abiding by he ules o socie y by law en o cemen , mo e
pa icula ly p ope y igh s, police, and cou s; and las ly, con ol o co up ion, which e lec s he pe cei ed
ex en o which public powe is exe cised o p i a e gain.
All go e nance indica o es ima es a e bound o a ange o alues o -2.5 and 2.5, wi h highe alues
co esponding o be e go e nance. The ule o law in pa icula has commonly been ci ed as pa icula ly
ele an o add ess acco ding o Vedde and Gallaway (1998) and Goh and Mohd Aznan (2023), none he-
less, he implemen a ion o all go e nance indica o s is econome ically in alid, gi en ha hey a e p one
o eplica e and ep esen simila aspec s o go e nance, and, as a di ec consequence, hey end o be
highly ulne able o mul icollinea i y issues.
Acco dingly, he econome ic guidelines o he Va iance In la ion Fac o (VIF) only allow o he imple-
men a ion o a ce ain amoun o wo ld go e nance indica o s, and as such, a new a iable is cons uc ed
o allow o he implemen a ion o he ull amoun o wo ld go e nance indica o s.
Using he da a o he six go e nance indica o s, an o e all go e nance indica o is cons uc ed by
making use o he p incipal componen analysis. The objec i e o p incipal componen analysis is o
educe he numbe o a iables in a da ase while p ese ing as much in o ma ion as possible. B oadly,
his echnique ans o ms a se o highly co ela ed a iables in o a smalle se o unco ela ed a iables,
which we e e o as p incipal componen s, wi hou losing he b oad s a is ical ele ance and signi icance
o he o iginal a iables.
Go e nance
i = (
Voice
i ,
Law
i ,
Regula o y
i ,
S abili y
i ,
E iciency
i ,
Co up ion
i )
In his app oach, only he i s p incipal componen is used o ep esen go e nance, gi en ha i is
capable o explaining up o oughly 80% o he a iance o he o iginal go e nance indica o s.
23
Gi en he dis inc na u e o g ow h and de elopmen dynamics, a bipola econome ic app oach will
la e be implemen ed, and acco dingly, also di e en da a. Fo economic g ow h dynamics, addi ional
da a om he WB is inco po a ed, mo e pa icula ly he g oss capi al o ma ion as a pe cen age o GDP,
ade openness, which e lec s he p opo ion o he mone a y alue o expo s and impo s on he GDP,
and las ly e ia y school en olmen (% g oss).
Fo de elopmen me ics, he da a employed sligh ly di e s, al hough mos o i also de i es om
he WB. In his case, complemen a y in o ma ion o he in an mo ali y a e (pe 1000 bi hs) is also
inco po a ed. The da a inco po a ed was s anda dised o depic alues in pe cen ages, and as such, mos
da a was, when app op ia e, mul iplied by 100 a he simila i y o de elopmen indica o s.
Addi ionally, i is ele an o e e o he lack o a uni o m designa ion o coun y names
and coun y codes, which made i essen ial o op o a uni o m app oach o depic he indi id-
uals being add essed, as such, his da abase also makes use o in o ma ion e ie ed om he
In e na ional O ganiza ion o S anda diza ion (ISO) o uni o mly designa e coun ies.
A e me ging he da a and ans o ming i o ensu e sa e speci ica ions o he a iables, i was la e
di ided in o h ee sepa a e da abases, which a e used o add ess he di icul in e play be ween s a is ical
ele ance and econome ic adequacy, which is common wi hin he s udy o op imal expendi u e due o a
consis en ly low numbe o obse a ions.
The i s da ase co esponds o an annual da abase ha anges om 1960 o 2023 o a se o
228 coun ies, which will la e be use ul o assess p edic i e and p elimina y dynamics, as well as some
addi ional es ima ions.
The second co esponds o a da abase wi h econome ic use ulness, in which he yea s we e d opped
and we e ans o med in o non-o e lapping 3-yea pe iod a e ages o mos a iables, excep GDP pe
capi a, HDI,IHDI, popula ion, and e ia y school en olmen .
XT=x +x −1+x −2
3o XT=x , T ={1, . . . , 21}, ={1962, . . . , 2023}
The la e da abase depic s in o ma ion o wen y-one non-o e lapping 3-yea pe iods (1962-1964,
1965-1967, 1968-1970, 1971-1973, 1974-1976, 1977-1979, 1980-1982, 1983-1985, 1986-1988, 1989-
1991, 1992-1994, 1995-1997, 1998-2000, 2001-2003, 2004-2006, 2007-2009, 2010-2012, 2013-2015,
2016-2018, 2019-2021, 2022-2023), and, simila ly o he annual da abase, i is also o a se o 228
coun ies.
The a ionale as o why his echnique was ca ied ou and he mo i e o only es ima e a e ages o
ce ain a iables a e ela ed o economic heo y and econome ic implemen a ion, and as such will only
24
be exhibi ed in he acco dan sec ion. Addi ionally, he hi d da abase co esponds o he same app oach,
bu only o 5-yea pe iods.
As o GDP pe capi a, HDI,IHDI, popula ion, and e ia y school en olmen , he alues a e no a e ages
since he alues ep esen he obse a ions o he las yea o he espec i e 3-yea o 5-yea pe iods.
Addi ionally, using such a iables, excep o school en olmen , loga i hmic g ow h a es we e cons uc ed.
log Y =log Y0+gY·
gY· =log Y −log Y0
gY·100 = 1
·(log Y −log Y0)·100
Gi en ha we use pe iods, he g ow h a es we e cons uc ed by using he loga i hmic change be ween
pe iods o hose a iables, di ided by he numbe o yea s o he pe iods. Such a p ocedu e allows o
he compu a ion o he a e age annual g ow h a es o such indica o s. Las ly, when mul iplied by 100, i
is possible o depic he pe cen age g ow h a e in con inuous ime.
3.1.2 Desc ip i e s a is ics and p elimina y esul s
In his sec ion, i is possible o add ess any issue ha migh be p esen in he da a used in he
empi ical me hodology. Table 2p esen s he join desc ip i e s a is ics o he non-o e lapping 3-yea
pe iod a e ages da abase, which will be used o compu e he op imal go e nmen expendi u e es ima es
o g ow h and de elopmen me ics. The i s no able dynamic is he numbe o obse a ions o di e en
economic indica o s, which a e no ably spa se. As men ioned abo e, he coun y’s dynamics o op imal
go e nmen expendi u e su e om da a a ailabili y issues, a phenomenon simila o mos o he p e ious
li e a u e.
Al hough such da a is mac oeconomically undamen al o unco e ing ela ionships ela ed o go e n-
men size, one mus accoun o i s na u e, as a low numbe o obse a ions can in oduce delicacy when
wo king wi h such da a.
The oughly 800 obse a ions o GDP pe capi a and HDI do no p esen he bigges issue, as his is
common in his ield o s udy, and when co ec ly add essed, should no pose isks o he alidi y o he
esul s. Da ies (2009), Ma ins and Veiga (2014), and Asimakopoulos and Ka a ias (2016), o ins ance,
p esen a simila o e en lowe numbe o obse a ions.
None heless, gi en he ecen enu e o IHDI, his me ic aises he mos p essing conce n, especially
as i is cen al o his s udy, wi h only 355 obse a ions. While o he wo ks occasionally use simila sample
sizes, ca e ul in e p e a ion is c ucial o mi iga e po en ial small sample bias.
25
li e a u e, o conduc a disagg ega ed sc u iny. A uni ied analysis ha igno es such he e ogenei y could
obscu e he A mey Cu e ela ionship.
I is impo an o no e ha he p e ious assessmen only shows end lines, which a e use ul o mak-
ing p edic ions and p elimina y assump ions abou po en ial u u e ela ionships. They do no , howe e ,
p o ide obus o s a is ically signi ican ools o ex apola ing empi ical in e ences on he op imal size o
go e nmen .
Thei applicabili y lies in guiding expec a ions ega ding ela ionships and challenges ha may a ise.
O e all, hey sugges ha u u e e idence may suppo he exis ence o he A mey Cu e o ha agg ega ion
o de eloped and de eloping coun ies may obscu e such ela ionships. Fu he mo e, hese p elimina y
indings sugges ha conclusions may a y depending on he me ic used, wi h pa icula aliena ion o
economic g ow h.
3.2 Econome ic me hodology
3.2.1 Econome ic app oach o economic g ow h indica o s
As p e iously exhibi ed, he subs an ial con as be ween economic g ow h and economic de elopmen
can imply he e ogenei y in he conclusions in e ed, and as such, dis inc econome ic app oaches a e
impe a i e. Consequen ly, a dual eg ession me hodology is applied, one o economic g ow h and he
o he o de elopmen indica o s, al hough bo h sha e pa icula cha ac e is ics in hei es ima ion.
Fo ins ance, as e e ed o in subsec ion 3.1.1, he da a is a e aged o e 3-yea pe iods o mos
a iables, and acco dingly, lags and loga i hmic g ow h a es a e implemen ed o cons uc app op ia e
dependen and explana o y a iables.
Wha hasn’ been discussed none heless is he a ionale o such a p ocedu e. The main mo i e o
such a echnique elies on con olling o he e ec s o business cycles in c oss-coun y and ime-se ies
s udies and emphasising he ele an a iables’ impac s on long- e m economic o de elopmen g ow h
(S u n and Eps ein,2021;Beck and Le ine,2004).
B oadly, he use o non-o e lapping pe iods allows o smoo hing ou sho - e m luc ua ions and o-
cusing on he impac o s uc u al changes by educing he noise om such anomalies and emphasising
long- e m economic dynamics. As ci ed by Beck and Le ine (2004), a e aging annual da a in o non-
o e lapping 5-yea pe iods allows o he pu ging o cyclical bias and consequen ly depic s mo e accu a e
conclusions.
The li e a u e, howe e , s esses he dilemma o how many yea s should be used o cons uc he
32

non-o e lapping pe iods, wi h some e en sugges ing ha a e aging in 5-yea pe iods may no be enough
o e ec i ely con ol o such cycles, gi en ha hey can commonly be pe sis en o la ge du a ions
(Schula ick and Solomou,2011;Loayza and Ranciè e,2006).
E en so, in he cu en amewo k, smalle pe iods o only 3-yea pe iods we e exe cised, conside ing
he ulne abili y and he necessi y o coun e balancing econome ic sui abili y wi h s a is ical ele ance.
Such judgemen ela es o he ac ha he inc emen in he obus ness o esul s esul ing om inc emen al
obse a ions om changing om 5-yea pe iods o 3-yea pe iods migh likely compensa e o he loss in
obus ness ha de i es om no con olling as accu a ely he business cycles.
Acco dingly, gi en his limi a ion, i is o o emos conce n o app oach ca e ully esul ing compu a ions,
and as such, u u e esul s will commonly be compu ed also o 5-yea pe iods o ensu e he s u diness o
he causal in e ence.
Addi ionally, mos li e a u e ocuses on economic g ow h and i s p obably quan i a i ely impo an con-
nec ion wi h business cycles, which may also ein o ce economic g ow h (Kaiha su e al.,2019;Pede sen
and Elme ,2003). None heless, e idence on business cycle dynamics wi h de elopmen indica o s is
lacking, and hus, he assump ion ha a e aging o 5-yea pe iods o bo h g ow h and de elopmen is
un ounded.
Consequen ly, adop ing he same non-o e lapping pe iods migh imply allacies, gi en ha i assumes
ha bo h indica o s o economic and de elopmen pe o mance a e p one o he same du a ion o business
cycles. By a ionale, gi en ha de elopmen implies mo e s uc u al changes, while economic g ow h has
mo e as -paced dynamics, i migh be ele an o app aise i longe a e ages may be mo e app op ia e o
he de elopmen indica o s, while o g ow h sho e a e ages migh be su icien , hus u he alida ing
he use o di e en pe iods o add ess business cycles incidence and i s mechanisms.
Wi h his opic so ed, he cons uc ion o he econome ic es ima ion can be assembled. Model 1
depic s he cu en es ima ion, which is used o assess he benchma k esul s o dynamics ela ing o he
size o go e nmen and economic g ow h. A he simila i y o mos li e a u e, his model a emp s o es i-
ma e an op imal go e nmen expendi u e le el o u ning poin , i exis en , acco ding o he maximisa ion
o economic g ow h, which, in his scena io, is ep esen ed by he GDP pe capi a annual a e age g ow h
a e o a coun y ia pe iod (∆GDP), and as s a ed in he p e ious sec ion, is compu ed using he
loga i hmic change a e be ween wo consecu i e pe iods.
33
∆
GDP
i, =αi+β0log
GDP
i, −1+β1
Educa ion
i, −1+β2
Size
i, +
β3
Size
2
i, +β4
GCF
i, +β5
Go e nance
i, +β6∆
Popula ion
i, +
β7
T ade
i, +γ +ϵi,
(Model 1)
Gi en he li e a u e on economic con e gence, which ends o suppo he exis ence o condi ional
con e gence, i is expec ed ha he p esen le el o economic ad ancemen migh explain a conside able
amoun o u u e g ow h, gi en ha coun ies end o end in a coun y-speci ic s eady s a e ha is depen-
den on hei speci ic labou o ce and sa ings speci ica ions (Ba o e al.,1991,1992;Cho and G aham,
1996;Kai ila,2004;Nonneman and Vanhoud ,1996;Mu hy and Chien,1997).
Consequen ly, he loga i hm o he las yea o he 3-yea pe iods o GDP pe capi a o coun y ia
pe iod −1(log GDP) was implemen ed as a con ol a iable. As sugges ed by Ba o (2016), his a iable is
an icipa ed o ha e a nega i e coe icien i condi ional con e gence holds, since wi h inc easing economic
p ospe i y, dec easing ma ginal e u ns o g ow h s a o ake o m, making i inc easingly ha de o
sus ain high a es o g ow h. Addi ionally, he inclusion o such a a iable allows he econome ic model o
be dynamic and, as such, depic s he pas alues o he dependen a iable o in luence i s p esen alue,
which helps cap u e he se ial pe sis ence o economic mechanisms o g ow h (Bha ga a and Sa gan,
1983;A ellano and Bond,1991).
As p e iously e e ed o, howe e , ha a iable only es s condi ional con e gence i con olled o
coun y-speci ic labou and sa ing se ings, gi en ha i ha does no hold, wha is being es ed is absolu e
con e gence. In his sense, o con ol o he labou o ce, he popula ion annual a e age g ow h a e o
coun y ia pe iod (Popula ion) compu ed by he loga i hmic change is also included, gi ing Ram (1986)
sugges ions ha i se es as a p oxy o labou o ce g ow h. Al hough a s able and obus labou o ce is
an essen ial ounda ional ool o sus ain a consis en ax base and wo king o ce, an abusi e inc emen al
popula ion migh make i ha de o keep up wi h a as - acked demand o jobs and cen al se ices, and
he e o e we can p ojec such a a iable o be non-signi ican o e en o s i le g ow h.
Analogously o Asimakopoulos and Ka a ias (2016), he a e age g oss capi al o ma ion as a pe cen -
age o GDP o coun y ia pe iod (GCF) is also inco po a ed, o mainly wo main easons. The i s
elies on he necessi y o con olling o he sa ings dynamics o in e condi ional con e gence, gi en he
p oposal o Hajamini and Falahi (2018) ha in es men as a pe cen age o GDP can be used as a subs i u e
o he sa ings a e.
Addi ionally, capi al o ma ion is also ele an o ac o ing in he ele ance o p i a e in es men and
34
i s mul iplying impac on economic g ow h.1Consequen ly, i is o eseen ha i will ha e an encou aging
and signi ican impac on he g ow h p ospec s.
The main explana o y a iable unde analysis e e s o go e nmen size, which is ep esen ed he e
as he a e age go e nmen expendi u e as a pe cen age o GDP o coun y ia pe iod (Size), which
is also inco po a ed in i s quad a ic o m (Size2). The quad a ic app oach is implemen ed as a de aul
conside ing ha mos ecen academics con on ed he adi ional app oach o linea ela ions and ound
he s a is ical pe inence o ca ying ou non-linea ela ionships (Hajamini and Falahi,2018).
As a esul , o he ela ionship o he A mey Cu e, which is expec ed, he go e nmen size should
depic a posi i e coe icien while i s squa e should exhibi a nega i e e ec , wi h he esul ing u ning poin
whe e inc eases in go e nmen expendi u e no longe imply ma ginal enhancemen s in economic g ow h,
con eying he op imal go e nmen size.
Fu he mo e, ade openness, he sha e o he mone a y alue o expo s and impo s on he o e all
GDP in coun y ia pe iod (T ade), is also included, ollowing he econome ic me hodology o bo h
A onso and Fu ce i (2010) and Asimakopoulos and Ka a ias (2016), o accoun o he e ec ha he
inse ion in in e na ional ade o e en ade libe alisa ion policies exe on economic g ow h. The e o e,
we migh an icipa e a posi i e and signi ican impac o ha a iable simila o A onso and Fu ce i (2010),
o a leas o de eloping coun ies alike Asimakopoulos and Ka a ias (2016).
Gi en he ele ance o con ol o he impac o educa ion on g ow h, he e ia y school en olmen ,
exp essed as a pe cen age o he popula ion o e ia y educa ion age, o he las yea o he 3-yea pe iods
o coun y ia pe iod −1(Educa ion) was also in eg a ed. Con a y o o he schola ship, such as Ma ins
and Veiga (2014), seconda y school en olmen was no used, gi en ha nowadays i lacks he e ogenei y
acco ding o he le el o economic p ospe i y gi en he ecen wo ldwide imp o emen in basic le els o
educa ion p ospec s. Gi en he p obable bene icial impac o a quali ied wo k o ce on g ow h, i is expec ed
ha he a iable will p o ide a posi i e coe icien .
Las ly, as depic ed in subsec ion 2.3.1, go e nance is o di e ele ance in explaining mos o he
he e ogenei y in economic p ospe i y be ween ich and poo coun ies (Acemoglu e al.,2001;No h
and Thomas,1973). Hence, he indica o o b oad go e nance o coun y ia pe iod (Go e nance),
cons uc ed using p incipal componen analysis, was also inco po a ed. Acco ding o in e ences om he
li e a u e, i should ha e a posi i e and s a is ically signi ican e ec on economic g ow h.
Gi en he subs an ial endogenei y p esen in dynamic models o economic g ow h, pa icula ly
due o he co ela ion be ween he lagged dependen a iable and he e o e m, a wo-s ep Sys em-
1G oss capi al o ma ion is used as a p oxy o p i a e in es men , gi en ha public sec o in es men ep esen s only a negligeable sha e o o al in es men .
35
Gene alised Me hod o Momen s (GMM) es ima o wi h ime Fixed E ec s (FE) is used o educe bias and
inc ease he es ima ion p ecision (Bond,2002).2
The use o Sys em-GMM is mos app op ia e when he panel uni s, such as coun ies, a e la ge, while
he pe iods a e ela i ely small, a cha ac e is ic sha ed by he p esen 3-yea and 5-yea pe iod da abase.
Addi ionally, he Di e ence-GMM has been ound o ha e poo ini e sample p ope ies, pa icula ly in
cases whe e he ime se ies a e highly pe sis en (Blundell and Bond,1998;Bond e al.,2001).
Fu he mo e, applying he Bond ule o humb sugges s ha Sys em-GMM is likely mo e app op ia e
han Di e ence-GMM. In Model 1, all explana o y a iables a e ea ed as endogenous, excep o he
pe iod dummies, which se e as exogenous ins umen s. Fo endogenous a iables, lagged alues om a
leas wo pe iods a e used as ins umen s. Mo eo e , o ensu e he obus ness o my empi ical in e ences,
I compu e he esul s using obus s anda d e o s o co ec o he e oskedas ici y and po en ial se ial
co ela ion in he e o e ms.
The in eg a ion o Sys em-GMM is also highly ele an because i helps add ess a mul i ude o sou ces
o endogenei y, such as unobse ed, simul anei y, and dynamic endogenei y (Ullah e al.,2018). Tack-
ling endogenei y de i ing om simul anei y is o special impo ance in his pa icula model, gi en ha
al hough go e nmen size migh heo e ically a ec g ow h, he e is also e idence ha g ow h can also
a ec go e nmen size (Peacock and Sco ,2000).
3.2.2 Econome ic app oach o economic de elopmen indica o s
In he es ima ion used o assess de elopmen and inclusi e de elopmen , he econome ic me hodol-
ogy is simila , as a di ec consequence o he pa allel na u e o HDI and IHDI, and as such a e explained
by he same a iables (excep o hei lags). Simila o he me hodology e e ed o in economic g ow h
modelling, he dynamics ela ed o business cycles imply he main use o 3-yea pe iods bu also 5-yea
pe iods due o obus ness issues and business cycle du a ion.
Model 2 exhibi s he cons uc ion o he econome ic model o de elopmen indica o s, which is simi-
la o Model 1 gi en he s uc u al simila i ies be ween economic g ow h and de elopmen . Consequen ly,
simila ly o Model 1, his new model also a emp s o compu e he u ning poin o he go e nmen size,
al hough p esen ly o bo h de elopmen and inclusi e de elopmen , he i s ep esen ed he e by he HDI
annual a e age g ow h a e o coun y iin pe iod (∆HDI) and he la e depic ed by he IHDI annual
a e age g ow h a e o coun y ia pe iod (∆IHDI), bo h compu ed by loga i hmic g ow h a es.
2Gi en he po en ial lack o s a is ical signi icance in Sys em-GMM es ima es, indi idual and ime FE eg essions a e also used o ensu e obus ness.
36
∆
HDI/IHDI
i, =αi+β0log
HDI/IHDI
i, −1+β1
Educa ion
i, −1+
β2
Size
i, +β3
Size
2
i, +β4
GCF
i, +β5
Go e nance
i, +
β6
In an
i, +γ +ϵi,
(Model 2)
Gi en he simila i y o bo h economic g ow h and de elopmen indica o s, co obo a ed by o he au-
ho s’ econome ic app oaches o de elopmen such as Da ies (2009) and Ma ins and Veiga (2014), we
could expec ha ce ain explana o y a iables should emain ele an when applied o assess de elopmen
dynamics.
In his sense, he e ia y school en olmen , g oss capi al o ma ion, go e nance, and su ely go e n-
men expendi u e in i s quad a ic o m a e all explana o y a iables ha we e also implemen ed in Model 2.
Gi en he simila i y o bo h models, we could an icipa e he s a is ical ele ance and he di ec ion o he
impac o such indica o s o be a close e lec ion o hei an icipa ed beha iou in Model 1.
Building on his a gumen , we could an icipa e ha e ia y school en ollmen , go e nance, and g oss
capi al o ma ion would ha e a posi i e and s a is ically signi ican impac on bo h me ics o de elopmen .
As o go e nmen size, he e ec should be quad a ic, simila o ha o he A mey Cu e, wi h Da ies
(2009) and Ma ins and Veiga (2014) ein o cing such plausibili y due o hei indings o quad a ic impac s
o go e nmen size on HDI.
Only he a iables o ade openness and popula ion g ow h we e d opped in he p esen model. The
ade openness was emo ed due o he in insic mone a y ele ance o ade, which does no anspose
as well o de elopmen indica o s as one should expec . Gi en ha inc emen s in de elopmen a e o en
a esul o s uc u al changes in a coun y, he oo-dilu ed e ec o ade on de elopmen leads o a lack
o s a is ical ele ance o ha a iable.
Gi en ha ade impac s economic p ospe i y, which is only one o many componen s o de elopmen ,
i s impac would only be obse ed h ough indi ec mechanisms, and consequen ly, he inclusion o ade
would imply mo e noise in he es ima ion han ac ual empi ical obus ness.
Simila ly, his pa ially holds o popula ion g ow h, gi en ha economic g ow h accoun ing o en dis-
plays popula ion g ow h as an impo an ounda ional ac o o economic g ow h, none heless, ha does
no con e well o de elopmen , in which said a iable seems o lack explana o y powe . The easibili y
o a oo o e ly ague ansmission mechanism be ween popula ion and de elopmen and i s o e i ing in
Model 2 ul ima ely led o i s wi hd awal.
To also apply dynamic models, along wi h he heo y o condi ional con e gence, he loga i hms o he
las yea o he pe iod o bo h HDI and IHDI o coun y ia pe iod −1(log HDI, log IHDI) we e also
37

inco po a ed as con ol a iables.
Al hough popula ion g ow h is now lacking in he model, he heo y o condi ional con e gence should
s ill be able o hold, and he e o e, i is an icipable ha he lags o he me ics o de elopmen p o ide a
nega i e coe icien , gi en ha as de elopmen g ows, dec easing ma ginal e u ns s a o punish de el-
opmen p ospec s, ul ima ely leading coun ies o hei own s eady s a es.
The p ac ice is also implemen ed by Da ies (2009) and Ma ins and Veiga (2014), in which he la e
ound e idence o con e gence. Gi en he heal h componen o he de elopmen indica o s, a new a iable
is added o Model 2 which was no p esen in Model 1. Tha a iable is he a e age in an mo ali y a e
o coun y ia pe iod −1(In an ), which is expec ed o exhibi a nega i e and s a is ical impac on
de elopmen , gi en ha he wo se he heal h condi ions o a coun y, he wo se i s de elopmen ou comes.
The econome ic dilemmas and he c i ical obs acle o endogenei y a e s ill p esen in he cu en
model o economic de elopmen , and as such, he same app oach o Model 1 is implemen ed he e,
pa icula ly by unning a wo-s ep Sys em-GMM wi h ime FE.3Model 2 also ea s all a iables as endoge-
nous and uses o ins umen al a iables he pe iod dummies as well as a leas he wo-pe iod lags o
he explana o y a iables. The esul s a e also exp essed wi h obus s anda d e o s o gua an ee highe
accu acy.
3Simila ly o he p e ious case, he po en ial lack o s a is ical ele ance in Sys em-GMM es ima ions also p omp s he use o indi idual and ime FE es ima ions
o ensu e obus ness.
38
4 Resul s
4.1 Main analysis
4.1.1 Benchma k esul s
Table 3 epo s he benchma k esul s de i ed om he baseline econome ic models o 3-yea pe-
iods. The no es desc ibe he econome ic speci ica ions, pa icula ly hei au oco ela ion alidi y, he
Hansen es , he numbe o ins umen s, and an addi ional numbe o s a is ical speci ica ions.
In column (1), we can obse e ha he go e nmen size is highly ela ed o inc emen s in he GDP
g ow h a e in a quad a ic o m. Consequen ly, he augmen in go e nmen expendi u e p o ides high and
signi ican bonuses o economic g ow h p ospec s a i s i s s ages bu la e ends o p o ide lowe and
lowe incen i es, e en ul ima ely eaching haza dous impac s on GDP, simila o he A mey Cu e.
Column (2), which also es s GDP g ow h bu uses FE depic s simila ela ionships. Consequen ly, by
de i ing he equa ion on go e nmen size, we can in e he op imal go e nmen size acco ding o economic
g ow h maximisa ion.
Consequen ly, we can depic ha o he Sys em-GMM es ima ion, he esul s indica e ha he op imum
go e nmen size o expendi u e should be alloca ed a oughly 32.54% o GDP, while o FE ha size is
subs an ially simila , si ua ed a oughly 27.22% o GDP.
Fo bo h columns, he es o he explana o y a iables seem o beha e as expec ed, wi h highe alues
o pas GDP implying lowe g ow h a es in he p esen , simila o he a ionale o condi ional con e gence.
The g oss capi al o ma ion also exhibi s a subs an ial and posi i e impac on economic g ow h. The
bene icial and closely ela ed impac o ade on g ow h is also exhibi ed by i s posi i e coe icien . The
popula ion g ow h nega i e e ec was also an icipa ed.
Educa ion also seems o beha e simila ly o wha was p e iously p ojec ed, wi h a bene icial impac
on g ow h p ospec s, none heless, only o he Sys em-GMM. Go e nance beha es con a ily, gi en ha i
p o ides a subs an ial posi i e impac on g ow h, none heless, only on he FE eg ession.
39
Table 3: Go e nmen size, g ow h, de elopmen , and inclusi e de elopmen o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −1-2.393*** -5.478***
(0.850) (0.838)
Log HDI −1-8.406*** -10.174***
(3.038) (1.038)
Log IHDI −18.483** -18.145***
(3.342) (2.467)
Educa ion −10.043** 0.012 0.016** 0.002 0.006 0.002
(0.020) (0.011) (0.007) (0.003) (0.016) (0.005)
Size 0.621*** 0.304*** 0.026 0.053** -0.223** -0.058
(0.218) (0.098) (0.049) (0.021) (0.103) (0.059)
Size2-0.010*** -0.006*** 0.000 -0.001** 0.003** 0.001
(0.003) (0.001) (0.001) (0.000) (0.001) (0.001)
GCF 0.225*** 0.145*** 0.002 0.015*** 0.049* 0.019
(0.047) (0.027) (0.013) (0.005) (0.029) (0.014)
Go e nance 0.256 0.739** 0.328* 0.059 -0.470* 0.315
(0.608) (0.288) (0.175) (0.067) (0.274) (0.218)
Popula ion -0.682** -0.446***
(0.279) (0.160)
T ade 0.029** 0.017*
(0.012) (0.009)
In an -0.021 -0.025*** 0.141** -0.060
(0.021) (0.006) (0.067) (0.037)
Obse a ions 838 834 820 814 355 343
Coun ies 130 126 130 124 112 124
AR(1) 0.000267 0.00884 0.00267
AR(2) 0.947 0.136 0.996
Hansen es 0.408 0.458 0.440
Ins umen s 122 102 44
Adj. R20.498 0.699 0.586
No es: Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. GMM es ima ions use obus
s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All a iables a e
ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he
a e age annual g ow h a e o GDP pe capi a o (1) and (2), o HDI o (3) and (4), and o IHDI o (5) and (6). Signi icance
le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
40
Columns (3) and (4) display HDI dynamics o Sys em-GMM and FE, espec i ely, wi h a no iceable
educ ion in he s a is ical ele ance o such a model, a leas o 3-yea pe iods. Column (3) does no e en
depic a s a is ical e ec o go e nmen size, wi h he only a iables epo ing s a is ical signi icance being
he lagged dependen a iable, wi h a nega i e impac , and educa ion and go e nance wi h a posi i e one.
Using FE, he s a is ical signi icance seems o be augmen ed, wi h he quad a ic e ec o go e nmen
size being displayed he e. The e o e, by decomposing i , we can de ec ha he op imal go e nmen size
o HDI, a leas in his case, is si ua ed a a ound 42.20% o GDP. As o he es o he a iables, a leas
he signi ican ones, hey all seem o beha e as an icipa ed in he p e ious sec ion.
Fo columns (5) and (6), we can also in e he mechanisms o IHDI, howe e , as displayed p e iously,
wi h cau ion gi en he lacking numbe o obse a ions. Fo bo h Sys em-GMM and FE, i is i s isible
ha condi ional con e gence applies, gi en ha he lags o he dependen p o ide a haza dous impac
on inclusi e de elopmen g ow h. We can obse e, howe e , ha p obably due o small sample bias, he
e ec s o mos a iables appea as a he peculia in Sys em-GMM. Al hough he g oss capi al o ma ion
appea s as posi i e, as an icipa ed, he go e nance displays a nega i e impac on IHDI, wi h go e nmen
size also displaying an awkwa d ela ionship.
Pa icula ly, go e nmen size, al hough p o iding a quad a ic e ec , is in he opposi e di ec ion, whe e
size inc emen s always imply nega i e e ec s on IHDI un il a ce ain poin , om whe e any inc emen
always leads o bene icial e ec s. No only does his e ec con adic mos heo ies, bu i also implies
ha he e is no maximisa ion poin o IHDI.
Con adic o ily, i p o ides a minimising poin whe e he nega i e e ec o go e nmen size on IHDI
g ow h peaks, co esponding in his case o 44.10% o GDP. As o FE, he esul s p o ide almos null
s a is ical ele ance, excep o he lag o IHDI, which implies a nega i e impac o he p esen le el o IHDI
on u u e IHDI g ow h p ospec s.
As p e iously depic ed, he business cycle incidence o economic g ow h and de elopmen migh di e ,
and as such, esul s o 5-yea pe iods a e also exhibi ed in Table 4, in an a emp o assess mo e obus
esul s, gi en he commonly slippe y and s a is ically medioc e esilience o his ype o da a.
Columns (1) and (2) depic he 5-yea pe iods esul s o economic g ow h, whe e, o bo h, he
s a is ical implica ions o condi ional con e gence also seem o be p esen . The Sys em-GMM eg ession
appea s as s a is ically weake han when compu ed o 3-yea pe iods in Table 3. Excep o he lag o
GDP, only he g oss capi al o ma ion p o ides s a is ical and posi i e e ec s on economic g ow h. The FE,
howe e , seems s a is ically signi ican , wi h all s a is ically signi ican a iables p o iding esul s ha we e
p e iously an icipa ed. As o go e nmen size, we can also de i e ha he go e nmen size is quad a ic
41
un il oughly 40% o GDP, and beyond ha , al hough he di ec ion becomes nega i e, he e is no s a is ical
signi icance.
In Figu e 11, i is isible ha he impac o he size o he go e nmen on IHDI g ow h is puzzling,
wi h he ma ginal impac s p o iding no clea insigh in o why such an indica o beha es as i did in he
econome ic model. I is e iden , howe e , ha he e ec o go e nmen size on IHDI g ow h p obably does
no beha e like in a U-shaped o m, gi en he opposi e and andom di ec ions o he ma ginal models.
As such, i is plausible ha small sample bias may be damaging he di ec ion o he esul s, wi h he
consequence being somewha andom e ec s o go e nmen size on IHDI.
Conside ing hose ongoing conce ns, al e na i e econome ic app oaches a e implemen ed o judge
he de imen al e ec s o small sample biases on esul s ela ed o IHDI. Table 5 e lec s he equi alen
esul s, whe e some econome ic con ols a e elaxed o inc emen sample dimensions and consequen ly
depic possibly mo e accu a e esul s by ading o econome ic igou o mo e sample obus ness.
Column (1) loosens he 3-yea Sys em-GMM eg ession by d opping he g oss capi al o ma ion, go -
e nance, and e ia y en olmen , and as a esul , inc easing obse a ions by oughly 30%. In his case, he
quad a ic e ec akes o m, wi h an op imal go e nmen size o 39.04% o GDP. Column (2), which uses
FE, only depic s simila associa ions i i u he excludes in an mo ali y. In his case, he op imal alue
is oughly 40.69%.
I is clea ha he small sample pa ially conceals he e idence o he A mey Cu e, none heless, he
dismissal o such a la ge numbe o con ols migh imply oo much delicacy o he esul s, while a he
same ime, exp essing an academic allacy o selec ion bias by only emo ing he a iables ha a e i ing
o my na a i e.
Consequen ly, o depic he empi ical e aci y and no au ho ampe ing, columns (3) and (4) exe
he same p ac ice o 5-yea pe iods, wi h he i s only emo ing he educa ion en olmen , he a iable
ha allows us o ex ac he la ges numbe o obse a ions, while he la e also emo es g oss capi al
o ma ion. The esul s o bo h columns (3) and (4) also exhibi e idence o he A mey Cu e, wi h op imal
sizes o 37.72% and 31.23% o GDP, espec i ely.
Conside ing he e idence he e con eyed, he e is sa is ac o y eason o assume ha he esul s de-
pic ed exp ess he i s empi ical e idence o an A mey Cu e ela ionship o IHDI, none heless, a leas
o now, wi h subs an ially less s a is ical obus ness. Gi en ha he esul s o IHDI we e no compu ed
using p ocedu es equi alen o HDI, i is no accu a e o ep esen i he op imal go e nmen size o IHDI
is g ea e han HDI, ne e heless, i is clea ha he gene al endency is o IHDI o epo a leas sizes
simila o hose o HDI. Speci ically, we could in e ha he op imal go e nmen size o IHDI anges om
48

Table 5: Al e na i e es ima ions o inclusi e de elopmen
3-yea pe iods 5-yea pe iods
(1) (2) (3) (4)
Va iables Sys-GMM FE Sys-GMM FE
Log IHDI −11.311 -14.910*** 4.761 -16.544***
(2.362) (1.160) (3.142) (1.816)
Size 0.214*** 0.110** 0.150* 0.115*
(0.059) (0.051) (0.078) (0.069)
Size2-0.003*** -0.001* -0.002** -0.002*
(0.001) (0.001) (0.001) (0.001)
GCF 0.049*
(0.028)
Go e nance -0.212 0.430*
(0.299) (0.231)
In an 0.064 0.107** -0.039
(0.046) (0.043) (0.036)
Obse a ions 461 462 327 335
Coun ies 128 126 121 125
AR(1) 0.000147 0.0340
AR(2) 0.193
Hansen es 0.126 0.324
Ins umen s 65 35
Adj. R20.515 0.404
No es: Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea o 5-yea pe iod dummies. GMM es ima ions
use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All
a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen
a iable is he a e age annual g ow h a e o IHDI o all eg essions. Signi icance le el a which he null hypo hesis is
ejec ed: ***, 1%; **, 5%, and *, 10%.
49
31.23% o 40.69% o GDP.
In o de no o implemen oo much noise in he analysis and a oid o e complica ion o esul s by
al e na ing be ween a iables, u u e esul s epo ed o IHDI will be compu ed using he o iginal con ols
used in HDI, none heless, he conce n o small sample is no igno ed, and con using o lacking e idence
may be exp essed as a di ec consequence o small sample, and no o causali y.
4.1.2 Componen s o go e nance
As depic ed in he li e a u e, his disse a ion also ies o elucida e some o he in e ences and mech-
anisms by which go e nance a ec s wel a e, which in he p e ious esul s has al eady been empi ically
highly bene icial and a ou able o many economic indica o s.
Tables 6and 20 depic esul s o add ess which pa icula componen s o go e nance seem o a ec
mos economic g ow h. One conclusion, which is pa allel o bo h ables, is easily no iceable. Tha ela es
o he ac ha wo indica o s o go e nance appea o be pa icula ly ele an o GDP g ow h, pa icula ly
go e nmen e ec i eness and poli ical s abili y and absence o iolence/ e o ism.
As o he es o he a iables, hey do no seem as ele an o explain g ow h, a leas when add essed
indi idually. The esul s o HDI and IHDI a e no depic ed due o he s a is ical insigni icance o all
indi idual go e nance indica o s, which migh lead us o ques ion i we can ex apola e ha go e nance
is only b oadly ele an o bo h measu es o de elopmen and ha he pa icula emphasis on speci ic
componen s appea s as i ele an , con a y o GDP, whe e a ocus on he speci ied wo ld go e nance
indica o s is mo e pe inen .
This migh p omp us o inqui e whe he e ec i eness and poli ical s abili y a e p imo dial o p omo e
p i a e sec o ac i i y and economic p ospe i y, none heless, only ocusing on hose indica o s may no
bene i e e y indi idual equally, and as such, pa icula go e nance componen s appea i ele an o de-
elopmen me ics. Consequen ly, o p omo e dis ibu ed wel a e, he equal and widesp ead ocus on all
go e nance appea s as mo e app op ia e.
50
Table 6: Go e nance componen s and economic g ow h o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables FE FE FE FE FE FE
Log GDP −1-4.809*** -4.795*** -4.961*** -5.327*** -5.525*** -5.100***
(0.692) (0.788) (0.756) (0.768) (0.777) (0.765)
Educa ion −10.011 0.010 0.011 0.011 0.009 0.012
(0.011) (0.010) (0.011) (0.011) (0.011) (0.011)
Size 0.327*** 0.331*** 0.327*** 0.308*** 0.293*** 0.321***
(0.103) (0.101) (0.099) (0.102) (0.094) (0.102)
Size2-0.006*** -0.006*** -0.006*** -0.006*** -0.005*** -0.006***
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
GCF 0.148*** 0.147*** 0.146*** 0.138*** 0.146*** 0.146***
(0.027) (0.027) (0.027) (0.028) (0.027) (0.027)
Popula ion -0.438*** -0.426*** -0.432*** -0.446*** -0.419** -0.435***
(0.152) (0.153) (0.157) (0.153) (0.161) (0.153)
T ade 0.016* 0.016* 0.016* 0.019** 0.018* 0.016*
(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)
Voice 0.495
(0.540)
Law 0.064
(0.600)
Regula o y 0.544
(0.479)
S abili y 1.203***
(0.346)
E ec i eness 1.418***
(0.484)
Co up ion 0.771
(0.499)
Obse a ions 834 834 834 834 834 834
Coun ies 126 126 126 126 126 126
Adj. R20.491 0.490 0.491 0.505 0.498 0.492
No es: Fixed e ec s eg essions, including 3-yea pe iod dummies. FE es ima ions use s anda d e o s clus e ed by coun y
and obus o he e oskedas ici y. The dependen a iable is he a e age annual g ow h a e o GDP pe capi a o all
eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
51
4.1.3 The in e play o go e nmen quali y and quan i y
Al hough go e nmen size is he p ima y ocus o his s udy, i s implica ions and in e connec ion wi h
ins i u ional quali y, as highligh ed by Coo ay (2009), a e also highly ele an . Acco dingly, his sec ion
examines his dynamic.
In pa icula , ollowing a me hodology simila o Coo ay (2009), Table 7p esen s he esul s o he
in e ac ion be ween go e nmen size and a ca ego ical a iable o quan iles o go e nance, namely, low,
middle, and high go e nance coun ies. The empi ical e idence o IHDI, in pa icula , is no ably lacking.
Howe e , o GDP and HDI, s a is ical signi icance is abundan , pa icula ly in low go e nance coun ies,
which, su p isingly, show s onge e idence o he A mey Cu e ela ionship o bo h indica o s.
This inding is no only peculia bu also di ec ly con as s wi h he e idence o Coo ay (2009) and
A onso and Jalles (2016), which sugges ha coun ies wi h weake ins i u ions should be less e icien in
managing public expendi u e.
Fu he mo e, he e idence o he A mey Cu e weakens as go e nance imp o es. As a esul , he
e ec is less p onounced o middle go e nance coun ies and e en comple ely disappea s o high go -
e nance coun ies.
Speci ically, o low go e nance coun ies, he es ima ed op imal go e nmen size anges om 27.38%
o 28.38% o GDP and om 28.15% o 30.97% o HDI. Fo middle go e nance coun ies, he op imal size
is 26.74% o GDP and anges om 29.60% o 37.97% o HDI.
Al hough hese esul s may appea pa adoxical, hei s a is ical consis ency suppo s hei po en ial
alidi y. I e i ied, hese indings could sugges he need o econside cu en de elopmen s a egies
employed by na ions wo ldwide. The causal ela ionship depic ed he e highligh s ha while go e nance is
c ucial o p ospe i y, adjus ing go e nmen size as a policy o s imula e p ospe i y migh be e en mo e im-
po an han wha p e iously hough in o e coming po e y aps, e en in coun ies wi h weak ins i u ional
ounda ions.
52
Table 7: In e ac ions o go e nmen size and go e nmen quali y
3-yea pe iods 5-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables FE FE FE FE FE FE
Log GDP −1-4.759*** -5.019***
(0.715) (0.698)
Log HDI −1-10.070*** -11.134***
(0.971) (1.132)
Log IHDI −1-18.412*** -17.646***
(2.650) (3.125)
Educa ion −10.012 0.003 -0.000 0.034*** 0.006*** 0.007
(0.009) (0.003) (0.005) (0.010) (0.002) (0.006)
Size·Go e n=0 0.677*** 0.131*** -0.028 0.696*** 0.147*** 0.153
(0.134) (0.031) (0.121) (0.153) (0.035) (0.116)
Size2·Go e n=0 -0.012*** -0.002*** 0.000 -0.012*** -0.003*** -0.004***
(0.002) (0.001) (0.002) (0.003) (0.001) (0.001)
Size·Go e n=1 0.207 0.049** -0.074 0.242** 0.063** -0.000
(0.137) (0.022) (0.057) (0.100) (0.025) (0.105)
Size2·Go e n=1 -0.004 -0.001** 0.001 -0.005** -0.001*** -0.001
(0.003) (0.000) (0.001) (0.002) (0.000) (0.002)
Size·Go e n=2 -0.170 -0.016 -0.100 -0.142 -0.013 -0.100
(0.147) (0.032) (0.061) (0.144) (0.032) (0.084)
Size2·Go e n=2 -0.000 0.000 0.001* 0.000 0.000 0.001
(0.002) (0.000) (0.001) (0.002) (0.000) (0.001)
GCF 0.146*** 0.017*** 0.025** 0.144*** 0.012** 0.014
(0.026) (0.004) (0.011) (0.034) (0.005) (0.017)
Go e n 7.301*** 1.067** 1.525 6.929*** 1.086*** 2.179
(1.903) (0.420) (1.196) (1.898) (0.402) (1.544)
Popula ion -0.437*** -0.249
(0.159) (0.200)
T ade 0.015* 0.013
(0.009) (0.009)
In an -0.023*** -0.074** -0.029*** -0.051
(0.005) (0.035) (0.007) (0.041)
Obse a ions 834 814 343 565 553 247
Coun ies 126 124 100 121 119 93
Adj. R20.509 0.705 0.597 0.554 0.729 0.485
No es: Fixed e ec s eg essions, including 3-yea o 5-yea pe iod dummies. FE es ima ions use s anda d e o s clus e ed
by coun y and obus o he e oskedas ici y. The dependen a iable is he a e age annual g ow h a e o GDP pe capi a
o (1) and (4), o HDI o (2) and (5), and o IHDI o (3) and (6). Go e n co esponds o quan iles acco ding o he a iable
o go e nance (=0 i low go e nance, =1 i middle, =2 i high). Signi icance le el a which he null hypo hesis is ejec ed:
***, 1%; **, 5%, and *, 10%.
53

4.1.4 Le els o de elopmen and income
The li e a u e also s esses ha he e ec s o de elopmen , pa icula ly when clus e ed by de eloped o
de eloping coun ies, a e ele an o add essing he di e en impac s o go e nmen size. Asimakopoulos
and Ka a ias (2016) and Ma ins and Veiga (2014) depic ha he quad a ic e ec o de eloped coun ies
is highe han ha o he de eloping ones, pa icula ly because de eloped coun ies possess ins i u ions o
highe quali y, and hus he cen al esou ces a e be e employed. As s essed in he li e a u e, howe e ,
ha somewha hin s a an possible e o in he easoning, gi en ha hey a e cap u ing he e ec o
ins i u ional quali y, he e p oxied by go e nance, and no he mechanisms o de elopmen s a us.
F om his pe spec i e, con olling o go e nance should lead o dis inc esul s, some hing which Goh
and Mohd Aznan (2023) had al eady ad oca ed. Consequen ly, Tables 8and 21 depic a simila app oach
o Ma ins and Veiga (2014), bu now con olling o ins i u ional quali y. Table 8uses 3-yea pe iods and
exhibi s di e en in e ences han wha mos li e a u e depic ed. Pa icula ly, he esul s depic in e ed
U-shaped e ec s o go e nmen size on economic g ow h and de elopmen o de eloping coun ies, wi h
espec i e op imal go e nmen sizes anging om 28.45% o 31.99% o GDP and 33.94% o HDI. De el-
oped coun ies appea o beha e quad a ically, none heless uniquely o economic g ow h, epo ing an
op imal size o go e nmen o 39.26%.
Table 21 depic s he esul s o 5-yea pe iods, whe e he quad a ic signi icance o de eloping coun-
ies s ill holds, ne e heless, now no only o GDP and HDI bu also o IHDI. The espec i e esul s depic
an op imal size o go e nmen o 29.55% o GDP, 29.26% o HDI, and 19.30% o IHDI.
Fo de eloped ones, he e ec is also quad a ic o all me ics, none heless, he di ec ion is qui e
dis inc . Peculia ly, he esul s depic U-shaped e ec s o go e nmen size o GDP and IHDI, while only
he HDI depic s he expec ed in e ed U-shaped e ec wi h an op imal go e nmen size o 29.26%.
O e all, he e is s ong co obo a ion ha a e con olling o go e nance, he e is e idence ha de el-
oping coun ies bene i mo e om inc emen s in go e nmen size, none heless, hey a e la e also mo e
penalised o oo much go e nmen , which depic s a s onge A mey Cu e o de eloping coun ies. As o
de eloped coun ies, i is clea ha he e ec is a lo weake , wi h he quad a ic e ec being subs an ially
andom and o en pa adoxical acco ding o he me ic and pe iod es ablished.
The a ionale o such s uc u e may ely on he ac ha de eloping coun ies lack he basic s uc u e
o p omo e g ow h, and consequen ly, any ma ginal imp o emen in go e nmen size is mo e signi ican ly
el , none heless, only when go e nance is con olled.
This may also e lec ha since he p i a e sec o is ex emely ine icien in ha clus e , he e is oom
o g ow he public sec o wi hou oo much o a cos o p i a e sec o e iciency.
54
Table 8: De elopmen dynamics o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −1-3.587** -5.604***
(1.699) (0.774)
Log HDI −112.340** -10.239***
(4.987) (1.016)
Log HDI −2-12.422***
(3.441)
Log IHDI −17.770** -18.156***
(3.151) (2.408)
Educa ion −10.074* 0.017* 0.005 0.002 -0.013 0.001
(0.040) (0.010) (0.007) (0.003) (0.014) (0.005)
De eloping·Size 0.837* 0.473*** 0.054 0.093*** -0.054 0.020
(0.431) (0.161) (0.091) (0.027) (0.189) (0.078)
De eloping·Size2-0.013* -0.008*** -0.000 -0.001*** -0.001 -0.001
(0.007) (0.003) (0.001) (0.000) (0.003) (0.001)
De eloped·Size 0.690* -0.431** 0.094 -0.017 -0.103 -0.079
(0.393) (0.215) (0.077) (0.036) (0.166) (0.055)
De eloped·Size2-0.009* 0.003 -0.001 0.000 0.002 0.001*
(0.005) (0.002) (0.001) (0.000) (0.002) (0.001)
GCF 0.186** 0.137*** -0.001 0.016*** 0.069** 0.020
(0.076) (0.028) (0.016) (0.005) (0.032) (0.014)
Go e nance 0.238 0.752*** -0.239 0.061 -0.757* 0.318
(0.841) (0.285) (0.189) (0.068) (0.423) (0.216)
Popula ion -0.673* -0.456***
(0.374) (0.160)
T ade 0.036*** 0.016*
(0.014) (0.009)
In an 0.031* -0.024*** 0.094* -0.059
(0.016) (0.006) (0.050) (0.037)
Obse a ions 838 834 807 814 355 343
Coun ies 130 126 130 124 112 100
AR(1) 0.000502 0.000357 0.00482
AR(2) 0.778 0.230
Hansen es 0.126 0.271 0.551
Ins umen s 49 53 31
Adj. R20.508 0.703 0.588
No es: See Table 3
55
As o de eloped coun ies, he lack o consis ency may exp ess ha hey a e de eloped o a eason,
gi en ha hey possess s uc u es and in as uc u e enough o sus ain g ow h, and as such, e en i
inc emen al go e nmen expendi u e is applied, con a y o de eloping coun ies, i comes a oo high o
a cos due o he de imen al e ec s on hei highly e icien p i a e sec o s. B oadly, i pa ially e lec s
he beha iou o no al e ing a winning o mula.
This b oadly sugges s ha al hough he A mey Cu e may s ill exis in de eloped coun ies, none he-
less, pu ely go e nmen expendi u e dynamics a e mo e indi ec and hus ha de o assess.
Acco dingly, his depic ion sligh ly con adic s he impac o de eloped coun ies exhibi ed by Asi-
makopoulos and Ka a ias (2016) and Ma ins and Veiga (2014). Fu he mo e, his inding migh s em
om hese s udies add essing he e ec o go e nmen expendi u e wi hou accoun ing o he possible
bias ha may a ise om he omission o ins i u ional quali y. In con as , he e he impac o go e nance
is accoun ed o , and as such, consequen ial esul s p o ide inno a i e in e ences, whose policymaking
ele ance will be ou lined la e .
Simila o Ma ins and Veiga (2014), he income ca ego isa ion is also add essed, none heless, gi en
he lack o s a is ical signi icance, ou ca ego ies o income we e used a he han wo. Tables 9and 22
depic such dynamics o 3-yea and 5-yea pe iods, espec i ely.
Assuming he in insic co ela ion be ween de elopmen and income le els, one could an icipa e ha
low income coun ies could also be mo e a ec ed by changes in go e nmen expendi u e, none heless,
his does no appea o be he case. Fo GDP in pa icula , he quad a ic e ec o go e nmen expendi u e
appea s a leas in one speci ica ion p esen in e e y clus e o income, hus obs uc ing he in e ence o
causali y acco ding o le el o income.
The ela ion o HDI is simila , none heless, in his case, he clus e in which go e nmen expendi u e
has no s a is ical signi icance in lowe -income coun ies. Fo IHDI, ha clus e is middle-lowe income
coun ies, al hough in his case, i is common o he ela ionships o be odd, pa icula ly since i is no
uncommon o he quad a ic ela ionship o be in a U-shape.
In a b oad sense, i is ha d o unco e i he uppe bound o he lowe bound o he pe cen ile o
income p o ides di e en impac s on go e nmen expendi u e, howe e , in a mo e elas ic app oach, we
could s a e ha i appea s o be a e y sligh s a is ical supe io i y o uppe middle and high-income
coun ies han o he es . I g ea ly accommoda ing and choosing he la e in e ence, we could depic
ha he esul s a e on pa wi h Ma ins and Veiga (2014), none heless, p ecau ion is sugges ed, gi en he
lack o s a is ical obus ness o such a s a emen .
56
Table 9: Income dynamics o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −1-3.912*** -5.851***
(1.318) (0.753)
Log HDI −11.002 -10.468***
(4.193) (1.156)
Log GDP −2-7.899**
(3.455)
Log IHDI −1-4.900 -18.757***
(3.417) (2.286)
Educa ion −10.007 0.020* 0.009 0.003 -0.004 0.001
(0.027) (0.011) (0.006) (0.003) (0.013) (0.005)
Low.I·Size 0.393 0.216 -0.089 0.084 -0.461** -0.424**
(0.384) (0.222) (0.118) (0.089) (0.229) (0.166)
Low.I·Size2-0.009 -0.004 0.002 -0.001 0.007 0.008**
(0.009) (0.005) (0.002) (0.002) (0.006) (0.003)
L.M.I·Size 0.540* 0.667** 0.060 0.125** -0.316* 0.147
(0.306) (0.292) (0.100) (0.060) (0.170) (0.117)
L.M.I·Size2-0.008 -0.011** 0.000 -0.002* 0.004 -0.002
(0.007) (0.005) (0.003) (0.001) (0.003) (0.002)
U.M.I·Size 0.725*** 0.715*** 0.152*** 0.083** -0.070 -0.023
(0.248) (0.138) (0.057) (0.041) (0.122) (0.112)
U.M.I·Size2-0.009** -0.012*** -0.002** -0.002** -0.000 -0.001
(0.004) (0.002) (0.001) (0.001) (0.002) (0.001)
High.I·Size 0.908*** -0.387** 0.167** 0.004 -0.093 -0.144**
(0.289) (0.174) (0.068) (0.029) (0.111) (0.055)
High.I·Size2-0.013*** 0.002 -0.002* 0.000 0.001 0.002***
(0.004) (0.002) (0.001) (0.000) (0.001) (0.001)
GCF 0.171*** 0.136*** 0.015 0.016*** 0.044 0.018
(0.052) (0.026) (0.016) (0.005) (0.029) (0.014)
Go e nance 0.134 0.734*** 0.000 0.071 0.043 0.293
(0.530) (0.278) (0.127) (0.070) (0.346) (0.211)
Popula ion -0.412 -0.440***
(0.302) (0.148)
T ade 0.043*** 0.015*
(0.012) (0.008)
In an 0.006 -0.024*** -0.006 -0.065**
(0.016) (0.006) (0.045) (0.031)
Obse a ions 838 834 806 814 355 343
Coun ies 130 126 129 124 112 100
AR(1) 0.000150 0.00121 0.00868
AR(2) 0.698 0.240
Hansen es 0.405 0.253 0.399
Ins umen s 121 73 104
Adj. R20.515 0.706 0.608
No es: See Table 3
57
Table 12: Economic g ow h and COFOG componen s o 3-yea pe iods
(1) (2) (3)
Va iables Sys-GMM FE FE
De ense 2.855***
(1.049)
De ense2-0.375***
(0.122)
Sa e y 5.250**
(2.516)
Sa e y2-1.354**
(0.566)
Social -0.631*
(0.369)
Obse a ions 424 420 427
Coun ies 68 63 65
AR(1) 0.0481
AR(2) 0.374
Hansen es 0.833
Ins umen s 25
Adj. R20.560 0.487
No es: S a is ically ele an Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. Con ols equal o Table 1, bu omi ed due
o size cons ain s. GMM es ima ions use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All
a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he a e age annual g ow h a e
o GDP pe capi a o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
64

Table 13: Economic de elopmen and COFOG componen s o 3-yea pe iods
(1) (2) (3) (4) (5) (6) (7)
Va iables FE Sys-GMM FE Sys-GMM Sys-GMM FE FE
De ense -0.140**
(0.054)
Public -0.051* -0.041***
(0.027) (0.012)
Public20.001* 0.001***
(0.001) (0.000)
Heal h -0.288*
(0.164)
Heal h20.036***
(0.013)
Sa e y 1.079** 0.602***
(0.470) (0.166)
Sa e y2-0.220** -0.132***
(0.087) (0.034)
Social -0.040**
(0.018)
Social20.002**
(0.001)
Obse a ions 404 419 413 418 413 406 413
Coun ies 60 69 63 68 68 61 63
AR(1) 0.00216 0.00520 0.00446
AR(2) 0.932 0.929 0.853
Hansen es 0.153 0.294 0.124
Ins umen s 65 51 65
Adj. R20.695 0.673 0.696 0.671
No es: S a is ically ele an Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. Con ols equal o Table 1, bu omi ed due
o size cons ain s. GMM es ima ions use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All
a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he a e age annual g ow h a e
o HDI o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
65
Table 14: Inclusi e economic de elopmen and COFOG componen s o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM Sys-GMM
De ense -0.362*** -0.196***
(0.135) (0.058)
Economic 0.244**
(0.114)
Economic2-0.006***
(0.002)
Go .Educa ion -0.148*
(0.074)
En i onmen -4.100**
(1.986)
En i onmen 22.778**
(1.246)
Heal h -0.478**
(0.241)
Heal h20.046***
(0.015)
Obse a ions 205 200 207 203 207 207
Coun ies 62 57 62 58 62 62
AR(1) 0.0592 0.00974 0.00695 0.0173
AR(2) 0.872 0.730 0.203 0.450
Hansen es 0.421 0.323 0.193 0.233
Ins umen s 52 38 32 44
Adj. R20.652 0.621
No es: No es: S a is ically ele an Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. Con ols equal o Table 1, bu omi ed
due o size cons ain s. GMM es ima ions use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y.
All a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he a e age annual g ow h
a e o IHDI o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
66
Table 15: Inclusi e economic de elopmen and COFOG componen s o 3-yea pe iods (ex ension)
(1) (2) (3) (4) (5)
Va iables FE Sys-GMM FE Sys-GMM FE
Housing -0.284***
(0.102)
Sa e y 1.092* 0.612*
(0.657) (0.364)
Sa e y2-0.260*** -0.151***
(0.100) (0.047)
Cul u e 1.715**
(0.759)
Cul u e2-0.352*
(0.182)
Social -0.129***
(0.038)
Social20.006**
(0.002)
Obse a ions 203 205 200 207 203
Coun ies 58 62 57 62 58
AR(1) 0.0295 0.0157
AR(2) 0.748 0.859
Hansen es 0.238 0.238
Ins umen s 56 56
Adj. R20.621 0.657 0.631
No es: No es: S a is ically ele an Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. Con ols equal o Table 1, bu omi ed
due o size cons ain s. GMM es ima ions use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y.
All a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he a e age annual g ow h
a e o HDI o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
67
4.2 Robus ness analysis
As sugges ed in Pa 2,Labon e (2010) and Vedde and Gallaway (1998) depic ha go e nmen
expendi u e is a mo e app op ia e me ic o measu e go e nmen size han go e nmen e enue, none he-
less, using he la e is s ill easonable. The e o e, his sec ion depic s he basic esul s using go e nmen
e enue as a pe cen age o GDP as a p oxy o go e nmen size o ensu e ha he basic esul s in Tables
3and 4hold.
Table 16 displays he esul s using 3-yea pe iods, wi h he undamen al esul s co obo a ing he
hypo hesis o condi ional con e gence, implying ha a e con olling o coun y-speci ic cha ac e is ics,
hey end o hei s eady s a es, independen ly o he me ic adop ed as dependen . Al hough a imes less
signi ican , he es o he explana o y a iables s ill hold hei e ec s on economic indica o s. Pa icula ly,
educa ion and ade s ill p esen bene icial impac s on p ospe i y.
Go e nance and g oss capi al o ma ion emain bene icial o GDP and HDI. Addi ionally, he popula ion
g ow h a e and in an mo ali y emain ne a ious, he i s o GDP g ow h and he la e o HDI and IHDI.
As o go e nmen size, he quad a ic e ec o he A mey Cu e is s ill p esen , a leas o HDI, wi h he
espec i e op imum o 43.73% o Sys em-GMM and 36.71% o FE.
In Table 29, explana o y a iables beha e as an icipa ed. Rega ding go e nmen size, he quad a ic
e ec is signi ican o bo h GDP and HDI, wi h op imums o 42.46% and 35.96% o GDP o he la e .
Simila ly o subsec ion 4.1.1, he lack o s a is ical ele ance may no necessa ily indica e he inexis-
ence o he A mey Cu e bu a he a di ec epe cussion o he small sample size. As such, he same
exe cise is p ac ised he e, whe e con ols a e disca ded o augmen sample size, none heless, now using
e enues.
In Table 17, he main in e ences a e clea , pa icula ly ha , in mos eg essions, he quad a ic e ec
o go e nmen size appea s, simila o he A mey Cu e. Respec i e op imums a e 39.78% and 32.24%
o 3-yea Sys em-GMM and FE es ima ions. Fo 5-yea pe iods, only Sys em-GMM p o ides esul s, wi h
an op imal size o 43.69%.
Summa ising, i is clea ha go e nmen e enue p esen s a pa ial loss o s a is ical capabili y han
when go e nmen size is ep esen ed by expendi u e, none heless, he obus empi ical e idence, o a
leas economic g ow h and de elopmen , and less obus ly, o inclusi e economic de elopmen , s ongly
suppo s he exis ence o he A mey Cu e.
68
Table 16: Robus ness check using go e nmen e enue o 3-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −1-3.006*** -5.487***
(0.854) (0.852)
Log HDI −1-8.590*** -9.843***
(3.124) (1.025)
Log IHDI −15.844* -18.326***
(3.141) (2.465)
Educa ion −10.025 0.007 0.014** 0.002 0.013 0.002
(0.019) (0.012) (0.007) (0.002) (0.018) (0.005)
Size ( e enue) 0.458** 0.213** 0.248*** 0.058*** -0.112 0.024
(0.193) (0.106) (0.075) (0.020) (0.095) (0.069)
Size ( e enue)2-0.005 -0.002 -0.003** -0.001*** 0.002 -0.000
(0.003) (0.002) (0.001) (0.000) (0.001) (0.001)
GCF 0.230*** 0.138*** 0.005 0.015*** 0.049 0.019
(0.051) (0.031) (0.014) (0.005) (0.045) (0.014)
Go e nance 0.177 0.890*** 0.358* 0.063 -0.279 0.292
(0.536) (0.309) (0.195) (0.065) (0.354) (0.215)
Popula ion -0.755* -0.432**
(0.403) (0.170)
T ade 0.025** 0.018*
(0.012) (0.010)
In an -0.006 -0.023*** 0.127** -0.062*
(0.025) (0.006) (0.064) (0.033)
Obse a ions 838 834 820 814 355 343
Coun ies 130 126 130 124 112 100
AR(1) 0.000291 0.0216 0.00530
AR(2) 0.610 0.254 0.751
Hansen es 0.366 0.258 0.291
Ins umen s 129 102 38
Adj. R20.480 0.700 0.585
No es: Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea pe iod dummies. GMM es ima ions use obus
s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All a iables a e
ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he
a e age annual g ow h a e o GDP pe capi a o (1) and (2), o HDI o (3) and (4), and o IHDI o (5) and (6). Signi icance
le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
69

Table 17: Al e na i e es ima ions o inclusi e de elopmen using go e nmen e enue
3-yea pe iods 5-yea pe iods
(1) (2) (3) (4)
Va iables Sys-GMM FE Sys-GMM FE
Log IHDI −10.332 -14.913*** 4.616 -13.660***
(2.471) (1.543) (5.186) (1.823)
Size ( e enue) 0.156*** 0.131* 0.141** 0.069*
(0.051) (0.074) (0.056) (0.041)
Size ( e enue)2-0.002* -0.002* -0.002* -0.001
(0.001) (0.001) (0.001) (0.001)
GCF 0.021* 0.048*
(0.011) (0.025)
Go e nance 0.247 -0.448
(0.206) (0.448)
In an 0.050 0.094
(0.051) (0.068)
Obse a ions 461 441 327 338
Coun ies 128 119 121 122
AR(1) 0.000115 0.0334
AR(2) 0.139
Hansen es 0.102 0.276
Ins umen s 59 35
Adj. R20.523 0.385
No es: Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 3-yea o 5-yea pe iod dummies. GMM es ima ions
use obus s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All
a iables a e ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen
a iable is he a e age annual g ow h a e o IHDI o all eg essions. Signi icance le el a which he null hypo hesis is
ejec ed: ***, 1%; **, 5%, and *, 10%.
70
4.3 Policy and esea ch ecommenda ions
The inno a i e ela ions his disse a ion depic s allow us o ponde he cu en si ua ion ela ed o
go e nmen expendi u e wo ldwide. The wo ldwide a e age go e nmen size, a ound 32%, is no much
u he om he oughly 30% p o ided by he GDP, while i is s ill somewha lowe han he 35% o 40%
p o ided by he HDI o IHDI. Conside ing his, he e a e no wo ldwide applicable policy ecommenda ions
in his aspec .
Ra he , he wo ldwide he e ogenei y makes i clea ha ce ain clus e s should ollow ce ain policies.
Fo ins ance, i is clea ha he Wes e n Wo ld, mainly Eu opean coun ies, epo much la ge go e nmen
sizes han he op imum, commonly e en la ge han 50%.
Table 18: Op imal go e nmen size by indica o
Op imal Go e nmen Size
GDP HDI IHDI
27.22% - 32.54% 33.34% - 42.20% 31.23% - 40.69%
Table 19: Ac ual go e nmen size as o 2022 ac oss di e en egions
Ac ual Go e nmen Size as o 2022
De eloped De eloping Eu opean Union
43.20% 28.33% 45.36%
Sub-Saha an A ica Eu ope Wes e n Wo ld
24.48% 43.29% 44.67%
E en hough each coun y should conside hei speci ic cases, he e idence o o e g own go e n-
men s is eal o he Wes e n Wo ld, and as such, i is highly ecommended o such go e nmen s o
s a o conside imming go e nmen al p esence in he economy, gi en ha , independen ly o he me ic
used, whe he i is mo e economic o social, he e does no seem o be much o gain om inc emen al
go e nmen al p esence.
Such policymaking could be a ool o oppose he ecen his o y o de icien g ow h ha has been
aking place in Eu ope, pa icula ly in he Eu opean Union (EU) membe s a es, and b oad e o s o lowe
go e nmen size o oughly 30-40% would be op imal, howe e , aking in o accoun he p i a e sec o and
social esponses, gi en ha i is impo an o gua an ee ha go e nmen cu s do no impose unequal
71
social damage on he mos ulne able laye s o socie y.
Ins i u ing Eu opean ma gins con olling he size o go e nmen in s a e membe s could be a me hod in
which some ine icien cen al in e e ence is eplaced by p i a e sec o ac i i y in an a emp o s imula e
co po a e s eng h and compe i i eness. The esul s, howe e , also s ess ha go e nance s ill akes a
majo ole, and consequen ly, he Eu opean ins i u ions, such as he EU and i s b anches, should con inue
o ocus on u he gua an eeing ins i u ional obus ness and e iciency, al hough ha ecommenda ion
applies o he es o he wo ld.
The si ua ion is he opposi e o de eloping coun ies, whe e he go e nmen size is la gely lowe han
any o he op imums compu ed, whe e i is common, pa icula ly o egions such as sub-Saha an A ica,
o epo alues o only 15% o go e nmen size.
Con a y o he expec a ions, he e idence does no indica e ha manipula ions in go e nmen size
a e only aluable when accompanied by ins i u ional quali y, al hough he la e emains highly ele an o
p ospe i y. In o he wo ds, while puzzling, i appea s ha coun ies se e ely hinde ed by weak ins i u ional
se ings may, in some cases, be be e posi ioned o ex ac g ea e u ili y om adjus men s in go e nmen
p esence.
Consequen ly, a leas heo e ically, coun ies ha lack su icien go e nance esilience o e ec i ely
and signi ican ly p omo e p ospe i y could u ilize de elopmen policies in ol ing adjus men s o he size o
go e nmen o enhance hei o e all economic p ospe i y. This, in u n, could enable coun ies ha could
no ini ially a o d good go e nance o o e come s agna ion and gain momen um.
Rega ding o e all go e nance, he mos signi ican akeaway o policymake s, pa icula ly in e na ional
aid ins i u ions, is he need o shi hei ocus away om me ely unding p eca ious na ions. Ins ead,
hey should p io i ize implemen ing and sha ing knowledge o help s eng hen hei lacking ins i u ions,
eso ing o inancial assis ance only when absolu ely necessa y. S uc u ing de elopmen policies in his
way would enable de eloping coun ies o enhance ins i u ional e iciency in a manne ha os e s g ow h
and de elopmen . A he same ime, in e na ional o e sigh would help ensu e ha he newly es ablished
ins i u ions a e s imula i e a he han ex ac i e.
Such a dis inc app oach could be a me hod by which de eloping coun ies o e come he po e y ap
ha has been sabo aging he con e gence o some egions, pa icula ly sub-Saha an A ica. In gene al, his
p ac ice would ocus on in es ing in eaching a he han jus p o iding and hus enabling independence
a he han dependence.
Addi ionally, o gua an ee p ospe i y, bo h de eloped and de eloping coun ies should ake in o accoun
he dis inc dynamics o ce ain expendi u es and gua an ee ha he inancial needs o public o de and
72
sa e y a e he i s o be sa is ied, gi en hei pi o al ole in he p omo ion o economic and social wel a e.
I is, howe e , impo an o s ess ha no me ic o wel a e is pe ec , consequen ly, i is essen ial
o con inue o elabo a e mo e pe ec ed measu es o ep esen quali y o li e. Pa icula ly, al hough HDI
and IHDI ep esen , a leas in heo y, enhancemen s om GDP, i is clea ha hey also ha e hei
sho comings. The eason ha mos economis s s ill use GDP o cap u e wel a e may indica e ha he
new de elopmen indica o s canno s ill do wha hey we e supposed o do, pa icula ly o eplace GDP
in mos s udies o de elopmen . Fo HDI and IHDI, speci ically, hei bounding om ze o o one may
p esen a disad an age a he han an ad an age, gi en ha i u he s i sel oo much om he nume ic
and bene ic na u e o GDP, which allows i o p o ide mo e accu a e esul s.
Such anges do no allow so much he e ogenei y and, consequen ly, commonly do no ha e as much
s a is ical ele ance as one hopes o . While GDP dis ega ds inequali y, he HDI and IHDI may be ocused
oo much on he social aspec by disconnec ing hemsel es oo much om he in insic economic alidi y
o no ma i e indica o s.
Addi ionally, con olling o inequali y may e en be mo e app op ia e han including heal h and edu-
ca ion in he cons uc ion o HDI and IHDI. The a ionale ocuses on he ac ha he inclusion o heal h
p ospec s and educa ion p ospec s does no allow economic a iabili y, while a he same ime, i is ex-
emely a e o coun ies o ha e good incomes and sa is ac o y income dis ibu ions while lacking in
le els o educa ion and heal h.
Consequen ly, including heal h and educa ion could cause mo e ha m han bene i s, and hus, o
u u e analysis, i could be mo e in e es ing o assess dynamics o a new a iable ha connec s GDP
and income inequali y. The IHDI has a simila cons uc ion o i s income componen , bu un o una ely, i
compu es i o be in he ange om ze o o one.
Inequali y-adjus ed GDP pe capi a
=
Real GDP pe capi a
1 + (
Gini Index
100 )
Fo new dynamics, I ecommend he cons uc ion o a new a iable o GDP pe capi a adjus ed o
inequali y, which allows us o gua an ee ha we ocus on he nume ical na u e o GDP while a he same
ime gua an eeing ha we a e cap u ing ci izens wel a e, penalising an unequal dis ibu ion o economic
powe . A simple compu a ion o GDP pe capi a discoun ed acco ding o he Gini index could be a good
s a ing poin o he cons uc ion o ha a iable.
The s udy o go e nmen size is highly ele an o assess he ele ance o a cen al go e nmen and
i s p esence in socie y and consequen ly add essing how he mechanisms change when a emp ing o
73
Ba o, R. J., Sala-I-Ma in, X., Blancha d, O. J., and Hall, R. E. (1991). Con e gence ac oss s a es and
egions.
B ookings Pape s on Economic Ac i i y
, 1991(1):107–182. h ps://doi.o g/10.2307/
2534639.
Beck, T. and Le ine, R. (2004). S ock ma ke s, banks, and g ow h: Panel e idence.
Jou nal o Banking &
Finance
, 28(3):423–442. h ps://doi.o g/10.1016/S0378-4266(02)00408-9.
Be gh, A. and Hen ekson, M. (2011). Go e nmen size and g ow h: a su ey and in e p e a ion o
he e idence.
Jou nal o Economic Su eys
, 25(5):872–897. h ps://doi.o g/10.1111/j.
1467-6419.2011.00697.x.
Be nanke, B. S. (2000).
Essays on he G ea Dep ession
. P ince on Uni e si y P ess, s u - s uden edi ion
edi ion. h p://www.js o .o g/s able/j.c 7s5dh.
Be y, W. D. and Lowe y, D. (1984). The measu emen o go e nmen size: Implica ions o he s udy
o go e nmen g ow h.
The Jou nal o Poli ics
, 46(4):1193–1206. h ps://doi.o g/10.2307/
2131249.
Bha ga a, A. and Sa gan, J. D. (1983). Es ima ing dynamic andom e ec s models om panel da a
co e ing sho ime pe iods.
Econome ica: Jou nal o he Econome ic Socie y
, pages 1635–1659.
h ps://doi.o g/10.2307/1912110.
Blundell, R. and Bond, S. (1998). Ini ial condi ions and momen es ic ions in dynamic panel da a mod-
els.
Jou nal o econome ics
, 87(1):115–143. h ps://doi.o g/10.1016/S0304-4076(98)
00009-8.
Bold in, M. and Cano a, F. (2001). Inequali y and con e gence in Eu ope’s egions: econside ing
Eu opean egional policies.
Economic Policy
, 16(32):206–253. h ps://doi.o g/10.1111/
1468-0327.00074.
Bond, S. (2002). Dynamic panel da a models: A guide o mic o da a me hods and p ac ice.
Po uguese
Economic Jou nal
.h ps://doi.o g/10.1007/s10258-002-0009-9.
Bond, S., Hoe le , A., and Temple, J. (2001). Gmm es ima ion o empi ical g ow h models.
Cen e o
Economic Policy Resea ch Discussion Pape s
.h ps://ss n.com/abs ac =290522.
Bose, N., Haque, M. E., and Osbo n, D. R. (2007). Public expendi u e and economic g ow h: A dis-
agg ega ed analysis o de eloping coun ies.
The Manches e School
, 75(5):533–556. h ps:
//doi.o g/10.1111/j.1467-9957.2007.01028.x.
80

Cano a, F. and Ma ce , A. (1995). The poo s ay poo : Non-con e gence ac oss coun ies and egions,
discussion pape no. 1265.
Cen e o Economic PolicyResea ch (CEPR), London.
h ps://ss n.
com/abs ac =289497.
Chen, S.-T. and Lee, C.-C. (2005). Go e nmen size and economic g ow h in aiwan: A h eshold eg es-
sion app oach.
Jou nal o Policy Modeling
, 27(9):1051–1066. h ps://doi.o g/10.1016/j.
jpolmod.2005.06.006.
Chiou-Wei, S.-Z., Zhu, Z., and Kuo, Y.-H. (2010). Go e nmen size and economic g ow h: an applica ion
o he smoo h ansi ion eg ession model.
Applied Economics Le e s
, 17(14):1405–1415. h ps:
//doi.o g/10.1080/13504850902984295.
Cho, D. and G aham, S. (1996). The o he side o condi ional con e gence.
Economics Le e s
, 50(2):285–
290. h ps://doi.o g/10.1016/0165-1765(95)00745-8.
Coo ay, A. (2009). Go e nmen expendi u e, go e nance and economic g ow h.
Compa a i e economic
s udies
, 51:401–418. h ps://doi.o g/10.1057/ces.2009.7.
Da , A. A. and Ami Khalkhali, S. (2002). Go e nmen size, ac o accumula ion, and economic g ow h:
e idence om oecd coun ies.
Jou nal o Policy Modeling
, 24(7):679–692. h ps://doi.o g/10.
1016/S0161-8938(02)00163-1.
Da ies, A. (2009). Human de elopmen and he op imal size o go e nmen .
The Jou nal o Socio-
Economics
, 38(2):326–330. h ps://doi.o g/10.1016/j.socec.2008.07.015.
De iş, K. and Qu eshi, Z. (2016). Income dis ibu ion wi hin coun ies: Rising inequali y.
Global Economy and De elopmen . B ookings
.h ps://www.b ookings.edu/a icles/
income-dis ibu ion-wi hin-coun ies- ising-inequali y/.
De a ajan, S., Swa oop, V., and u Zou, H. (1996). The composi ion o public expendi u e and eco-
nomic g ow h.
Jou nal o Mone a y Economics
, 37(2):313–344. h ps://doi.o g/10.1016/
S0304-3932(96)90039-2.
Di Ma eo, L. (2013).
Measu ing go e nmen in he 21s cen u y: An in e na ional o e iew o he size and
e iciency o public spending
. F ase Ins i u e. h ps://www. ase ins i u e.o g/s udies/
measu ing-go e nmen -21s -cen u y.
81
DiPei o, W. R. and Ano uo, E. (2012). Go e nmen size, public deb and eal economic g ow h:
a panel analysis.
Jou nal o Economic S udies
, 39(4):410–419. h ps:doi.o g/10.1108/
01443581211255620.
D yzek, J. and Goodin, R. E. (1986). Risk-sha ing and social jus ice: The mo i a ional ounda ions o he
pos -wa wel a e s a e.
B i ish Jou nal o Poli ical Science
, 16(1):1–34. h ps://doi.o g/10.
1017/S0007123400003781.
Eas e lin, R. A. (2001). Income and happiness: Towa ds a uni ied heo y.
The Economic Jou nal
,
111(473):465–484. h ps://doi.o g/10.1111/1468-0297.00646.
Eas e lin, R. A. (2015).
Happiness and Economic G ow h – The E idence
, pages 283–299. Sp inge
Ne he lands, Do d ech . h ps://doi.o g/10.1007/978-94-017-9178-6_12.
F iedman, M. (2016).
61. Capi alism and F eedom
, pages 344–349. Columbia Uni e si y P ess, New Yo k
Chiches e , Wes Sussex. h ps://doi.o g/10.7312/blau17412-074.
Goel, R. K. and Nelson, M. A. (1998). Co up ion and go e nmen size: A disagg ega ed analysis.
Public
choice
, 97(1):107–120. h ps://doi.o g/10.1023/A:1004900603583.
Goh, H.-H. and Mohd Aznan, M. S. B. (2023). The op imal go e nmen size and economic g ow h: A
compa a i e s udy be ween malaysia and sou h ko ea.
Heliyon
, 9(12):e22834. h ps://doi.o g/
10.1016/j.heliyon.2023.e22834.
Guseh, J. S. (1997). Go e nmen size and economic g ow h in de eloping coun ies: A poli ical-
economy amewo k.
Jou nal o Mac oeconomics
, 19(1):175–192. h ps://doi.o g/10.1016/
S0164-0704(97)00010-4.
Gwa ney, J. D., Lawson, R., and Holcombe, R. G. (1998).
The size and unc ions o go e nmen and
economic g ow h
. Join Economic Commi ee, Washing on. h ps://ohu.people.ysu.edu/
a icles/go _size_economic_g ow h.pd .
Hajamini, M. and Falahi, M. A. (2018). Economic g ow h and go e nmen size in de eloped eu opean
coun ies: A panel h eshold app oach.
Economic Analysis and Policy
, 58:1–13. h ps://doi.
o g/10.1016/j.eap.2017.12.002.
Hausmann, R., Rod ik, D., and Velasco, A. (2008). 32415 G ow h Diagnos ics. In
The Washing on
Consensus Reconside ed: Towa ds a New Global Go e nance
. Ox o d Uni e si y P ess. h ps://
doi.o g/10.1093/acp o :oso/9780199534081.003.0015.
82
Hibbs, D. A. (1977). Poli ical pa ies and mac oeconomic policy.
Ame ican Poli ical Science Re iew
,
71(4):1467–1487. h ps://doi.o g/10.2307/1961490.
Hicks, D. A. (1997). The inequali y-adjus ed human de elopmen index: A cons uc i e p oposal.
Wo ld
De elopmen
, 25(8):1283–1298. h ps://doi.o g/10.1016/S0305-750X(97)00034-X.
Hicks, J. R. (1948).
Value and Capi al (2nd ed.)
. Ox o d Uni e si y P ess. h ps://global.oup.
com/academic/p oduc / alue-and-capi al-9780198282693?cc=p &lang=en&.
Kaiha su, S., Koga, M., Saka a, T., Ha a, N., e al. (2019). In e ac ion be ween business cycles and
economic g ow h.
Mone a y and Economic S udies
, 37(1):99–126. h ps://www.imes.boj.o .
jp/ esea ch/pape s/english/me37-6.pd .
Kai ila, V. (2004). In eg a ion and condi ional con e gence in he enla ged eu a ea. ETLA Discussion
Pape s 935, ETLA Discussion Pape s, Helsinki. h ps://hdl.handle.ne /10419/63931.
Ka as, G. (1996). The op imal go e nmen size: Fu he in e na ional e idence on he p oduc i i y
o go e nmen se ices.
Economic Inqui y
, 34(2):193–203. h ps://doi.o g/10.1111/j.
1465-7295.1996. b01372.x.
Ka as, G. (1997). On he op imal go e nmen size in eu ope: Theo y and empi ical e idence.
The
Manches e School
, 65(3):280–294. h ps://doi.o g/10.1111/1467-9957.00057.
Keynes, J. M. (1923).
A ac on mone a y e o m
. Cosimo Classics. h ps://hdl.handle.ne /
2027/mdp.35112104085974.
Kim, D.-H., Wu, Y.-C., and Lin, S.-C. (2018). He e ogenei y in he e ec s o go e nmen size and go e -
nance on economic g ow h.
Economic Modelling
, 68:205–216. h ps://doi.o g/10.1016/j.
econmod.2017.07.014.
Kuzne s, S. (1941).
Na ional income
. Na ional Bu eau o Economic Resea ch New Yo k. h p://www.
nbe .o g/books/kuzn41-2.
Labon e, M. (2010). The size and ole o go e nmen : Economic issues, june 14, cong essional esea ch
se ice 7-5700, l32162. h ps://hdl.handle.ne /1813/77640.
Lae en, M. L., Igan, M. D., Claessens, M. S., and Dell’A iccia, M. G. (2010). Lessons and policy im-
plica ions om he global inancial c isis.
IMF Wo king Pape s
.h ps://doi.o g/10.5089/
9781451963021.001.
83
Landau, D. (1983). Go e nmen expendi u e and economic g ow h: A c oss-coun y s udy.
Sou he n
Economic Jou nal
, 49(3):783–792. h ps://doi.o g/10.2307/1058716.
Lin, S. A. (1994). Go e nmen spending and economic g ow h.
Applied Economics
, 26(1):83–94. h ps:
//doi.o g/10.1080/00036849400000064.
Loayza, N. V. and Ranciè e, R. (2006). Financial de elopmen , inancial agili y, and g ow h.
Jou nal o
Money, C edi and Banking
, 38(4):1051–1076. h p://www.js o .o g/s able/3838993.
Mahle , V. A. (1992). Measu ing public sec o size in he ad anced ma ke economy coun ies: The
p oblem o de la o s.
Social Indica o s Resea ch
, 27:311–325. h ps://doi.o g/10.1007/
BF00303852.
Makin, A. J., Pea ce, J., and Ra nasi i, S. (2019). The op imal size o go e nmen in aus alia.
Economic
Analysis and Policy
, 62:27–36. h ps://doi.o g/10.1016/j.eap.2018.12.001.
Ma ins, S. and Veiga, F. J. (2014). Go e nmen size, composi ion o public expendi u e, and economic
de elopmen .
In e na ional Tax and Public Finance
, 21:578–597. h ps://doi.o g/10.1007/
s10797-014-9313-4.
Ma hu , S. K. (2005). Economic g ow h & condi ional con e gence: I s speed o selec ed egions o 1961-
2001.
Indian Economic Re iew
, 40(2):185–208. h p://www.js o .o g/s able/29793843.
McKinley, T. (2010). Inclusi e g ow h c i e ia and indica o s: An inclusi e g ow h index o diagno-
sis o coun y p og ess. Technical epo , Asian De elopmen Bank. h ps://www.adb.o g/
publica ions.
Muelle , D. C. (2003).
Public Choice III
. Camb idge Uni e si y P ess, 3 edi ion. h ps://doi.o g/
10.1017/CBO9780511813771.
Mu hy, N. V. and Chien, I. (1997). The empi ics o economic g ow h o oecd coun ies: Some new
indings.
Economics Le e s
, 55(3):425–429. h ps://doi.o g/10.1016/S0165-1765(97)
00092-X.
Naha , S. and Inde , B. (2002). Tes ing con e gence in economic g ow h o oecd coun ies.
Applied
Economics
, 34(16):2011–2022. h ps://doi.o g/10.1080/00036840110117837.
84
Nell, K. S. (2020). E alua ing he condi ional con e gence hypo hesis in he pos -1989 globaliza ion pe-
iod.
Applied Economics
, 52(30):3308–3326. h ps://doi.o g/10.1080/00036846.2019.
1710451.
Ni ola, N. and Sahu, S. (2019). The in e ac i e impac o go e nmen size and quali y o ins i u ions on
economic g ow h- e idence om he s a es o india.
Heliyon
, 5(3):e01352. h ps://doi.o g/10.
1016/j.heliyon.2019.e01352.
Nonneman, W. and Vanhoud , P. (1996). A Fu he Augmen a ion o he Solow Model and he Empi ics
o Economic G ow h o OECD Coun ies*.
The Qua e ly Jou nal o Economics
, 111(3):943–953.
h ps://doi.o g/10.2307/2946677.
No dhaus, W. D. and Tobin, J. (1972).
Economic Resea ch: Re ospec and P ospec , Volume 5, Economic
G ow h
. NBER. h p://www.nbe .o g/chap e s/c7620.
No h, D. C. and Thomas, R. P. (1973).
The Rise o he Wes e n Wo ld: A New Economic His o y
. Camb idge
Uni e si y P ess. h ps://doi.o g/10.1017/CBO9780511819438.
OECD (2024).
Socie y a a Glance 2024: OECD Social Indica o s
. OECD Publishing, Pa is. h ps:
//doi.o g/10.1787/918d8db3-en.
Peacock, A. and Sco , A. (2000). The cu ious a ac ion o wagne ’s law.
Public choice
, 102(1):1–17.
h ps://doi.o g/10.1023/A:1005032817804.
Pede sen, T. M. and Elme , A. M. (2003). In e na ional e idence on he connec ion be ween business
cycles and economic g ow h.
Jou nal o Mac oeconomics
, 25(2):255–275. h ps://doi.o g/
10.1016/S0164-0704(03)00028-4.
Quah, D. (1993). Gal on’s allacy and es s o he con e gence hypo hesis.
The Scandina ian Jou nal o
Economics
, 95(4):427–443. h ps://doi.o g/10.2307/3440905.
Quah, D. T. (1997). Empi ics o g ow h and dis ibu ion: S a i ica ion, pola iza ion, and con e gence
clubs.
Jou nal o economic g ow h
, 2:27–59. h ps://doi.o g/10.1023/A:1009781613339.
Ram, R. (1986). Go e nmen size and economic g ow h: A new amewo k and some e idence om
c oss-sec ion and ime-se ies da a.
The Ame ican Economic Re iew
, 76(1):191–203. h p://www.
js o .o g/s able/1804136.
85

Rauniya , G. and Kanbu , R. (2010). Inclusi e g ow h and inclusi e de elopmen : a e iew and syn hesis
o asian de elopmen bank li e a u e.
Jou nal o he Asia Paci ic Economy
, 15(4):455–469. h ps:
//doi.o g/10.1080/13547860.2010.517680.
Rome , P. M. (1986). Inc easing e u ns and long- un g ow h.
Jou nal o Poli ical Economy
, 94(5):1002–
1037. h ps://doi.o g/10.1086/261420.
Sachs, J. D., McA hu , J. W., Schmid -T aub, G., K uk, M., Bahadu , C., Faye, M., and McCo d, G. (2004).
Ending a ica’s po e y ap.
B ookings Pape s on Economic Ac i i y
, 2004(1):117–216. h p://
www.js o .o g/s able/3217964.
Saga , A. D. and Najam, A. (1998). The human de elopmen index: a c i ical e iew.
Ecological Economics
,
25(3):249–264. h ps://doi.o g/10.1016/S0921-8009(97)00168-7.
Samuelson, P. A. (1961).
The E alua ion o ‘Social Income’: Capi al Fo ma ion and Weal h
, pages 32–57.
Palg a e Macmillan UK, London. h ps://doi.o g/10.1007/978-1-349-08452-4_3.
Schula ick, M. and Solomou, S. (2011). Ta i s and economic g ow h in he i s e a o globaliza ion.
Jou nal o Economic G ow h
, 16:33–70. h ps://doi.o g/10.1007/s10887-011-9061-6.
Se h, S. (2009). Inequali y, in e ac ions, and human de elopmen .
Jou nal o Human De elopmen and
Capabili ies
, 10(3):375–396. h ps://doi.o g/10.1080/19452820903048878.
Shea e , R. A. (1961). The concep o economic g ow h.
Kyklos
, 14(4):497–532. h ps://doi.o g/
10.1111/j.1467-6435.1961. b00368.x.
Smi h, A. (1776).
An Inqui y in o he Na u e and Causes o he Weal h o Na ions
. His o y o Economic
Though Books. McMas e Uni e si y A chi e o he His o y o Economic Though . h ps://doi.
o g/10.7208/chicago/9780226763750.001.0001.
Solow, R. M. (1956). A con ibu ion o he heo y o economic g ow h.
The Qua e ly Jou nal o Economics
,
70(1):65–94. h ps://doi.o g/10.2307/1884513.
S ee en, P. (1994). Human de elopmen : Means and ends.
The Ame ican Economic Re iew
, 84(2):232–
237. h p://www.js o .o g/s able/2117835.
S u n, S. and Eps ein, G. (2021). How much should we us i e-yea a e aging o pu ge business cycle e -
ec s? a eassessmen o he inance-g ow h and capi al accumula ion-unemploymen nexus.
Economic
Modelling
, 96:242–256. h ps://doi.o g/10.1016/j.econmod.2020.12.028.
86
Tanninen, H. (1999). Income inequali y, go e nmen expendi u es and g ow h.
Applied Economics
,
31(9):1109–1117. h ps://doi.o g/10.1080/000368499323599.
Tanzi, V. and Schuknech , L. (1997). Reconside ing he iscal ole o go e nmen : The in e na ional pe -
spec i e.
The Ame ican Economic Re iew
, 87(2):164–168. h ps://www.js o .o g/s able/
2950906.
Te asawa, K. L. and Ga es, W. R. (1998). Rela ionships be ween go e nmen size and economic g ow h:
Japan’s go e nmen e o ms and e idence om oecd.
In e na ional Public Managemen Jou nal
,
1(2):195–223. h ps://doi.o g/10.1016/S1096-7494(99)80091-9.
Ullah, S., Akh a , P., and Zae a ian, G. (2018). Dealing wi h endogenei y bias: The gene alized me hod o
momen s (gmm) o panel da a.
Indus ial Ma ke ing Managemen
, 71:69–78. h ps://doi.o g/
10.1016/j.indma man.2017.11.010.
an den Be gh, J. C. (2009). The gdp pa adox.
Jou nal o Economic Psychology
, 30(2):117–135. h ps:
//doi.o g/10.1016/j.joep.2008.12.001.
Vedde , R. K. and Gallaway, L. E. (1998).
Go e nmen Size and Economic G ow h
. The Commi ee.
h ps://hdl.handle.ne /2027/osu.32435069311363.
Whajah, J., Bokpin, G. A., and Ku u, S. (2019). Go e nmen size, public deb and inclusi e g ow h in a ica.
Resea ch in In e na ional Business and Finance
, 49:225–240. h ps://doi.o g/10.1016/j.
iba .2019.03.008.
Wol owi z, P. (2006). Good go e nance and de elopmen : a ime o ac ion, jaka a, indonesia. Rema ks
by P esiden , Paul Wol owi z (English). P esiden ial speech Washing on, D.C. : Wo ld Bank G oup.
h p://documen s.wo ldbank.o g/cu a ed/en/938181467987875500.
Wo ld Bank (2022). Wo ld de elopmen epo 2022: Finance o an equi able eco e y ( ol. 1
o 2) (english). Technical epo , Wo ld Bank G oup. Wo ld De elopmen Indica o s | Wo ld
De elopmen Repo , Washing on, D.C. h p://documen s.wo ldbank.o g/cu a ed/en/
408661644986413472.
87
A Annex
0
7.5
15
Loga i hm o GDP pe capi a
3 4 5
Loga i hm o HDI
-100
0
100
GDP pe capi a g ow h a e (%)
-25 0 25
HDI g ow h a e (%)
Obse a ions 95% Con idence in e al P edic ed alues
Figu e 12: T end lines o GDP pe capi a and HDI.
Sou ce:
UNDP/WB.
5
10
15
Loga i hm o GDP pe capi a
2.5 4 5.5
Loga i hm o IHDI
-100
0
100
GDP pe capi a g ow h a e (%)
-20 0 20
IHDI g ow h a e (%)
Obse a ions 95% Con idence in e al P edic ed alues
Figu e 13: T end lines o GDP pe capi a and IHDI.
Sou ce:
UNDP/WB.
88
3
4
5
Loga i hm o HDI
2.5 4 5.5
Loga i hm o IHDI
-20
0
20
HDI g ow h a e (%)
-20 0 20
IHDI g ow h a e (%)
Obse a ions 95% Con idence in e al P edic ed alues
Figu e 14: T end lines o HDI and IHDI.
Sou ce:
UNDP.
0
5
10
15
Loga i hm o GDP pe capi a
0 25 50 75 100
Go e nmen expendi u e (% o GDP)
-75
0
75
GDP pe capi a g ow h a e (%)
0 25 50 75 100
Go e nmen expendi u e (% o GDP)
Obse a ions 95% Con idence in e al P edic ed alues
Figu e 15: T end line o go e nmen size and he GDP pe capi a g ow h a e.
Sou ce:
WB/IMF.
89
Table 23: Consump ion and in es men o 5-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −10.194 -5.871***
(1.773) (1.029)
Log GDP −2-1.768
(1.658)
Log HDI −16.116 -11.042***
(5.128) (1.392)
Log HDI −2-6.492***
(2.104)
Log IHDI −1-3.655 -18.388***
(5.805) (4.374)
Educa ion −1-0.013 0.002 -0.006 -0.001 0.003 -0.008
(0.019) (0.013) (0.013) (0.002) (0.028) (0.008)
Consump ion 0.440 0.271*** 0.281** 0.024 0.384** 0.037
(0.317) (0.092) (0.123) (0.018) (0.183) (0.076)
Consump ion2-0.007 -0.005*** -0.004*** -0.000 -0.006** -0.000
(0.005) (0.001) (0.002) (0.000) (0.003) (0.001)
In es men -1.466** 0.082 -0.303* 0.072** 0.162 -0.245
(0.678) (0.175) (0.167) (0.030) (0.336) (0.293)
In es men 20.104** -0.012 0.037** -0.004* -0.022 0.027
(0.053) (0.010) (0.017) (0.002) (0.023) (0.031)
GCF 0.147** 0.144*** 0.014 0.007 0.060 0.049***
(0.069) (0.035) (0.024) (0.006) (0.048) (0.011)
Go e nance -0.014 1.230*** 0.424** 0.003 0.187 0.401
(0.533) (0.335) (0.201) (0.071) (0.538) (0.260)
Popula ion -1.747** -0.621***
(0.737) (0.207)
T ade 0.016* 0.015*
(0.009) (0.009)
In an 0.059 -0.034*** -0.020 -0.095
(0.049) (0.008) (0.064) (0.063)
Obse a ions 451 443 432 435 165 128
Coun ies 122 114 121 112 101 64
AR(1) 0.0428 0.00621
AR(2) 0.215 0.815
Hansen es 0.331 0.116 0.952
Ins umen s 34 31 20
Adj. R20.614 0.796 0.708
No es: See Table 4
96

Table 24: Economic g ow h quan ile eg ession
(1) (2) (3) (4) (5)
3-yea pe iods τ= 0.1τ= 0.25 τ= 0.5τ= 0.75 τ= 0.9
Log GDP −10.823* 0.829*** 0.836*** 0.842*** 0.846***
(0.445) (0.304) (0.142) (0.038) (0.100)
Educa ion −10.000 0.000 0.000 0.000 0.000
(0.006) (0.004) (0.002) (0.001) (0.001)
Size 0.016 0.013 0.009 0.006 0.003
(0.056) (0.039) (0.018) (0.005) (0.013)
Size2-0.000 -0.000 -0.000 -0.000 -0.000
(0.001) (0.001) (0.000) (0.000) (0.000)
GCF 0.003 0.004 0.004 0.005*** 0.005
(0.015) (0.010) (0.005) (0.001) (0.003)
Go e nance 0.030 0.026 0.022 0.018 0.016
(0.160) (0.109) (0.051) (0.014) (0.036)
Popula ion -0.004 -0.008 -0.014 -0.018* -0.021
(0.108) (0.074) (0.035) (0.009) (0.024)
T ade 0.000 0.000 0.001 0.001* 0.001
(0.004) (0.003) (0.001) (0.000) (0.001)
Obse a ions 838 838 838 838 838
(1) (2) (3) (4) (5)
5-yea pe iods τ= 0.1τ= 0.25 τ= 0.5τ= 0.75 τ= 0.9
Log GDP −10.687*** 0.693*** 0.703*** 0.713*** 0.718***
(0.128) (0.094) (0.052) (0.047) (0.066)
Educa ion −10.001 0.001 0.001 0.001 0.001
(0.002) (0.001) (0.001) (0.001) (0.001)
Size 0.015 0.014 0.014** 0.014** 0.014
(0.016) (0.012) (0.007) (0.006) (0.008)
Size2-0.000 -0.000 -0.000** -0.000** -0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
GCF 0.006 0.006* 0.006*** 0.007*** 0.007***
(0.005) (0.003) (0.002) (0.002) (0.002)
Go e nance 0.042 0.040 0.037* 0.033* 0.031
(0.047) (0.034) (0.019) (0.017) (0.024)
Popula ion -0.012 -0.012 -0.012 -0.012 -0.012
(0.030) (0.022) (0.012) (0.011) (0.016)
T ade 0.000 0.001 0.001 0.001* 0.001
(0.001) (0.001) (0.001) (0.000) (0.001)
Obse a ions 573 573 573 573 573
No es: Qua ile eg essions using he quan iles o 0.1, 0.25, 0.5, 0.75 and 0.9. The dependen a iable is he loga i hm o
GDP pe capi a o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
97
Table 25: Inclusi e economic de elopmen quan ile eg ession
(1) (2) (3) (4) (5)
3-yea pe iods τ= 0.1τ= 0.25 τ= 0.5τ= 0.75 τ= 0.9
Log IHDI −10.523*** 0.499*** 0.460*** 0.417*** 0.391***
(0.162) (0.125) (0.083) (0.099) (0.134)
Educa ion −10.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
Size -0.002 -0.002 -0.002 -0.002 -0.002
(0.003) (0.003) (0.002) (0.002) (0.003)
Size20.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
GCF 0.001 0.001 0.001 0.000 0.000
(0.001) (0.001) (0.000) (0.000) (0.001)
Go e nance 0.009 0.009 0.009 0.010 0.010
(0.012) (0.009) (0.006) (0.007) (0.010)
In an -0.002 -0.002 -0.002* -0.002 -0.002
(0.002) (0.002) (0.001) (0.001) (0.002)
Obse a ions 355 355 355 355 355
(1) (2) (3) (4) (5)
5-yea pe iods τ= 0.1τ= 0.25 τ= 0.5τ= 0.75 τ= 0.9
Log IHDI −10.098 0.106 0.121 0.136 0.142
(0.210) (0.164) (0.102) (0.127) (0.157)
Educa ion −10.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
Size 0.002 0.002 0.001 -0.001 -0.001
(0.005) (0.004) (0.002) (0.003) (0.004)
Size2-0.000 -0.000 -0.000 -0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000)
GCF 0.001 0.001 0.001 0.001 0.001
(0.001) (0.001) (0.001) (0.001) (0.001)
Go e nance 0.026 0.022* 0.015* 0.009 0.006
(0.016) (0.012) (0.008) (0.010) (0.012)
In an -0.003 -0.003 -0.002 -0.002 -0.002
(0.003) (0.002) (0.001) (0.002) (0.002)
Obse a ions 263 263 263 263 263
No es: Qua ile eg essions using he quan iles o 0.1, 0.25, 0.5, 0.75 and 0.9. The dependen a iable is he loga i hm o
IHDI o all eg essions. Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
98
Table 26: Economic g ow h and COFOG componen s o 5-yea pe iods
(1) (2) (3) (4) (5)
Va iables Sys-GMM FE FE Sys-GMM FE
De ense 1.912***
(0.619)
De ense2-0.254***
(0.073)
Public -1.507***
(0.416)
Public20.074***
(0.024)
Housing -3.138**
(1.403)
Housing20.881**
(0.334)
Sa e y 3.101** 3.621***
(1.578) (1.226)
Sa e y2-0.715*** -0.900***
(0.242) (0.152)
Obse a ions 297 294 294 297 290
Coun ies 69 63 63 69 62
AR(1) 0.0110 0.00179
AR(2) 0.818 0.449
Hansen es 0.428 0.458
Ins umen s 60 70
Adj. R20.503 0.479 0.578
No es: Equal app oach as Table 12 bu using 5-yea pe iod dummies.
99
Table 27: Economic de elopmen and COFOG componen s o 5-yea pe iods
(1) (2) (3) (4) (5) (6) (7)
Va iables FE FE FE FE Sys-GMM Sys-GMM FE
De ense -0.160***
(0.047)
Economic 0.174***
(0.059)
Economic2-0.015***
(0.004)
Go .Educa ion 0.419**
(0.166)
Go .Educa ion2-0.041***
(0.014)
Public -0.031*
(0.017)
Heal h -0.444**
(0.211)
Heal h20.037**
(0.016)
Housing -0.841*
(0.432)
Housing20.274**
(0.112)
Sa e y 0.572***
(0.150)
Sa e y2-0.132***
(0.022)
Obse a ions 282 286 286 286 291 291 282
Coun ies 61 62 62 62 67 67 61
AR(1) 0.0218 0.0282
AR(2) 0.530 0.402
Hansen es 0.208 0.130
Ins umen s 56 56
Adj. R20.637 0.592 0.582 0.583 0.651
No es: Equal app oach as Table 13 bu using 5-yea pe iod dummies.
100
Table 28: Inclusi e economic de elopmen and COFOG componen s o 5-yea pe iods
(1) (2) (3) (4) (5)
Va iables FE Sys-GMM FE Sys-GMM FE
De ense 0.332*
(0.188)
De ense2-0.038***
(0.012)
Go .Educa ion 2.243*
(1.249)
Go .Educa ion2-0.256**
(0.128)
Heal h 0.106*
(0.063)
Housing -3.135**
(1.487)
Housing20.991*
(0.567)
Sa e y 0.867**
(0.428)
Sa e y2-0.173***
(0.044)
Obse a ions 146 154 147 154 146
Coun ies 54 61 54 61 54
AR(1) 0.0884 0.0673
AR(2)
Hansen es 0.611 0.518
Ins umen s 30 17
Adj. R20.543 0.421 0.536
No es: Equal app oach as Table 15 bu using 5-yea pe iod dummies.
101

Table 29: Robus ness check using go e nmen e enue o 5-yea pe iods
(1) (2) (3) (4) (5) (6)
Va iables Sys-GMM FE Sys-GMM FE Sys-GMM FE
Log GDP −1-2.317*** -5.686***
(0.896) (0.749)
Log HDI −1-5.405*** -10.807***
(1.790) (1.098)
Log IHDI −10.956 -18.701***
(4.287) (3.018)
Educa ion −10.044 0.027*** 0.013*** 0.004* -0.004 0.013**
(0.029) (0.010) (0.005) (0.002) (0.013) (0.005)
Size ( e enue) 0.292* 0.242** 0.108*** 0.057*** 0.081 0.057
(0.168) (0.101) (0.040) (0.019) (0.073) (0.053)
Size ( e enue)2-0.004 -0.003* -0.001 -0.001*** -0.000 -0.001
(0.003) (0.002) (0.001) (0.000) (0.001) (0.001)
GCF 0.177*** 0.148*** 0.031** 0.012** 0.113** 0.004
(0.045) (0.034) (0.012) (0.005) (0.046) (0.023)
Go e nance 0.288 0.869*** 0.157 0.031 -0.179 0.357
(0.499) (0.266) (0.106) (0.050) (0.304) (0.305)
Popula ion -0.468 -0.207
(0.324) (0.246)
T ade 0.008 0.015
(0.008) (0.009)
In an 0.001 -0.029*** 0.039 -0.058
(0.013) (0.007) (0.065) (0.036)
Obse a ions 573 565 563 553 263 247
Coun ies 129 121 129 119 109 93
AR(1) 0.000844 0.000367 0.0573
AR(2) 0.106 0.837
Hansen es 0.294 0.379 0.759
Ins umen s 67 81 41
Adj. R20.539 0.717 0.429
No es: Two-s ep Sys em-GMM and Fixed e ec s eg essions, including 5-yea pe iod dummies. GMM es ima ions use obus
s anda d e o s. FE es ima ions use s anda d e o s clus e ed by coun y and obus o he e oskedas ici y. All a iables a e
ea ed as endogenous in Sys em-GMM es ima ions wi h he excep ion o pe iod dummies. The dependen a iable is he
a e age annual g ow h a e o GDP pe capi a o (1) and (2), o HDI o (3) and (4), and o IHDI o (5) and (6). Signi icance
le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
102
Table 30: Go e nmen size, g ow h, de elopmen , and inclusi e de elopmen o annual pe iods
(1) (2) (3)
Va iables FE FE FE
Log GDP −1-5.332***
(1.006)
Log HDI −1-9.907***
(1.336)
Log IHDI −1-22.120***
(2.964)
Educa ion −10.005 0.002 0.005
(0.011) (0.003) (0.007)
Size 0.226** 0.040* 0.001
(0.105) (0.021) (0.058)
Size2-0.005*** -0.001* -0.000
(0.001) (0.000) (0.001)
GCF 0.155*** 0.017*** 0.034***
(0.024) (0.005) (0.012)
Go e nance 0.979*** 0.037 0.329
(0.307) (0.091) (0.233)
Popula ion -0.345**
(0.133)
T ade 0.027***
(0.008)
In an -0.017** -0.059*
(0.008) (0.032)
Obse a ions 2,286 2,245 1,076
Cou ies 129 128 110
Adj. R20.471 0.485 0.309
Fixed e ec s eg essions, including yea dummies. FE es ima ions use s anda d e o s clus e ed by coun y and obus o
he e oskedas ici y. The dependen a iable is he annual g ow h a e o GDP pe capi a o (1), o HDI o (2), and o IHDI
o (3). Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
103
Table 31: De elopmen dynamics o annual pe iods
(1) (2) (3)
Va iables FE FE FE
Log GDP −1-5.546***
(0.948)
Log HDI −1-9.975***
(1.346)
Log IHDI −1-22.136***
(2.899)
Educa ion −10.011 0.003 0.004
(0.010) (0.003) (0.007)
De eloping·Size 0.452*** 0.077*** 0.090
(0.146) (0.026) (0.062)
De eloping·Size2-0.009*** -0.001*** -0.002**
(0.002) (0.000) (0.001)
De eloped·Size -0.501*** -0.022 -0.069
(0.188) (0.032) (0.054)
De eloped·Size20.003* 0.000 0.001
(0.002) (0.000) (0.001)
GCF 0.153*** 0.018*** 0.034***
(0.025) (0.005) (0.013)
Go e nance 0.984*** 0.038 0.323
(0.307) (0.091) (0.228)
Popula ion -0.379***
(0.117)
T ade 0.024***
(0.007)
In an -0.017** -0.059*
(0.008) (0.032)
Obse a ions 2,286 2,245 1,076
Coun ies 129 128 110
Adj. R20.478 0.488 0.316
Fixed e ec s eg essions, including yea dummies. FE es ima ions use s anda d e o s clus e ed by coun y and obus o
he e oskedas ici y. The dependen a iable is he annual g ow h a e o GDP pe capi a o (1), o HDI o (2), and o IHDI
o (3). Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
104
Table 32: Income dynamics o annual pe iods
(1) (2) (3)
Va iables FE FE FE
Log GDP −1-5.598***
(0.908)
Log HDI −1-10.101***
(1.389)
Log IHDI −1-22.406***
(2.911)
Educa ion −10.011 0.003 0.004
(0.010) (0.003) (0.006)
Low.I·Size 0.319 0.043 -0.262**
(0.340) (0.058) (0.116)
Low.I·Size2-0.006 -0.000 0.004**
(0.006) (0.001) (0.002)
L.M.I·Size 0.250 0.072 0.279
(0.164) (0.050) (0.174)
L.M.I·Size2-0.004* -0.001 -0.004
(0.002) (0.001) (0.003)
U.M.I·Size 0.652*** 0.066* 0.002
(0.211) (0.039) (0.076)
U.M.I·Size2-0.012*** -0.001** -0.001*
(0.003) (0.001) (0.001)
High.I·Size -0.386*** -0.000 -0.113**
(0.140) (0.033) (0.049)
High.I·Size20.002 0.000 0.001**
(0.001) (0.000) (0.001)
GCF 0.151*** 0.016*** 0.032**
(0.025) (0.005) (0.013)
Go e nance 0.975*** 0.057 0.356
(0.308) (0.093) (0.228)
Popula ion -0.411***
(0.105)
T ade 0.024***
(0.007)
In an -0.017** -0.057*
(0.008) (0.032)
Obse a ions 2,286 2,240 1,076
Coun ies 129 127 110
Adj. R20.480 0.492 0.325
Fixed e ec s eg essions, including yea dummies. FE es ima ions use s anda d e o s clus e ed by coun y and obus o
he e oskedas ici y. The dependen a iable is he annual g ow h a e o GDP pe capi a o (1), o HDI o (2), and o IHDI
o (3). Signi icance le el a which he null hypo hesis is ejec ed: ***, 1%; **, 5%, and *, 10%.
105