Economic G ow h and Inequali y: The
Role o Fiscal Policies
Leonel Muinelo,
O iol Roca-Sagalés
11.05
De
p
a amen d'Economia A
p
licada
Facul a d'Economia i Emp esa
Aques documen pe any al Depa amen d'Economia Aplicada.
Da a de publicació :
Depa amen d'Economia Aplicada
Edi ici B
Campus de Bella e a
08193 Bella e a
Telè on: (93) 581 1680
Fax:(93) 581 2292
E-mail: [email p o ec ed]
h p://www.ecap.uab.es
Feb e 2011
Economic G ow h and Inequali y: The Role o Fiscal Policies
Leonel Muinelo and O iol Roca-Sagalés
Depa amen d’Economia Aplicada, Uni e si a Au ònoma de Ba celona, Bella e a, 08193 Spain
ABSTRACT
This pape analyses he impac o di e en ins umen s o iscal policy on economic
g ow h as well as on income inequali y, using an unbalanced panel o 43 uppe -middle and
high income coun ies o he pe iod 1972-2006. We conside and es ima e wo indi idual
equa ions explaining g ow h and inequali y in o de o assess he incidence o di e en iscal
policies. Fi s ly, ou app oach conside s imposing o hogonal assump ions be ween g ow h
and inequali y in bo h equa ions, and secondly, i allows g ow h o be included in he
inequali y equa ion, and inequali y o be included in he g ow h equa ion. The empi ical
esul s sugges ha an inc ease in he size o go e nmen measu ed h ough cu en
expendi u es and di ec axes diminishes economic g ow h while educing inequali y, , being
public in es men he only iscal policy ha may b eak his ade-o be ween e iciency and
equi y, since inc eases in his i em educes inequali y wi hou ha ming ou pu . The e o e, he
esul s e lec ha he ade-o be ween e iciency and equi y ha go e nmen s o en con on
when designing hei iscal policies may be a oided.
Keywo ds: iscal policy, inequali y, g ow h, panel da a models
JEL classi ica ion numbe s: E62, D31, O47, C23
1. In oduc ion
This pape in es iga es o wha ex en , and by means o which componen s, iscal policy
has an impac on economic ac i i y and income inequali y in a sample o 43 uppe -middle and
high income coun ies.
The in e ac ions be ween economic g ow h and income inequali y ha e a ac ed a g ea
deal o a en ion in ecen yea s. While ea lie wo ks sugges ed a nega i e ade-o be ween
g ow h and inequali y in he i s s ages o de elopmen , mo e ecen s udies sugges
mechanisms by which inequali y is indeed inc eased by economic g ow h o by which income
inequali y a ec s g ow h (posi i ely o nega i ely).
Seminal s udies by Lewis (1954), Kuzne s (1955) and Kaldo (1956) sugges ed ha
income inequali y is mos ly de e mined by he le el o economic de elopmen . Mo e
p ecisely, hey analysed how economic de elopmen a ec s income dis ibu ion in he long-
un sugges ing a po en ial inc easing e ec o g ow h on income inequali y in he i s s ages
o economic de elopmen , and a dec easing e ec in he la e s ages (“in e ed-U
hypo hesis”).1 Mo e ecen s udies ha e pu o wa d new ideas abou he e ec s o economic
de elopmen on income inequali y. These wo ks ha e aken in o accoun h ee phenomena:
he apid g ow h o in e na ional ade (Wood and Ridao-Cano, 1996); he inc eased di usion
o new echnologies in di e en p oduc i e ac i i ies (Eiche , 1996, Galo and Tsiddon, 1997,
Aghion e al., 1999, and Hassle and Rod íguez-Mo a, 2000), and he eme gence o new
o ganiza ional o ms (Ca oli, 1999).
In con as , in he ex ensi e li e a u e on de elopmen ha has appea ed du ing he
1990s, he causa ion be ween inequali y and g ow h uns in he opposi e di ec ion. In ac , he
cen al conce n mainly ocuses on he ole o income and weal h inequali y in he p ocess o
economic g ow h. Two main g oups o s udies can be iden i ied in his heo e ical li e a u e.
1 Fo heo e ical s udies o his “in e ed-U hypo hesis”, see Robinson (1976), G eenwood and Jo ano ic (1990), and
Helpam (1998). In u n, Fields (2001) o e s a comple e su ey o he empi ical li e a u e.
2
One g oup sugges s a ious ansmission channels h ough which g ea e ini ial inequali y
os e s economic g ow h.2 The o he g oup sugges s se e al economic and poli ical channels
h ough which ini ial inequali y migh be ha m ul o g ow h.3
The di e si y o heo e ical app oaches ha ocus on he mac oeconomic e ec s o
iscal policies; he sho age o empi ical con ibu ions examining hei impac s o an
ex ended se o coun ies, and he sca ci y o wo ks ha ela e he g ow h and associa ed
dis ibu i e e ec s o di e en public policies, poin s o he need o an analysis measu ing
bo h e ec s. In his con ex , ou pape does he a o emen ioned and connec s hese wo
s ands o li e a u e.
To p o ide a comp ehensi e analysis o he g ow h and dis ibu i e e ec s o di e en
iscal ins umen s, we conside and es ima e sepa a ely equa ions o g ow h and inequali y
using an unbalanced panel o 43 uppe -middle and high income coun ies o he pe iod 1972-
2006. We begin by conside ing independen models o g ow h and inequali y (o hogonal
equa ions) ha inco po a e a limi ed se o con ol a iables commonly ound in he li e a u e,
and e alua e, independen ly, he impac s o di e en ins umen s o iscal policy on bo h
mac o-agg ega es. Acco ding o he abo e-men ioned li e a u e, he e a e solid economic
a gumen s o belie e ha income inequali y and economic g ow h de e mine each o he .
Consequen ly, ou empi ical s a egy also conside s he inclusion o income inequali y in he
g ow h equa ion and GDP g ow h in he inequali y equa ion (s uc u al equa ions). This
makes i possible o analyse he ela ionship be ween bo h agg ega es and o in es iga e he
ole o iscal policy, which has adi ionally been conside ed as an e ec i e ins umen o
gene a ing e enue and o edis ibu i e pu poses. The con ibu ion o his pape is hus
wo old. Fi s ly, we look simul aneously a GDP g ow h, income inequali y and iscal policies
2 See Rebelo (1991) and Dea on and Paxson (1997), among o he s.
3 Fo s udies ha use pu ely economic easons see, Aghion and Bol on (1992 and 1997), Galo and Zei a (1993), Pike y
(1997), Galo and Zan (1997), Mu phy, Shlei e and Vishny (1998), and Mookhe jee and Ray (2003). Fo s udies ha analyse
he in luence o poli ical channels see Gup a (1990), Be ola (1993), Sain -Paul and Ve die (1993), Alesina and Rod ik
(1994), and Bénabou (1996).
3

in an ex ended panel o coun ies. Secondly, we pe o m he analysis o a a ie y o
disagg ega ed iscal measu es, bo h in e ms o expense and e enue.
The esul s ob ained show ha iscal policies ha e signi ican e ec s on g ow h and
inequali y. Highe di ec axes and cu en expendi u es con ac economic g ow h while, a
he same ime, educe economic inequali y. These esul s clea ly e lec he ade-o be ween
e iciency and equi y ha go e nmen s con on when designing hei iscal policies:
inc easing he size o he go e nmen diminishes economic g ow h, al hough i achie es a
signi ican imp o emen in he equali y o incomes. The only iscal policy ha may b eak his
ade-o is public in es men since, acco ding o he esul s ob ained; inc eases in his i em
educe inequali y wi hou ha ming ou pu .
This pape is o ganised as ollows. Sec ion 2 p o ides he heo e ical amewo k, whe e
di e en hypo heses conce ning he in luence o iscal policies on economic g ow h and
inequali y a e discussed. Sec ion 3 de ails he da a base and discusses he empi ical
me hodology. In Sec ion 4 he empi ical esul s a e p esen ed, while in Sec ion 5 we es hei
obus ness o di e en assump ions. Finally, sec ion 6 con ains ou conclusions.
2. Theo e ical amewo k
In his sec ion, we i s p esen he heo e ical p io s unde lying he empi ical g ow h
equa ions, in pa icula hose ela ed o he ole o iscal policies in economic ac i i y;
secondly, we p esen hose ha allow he impac o iscal policies on economic inequali y o
be es ima ed; and las ly, we explain he iscal a iables conside ed in bo h models.
2.1 Fiscal policy and economic g ow h
The mac oeconomic analysis dis inguishes basically wo gene al heo e ical app oaches
when analysing he capaci y o iscal policy o a ec economic ac i i y. On he one hand,
om a neoclassical app oach, se e al models emphasise he sho - e m e ec s o di e en
ins umen s o iscal policy. In his app oach, he s eady-s a e g ow h is d i en by exogenous
4
ac o s, such as he dynamics o popula ion and he echnological p og ess. Thus, he
con en ional wisdom has been ha di e ences in ax and expendi u e policies can be
impo an de e minan s o he le el o ou pu , bu a e unlikely o ha e a signi ican pe manen
e ec on he economic g ow h a e.4
The public-policy neoclassical g ow h models con as wi h he p edic ions o he
endogenous g ow h models, whe e g ow h is no conduc ed by exogenous ac o s. In hese
models, in es men in human and physical capi al does a ec he s eady-s a e g ow h a e
and, consequen ly, he e is much mo e scope o ax and go e nmen expendi u e o play a
ole in he g ow h p ocess. These wo ks end o ans o m he empo a y g ow h e ec s o
iscal policy ha he neoclassical model in ol es, in o pe manen e ec s. Thus, endogenous
g ow h models ha inco po a e public policies p edic ha dis o ing axes, as well as
p oduc i e public expendi u es, a ec economic g ow h. I ollows ha iscal policy can
a ec he le el o ou pu as well as i s long- e m g ow h a e.5
In line wi h hese endogenous app oaches, ou benchma k equa ion o economic g ow h
is based on he models de eloped by Ba o (1990) and Ba o and Sala-i-Ma in (1992).
Addi ionally, and in o de o a oid he biases associa ed wi h an incomple e speci ica ion o
he go e nmen budge cons ain , we ollow Knelle e al’s (1999) s a egy conce ning he
inclusion o iscal a iables.
We conside an economy o n p oduce s, each one p oducing one p oduc (y), acco ding
o he p oduc ion unc ion:

gAky 
1 (1)
4 Sa o (1967), K zyzaniak (1967) and Felds ein (1974) use he neoclassical model o analyse he e ec s o di e en axes on
g ow h; Chamely (1986) and Judd (1985) use he model de eloped by Cass and Koopmans (1965) o s udy he e ec s o
iscal policy conside ing endogenous sa ing a es; Summe s (1981) and Aue bach and Ko liko (1987) adap he model o
o e lapping gene a ions o Diamond (1965) o analyse he dynamic e ec s o iscal policy.
5 Since he pionee ing con ibu ions o Ba o (1990), King and Rebelo (1990), and Lucas (1990), se e al pape s ha e
ex ended he analysis o axa ion, public expendi u e and g ow h. See, o example, Ga cía-Peñalosa and Tu no sky (2007).
5
whe e ep esen s p i a e capi al and
k
g
he public capi al used by he p oduce (wha we
conside he p oduc i e public expendi u e).
The go e nmen balances i s budge in each pe iod by aising a p opo ional ax on
ou pu a a e

(dis o iona y ax) and lump-sum axes (non-dis o iona y axes).
The e o e, he budge cons ain ha he go e nmen aces can be exp essed as:
L
nyLbCng


 (2)
whe e and g a e he non-p oduc i e and p oduc i e public expendi u e, espec i ely.
Because we allow o he case o an unbalanced budge , we include a a iable b ha collec s
he budge a y su plus (de ici ).
C
The economic g ow h a e o he coun y i du ing pe iod , i
y,

is a unc ion o a se o
non- iscal a iables , and a ec o o iscal a iables :
i
X, i
FP,
i
m
j jm
h
i i i uFPXy ,
1,
1,,   

(3)
Assuming ha ec o F
P
includes all he ele an elemen s, we deduce ha :
0
1,


m
j j
FP (4)
In o de o a oid pe ec collinea i y in he es ima ion o equa ion (3) we exclude one
elemen o ec o . The omi ed a iable is e ec i ely he assumed compensa ing elemen
wi hin he go e nmen ’s budge cons ain (Knelle e al., 1999). Thus, conside ing ha he
g ow h equa ion can be exp essed as:
FP
i
m
j mm jj
h
i i i uFPFPXy ,
1
1,,
1,,  



(5)
we omi he elemen o ob ain he new g ow h equa ion:
m
FP ,
(6)

i
m
j jmj
h
i i i uFPXy ,
1,
1,,   

6
Acco ding o his s a egy, he in e p e a ion o he es ima ed coe icien o each iscal
a iable is he e ec o a uni a y change in he ele an a iable (included in he eg ession)
o se by a uni a y change in he omi ed iscal a iable, which is he implici inancial
elemen . The in e p e a ion o he es ima ed coe icien s o he non-omi ed iscal a iables
a ies i he omi ed ca ego y is al e ed.
In o de o educe he speci ica ion e o bias, we conside wo g ow h o hogonal
equa ions con aining di e en se s o con ol a iables. Model 1 conside s a se o con ol
a iables based on he Solow g ow h model ha includes he ini ial le el o GDP pe capi a,
p i a e in es men ,6 and popula ion g ow h. Based on Mankiw-Rome -Weil (1992), and in
o de o con ol o he impac o human capi al accumula ion, Model 2 includes he o me
Solow se and inco po a es he a e age yea s o schooling o he popula ion aged 25 and
abo e.
P e ious s udies do no ake in o accoun inequali y when calcula ing he e ec s o
iscal policy on ou pu . As a gued abo e, we also conside he inclusion o an inequali y
measu e in he g ow h equa ions allowing o a join analysis o he mac oeconomic and
dis ibu i e e ec s o iscal policy, which cons i u es a no el ea u e o ou s udy. Thus,
Model 3 and 4 (s uc u al equa ions) expand ou benchma k g ow h equa ions wi h he
addi ion o an inequali y a iable (Gini index) in Model 1 and 2, espec i ely.
2.2 Fiscal policy and income edis ibu ion
In con as wi h he abundan heo e ical li e a u e ela ing iscal policy and economic
g ow h, con ibu ions abou he e ec s o such policies on income inequali y ha e been e y
sca ce un il ecen ly.
Fo economic inequali y, ou benchma k equa ion is based on he empi ical app oaches
o Li and Zou (1998), Li, Squi e and Zou (1998), Cas elló and Doménech (2002) and
6 The p i a e in es men a iable was specially cons uc ed o his s udy as a di e ence o o al in es men ( om Penn
Wo ld Table 6.1) and public in es men ( om Go e nmen Finance S a is ics o IMF), as a sha e o GDP.
7
Ba o (1990 and 2008) and Cas elló-Climen (2010), can be in e p e ed in se e al ways.
Fi s ly, i is possible ha pa o ha go e nmen spending is somehow was e ul. Secondly, i
is also possible ha go e nmen spending is jus a p oxy o he en i e se o go e nmen non-
p ice in e en ions (like employmen legisla ion, heal h and sa e y ules and p oduc
s anda ds), and, i may be ha hese non-p ice in e en ions a e esponsible o educing
g ow h, and no he le el o expendi u e.
Public in es men is posi i e and wi h a smalle coe icien han p i a e in es men bu
i is no s a is ically signi ican , ega dless o whe he i is inanced by di ec o indi ec axes.
Conside ing ha public in es men is cons i u ed mainly by in es men in in as uc u es, one
would expec ha i in luences he a e o economic g ow h bo h posi i ely and di ec ly.
Acco ding o his a gumen , ou esul would be somehow su p ising; howe e , as Romp and
Haan (2007) and Kamps (2005) summa ise, his posi i e bu non-signi ican impac could be
explained because he e ec o public in es men may di e conside ably ac oss he coun ies
included in ou sample.
A signi ican nega i e e ec is ound in he case o di ec axes. This esul , which is
also ob ained by Knelle e al. (1999), is consis en wi h economic heo y because o he
dis o ing e ec s o his ype o axes. In con as , indi ec axes do no ha e a signi ican
impac on g ow h.
An impo an addi ional esul de i ed om he s uc u al equa ions is ha income
inequali y, measu ed by he Gini coe icien , ha ms economic g ow h.15 This esul is
consis en wi h he ea ly 1990s empi ical g ow h li e a u e based on a c oss-coun y app oach
(see Pe sson and Tabellini (1994) and Pe o i (1994 and 1996)), and also wi h mo e ecen
s udies ha use a panel da a app oach (Huang e al., 2009). Howe e , he nega i e e ec o
inequali y on g ow h is con adic o y wi h he con en ional ex book indica ing ha inequali y
15 To educe any inconsis ency esul ing om he ac ha some Gini coe icien s a e based on income, whe eas a ew a e
based on expendi u e; in his s uc u al equa ion we ollow Deninige and Squi e’s sugges ion and add 6.6 o gini coe icien s
based on expendi u e.
14

is good o incen i es and he e o e good o g ow h,16 and also di e s om he empi ical
s udies o Pa idge (1997), Li and Zou (1998), Fo bes (2000), Ba o (2000), Lundbe g and
Squi e (2003), and Cas elló-Climen (2010). Despi e he ac ha hese la e pape s also use a
panel da a app oach, i is impo an o unde line some impo an di e ences wi h espec o
ou wo k he e. Fi s ly, hey conside ini ial inequali y as an explana o y a iable while we use
a con inuous Gini a iable cons uc ed aking in o conside a ion disposable income o
expendi u e; secondly, hey use di e en es ima ion me hods; and hi dly, he se s o con ol
a iables, coun ies and pe iod conside ed a e also di e en .
Gi en ha in his empi ical model we a e conside ing iscal policy a iables, we hink
ha he signi ican and nega i e impac o inequali y on g ow h canno be explained only by
he adi ional a gumen s based on he poli ical economy li e a u e (see Alesina and Rod ik
1994; and Benabou, 1996; among o he s). Addi ionally, gi en ha we also con ol o
in es men (in human and physical capi al), his esul would indica e ha income inequali y
has a di ec nega i e e ec on g ow h no coming om i s e ec on in es men decisions.17 In
a con ex o uppe -middle and high income economies, and conside ing ha capi al ma ke s
a e impe ec and he agen s a e he e ogeneous, one possible explana ion could a ise o
explain he es ima ed nega i e e ec s o inequali y on g ow h. Inequali y is de imen al o
g ow h, as bo owe s end o unde -in es in e o when i is unobse able; when he e a e
incen i e p oblems, he mo e unequal he income dis ibu ion is, he lowe he agg ega e le el
o e o will be (see, o example, Aghion and Bol on, 1997).
4.2. Dis ibu ional e ec s o iscal policy
The inequali y equa ions allow he dis ibu i e e ec s o iscal policies o be analysed.
In Table 3 we p esen he main esul s ob ained wi h he o hogonal and s uc u al equa ions
16 The adi ional isions p opose a posi i e e ec o inequali y on g ow h due o di e en a es o sa ing o ich and poo
indi iduals (Kaldo , 1956 and 1957), and due o incen i e e ec s (Rebelo, 1991).
17 Fo wo ks ha analyse he ole o in es men o explain a nega i e impac o inequali y on g ow h, see Bane jee and
Newman (1993); and Pike y (1997).
15
o inequali y de ailed in sec ion 2.2. In each model, he i s column shows he esul s
conside ing indi ec axes as he implici inancing elemen , while he second column uses
di ec axes. We only epo he es ima es o he ele an iscal a iables (o he e enues, ne
lending, and he su plus/de ici a iables a e nei he s a is ically no economically
signi ican ).
The esul s o he o hogonal and s uc u al equa ions a e e y simila and do no change
in e ms o signi icance, sign and magni ude o he con ol a iables. In bo h ypes o
equa ions, he con ol a iables a e signi ican and wi h he expec ed sign, which basically
coincides wi h he esul s o Li and Zou (1998), and Li e al. (1998). Thus, inc eases in ini ial
ci il libe ies index educe income inequali y while an inc ease in ini ial educa ional
inequali y inc eases income inequali y. I is ema kable ha educa ional inequali y, as a p oxy
o asse s inequali y, has a much g ea e in luence han hose associa ed wi h poli ical
economy a gumen s, as is also poin ed ou by Li e al. (1998). I is also impo an o
emphasise ha he dummy a iable ha con ols o he di e ences caused by he di e en
sou ce o he Gini indices, which is inco po a ed in Model 6 and 8, is signi ican and inc eases
he explana o y powe o bo h ypes o equa ions (o hogonal and s uc u al).
Conce ning he iscal a iables, i is no ewo hy ha cu en public expendi u e has a
signi ican and sizeable nega i e e ec on income inequali y; i educes income inequali y
ega dless o whe he i is inanced by di ec o indi ec axes. This esul is consis en wi h
Buli and Gulde (1995), Gus a sson and Johansson (1999), Li e al. (2000), Galli and an de
Hoe en (2001), and Al onso e al. (2010). One would expec ha his ype o expendi u e
educes income inequali y because i includes di e en social expenses wi h dis ibu i e
implica ions h ough he immedia e bene i s, o example, expenses in ans e s like pensions
o di e en subsidies.
16
The e ec o public in es men on inequali y is sligh and no s a is ically signi ican in
he o hogonal equa ions. Howe e , he e ec o his a iable is nega i e and s a is ically
signi ican in he s uc u al eg essions. This la e esul showing a educ ion in inequali y is
in line wi h he a gumen s o B ennenman and Ke (2002) and wi h he empi ical esul s
ob ained wi h a sample o La in-Ame ican coun ies by Calde ón and Se én (2004).
Concep ually, he de elopmen o public in as uc u es helps unde de eloped a eas o he
economy o be connec ed o he co es o economic ac i i y, allowing access o addi ional
p oduc i e oppo uni ies. Ano he a gumen along hese lines is ha in as uc u es also
imp o e access o help and educa ion se ices.
The e ec o di ec axes on inequali y is nega i e and signi ican ; ne e heless, he
es ima ed coe icien s a e much smalle han hose co esponding o cu en expendi u e. This
esul is in line wi h Chu, Da oodi and Gup a (2000) o de eloping coun ies!!! (wha we
do? quo o no ?). This nega i e impac may e lec he p og essi e s uc u e o he ax
sys ems o he analysed coun ies, many o hem wi h a mode n iscal sys em. Wi h a
p og essi e ax sys em, inc eases in di ec ax e enue – whe he h ough inc eases in he ax
base, in he o e all a e age ax a e o in he p og ession o he ax s uc u e – would yield a
la ge edis ibu i e e ec and hus lowe inequali y (Lambe , 2001). Finally, indi ec axes
ha e a posi i e coe icien bu no signi ican .
Las ly, he esul s ob ained wi h he s uc u al equa ions clea ly show ha economic
g ow h has a s a is ically signi ican and nega i e e ec on inequali y. Conside ing ha ou
sample con ains a g oup o high-income and uppe -middle-income coun ies, his esul is in
line wi h a Kuzne s- ype ela ionship.
5. Sensi i i y Analysis
In his sec ion we es he obus ness o ou main esul s by modi ying some aspec s o
he es ima ed g ow h and inequali y equa ions. In bo h cases, we begin by es ing i he
17
coe icien s o he iscal a iables a e sensi i e o he inclusion o new con ol a iables in
bo h equa ions.
Addi ionally, we use ins umen al a iables o examine he possibili y o simul anei y
be ween iscal a iables, g ow h and inequali y. We deal wi h endogenei y in bo h equa ions
by using he i e-yea lagged alues o ou endogenous explana o y a iables as ins umen s.
The use o i e-yea lagged alues as ins umen s is jus i ied o h ee easons (Mu ay,
2006). Fi s ly, i is unlikely ha economic g ow h and inequali y will oday a ec pas alues
o ou iscal policy a iables. Secondly, he lagged alues o hese a iables a e co ela ed
wi h he alues wi hou lags. Las ly, he only impac o hese lagged alues on economic
g ow h and inequali y mus pass h ough he endogenous a iables. This is sugges ed by he
ac ha including he explana o y endogenous a iables and hei i e-yea lagged alues in
he same eg ession yields no s a is ically signi ican e ec o he la e .
Finally, in o de o a hom whe he he esul s a e being d i en by one pa icula coun y
in ou sample, we epea he eg essions o g ow h and inequali y a e emo ing each o he
coun ies one a ime. The esul s, in bo h equa ions, a e s able indica ing ha no single
coun y is d i ing ou esul s.
5.1 Sensi i i y Analysis I: The g ow h equa ions.
In i s ins ance, we conduc a s epwise eg ession analysis by adding o he a iables
discussed in he g ow h li e a u e on Model 2 and 4 ( hose ha also include a human capi al
a iable). The wo ks o Ba o (1990), Mendoza e al. (1997) and Lundbe g and Squi e
(2003), p o ide he new a iables conside ed in his sensi i i y analysis which a e ade,
in la ion, inancial de elopmen , and a measu e o ci il libe ies. We epo he esul s
including only he a iables ha a e signi ican ( ade and in la ion).18 Columns 1 o 4 o
Table 4 epo he esul s o his sensi i i y exe cise.
18 The signi icance and signs o he ele an iscal a iables emain unchanged including all new con ol a iables.
18
Un o una ely he e a e well known p oblems wi h es ima ing g ow h eg essions. The
igh -hand-side a iables a e ypically endogenous,19 and measu ed wi h e o . Ano he
di icul y is ha o omi ed a iables. One a iable ha should be included in a condi ional
con e gence eg ession, he ini ial le el o e iciency, is no obse ed. One way o add ess
hese p oblems has been h ough a i s -di e ence gene alised me hod o momen s es ima o s
applied o dynamic panel da a models. The de elopmen s in dynamic panel da a models (e.g.,
A ellano and Bond, 1991; A ellano and Bo e , 1995; K ie , 1995) ha e ocused mainly on
hose applica ions o mic o da a se s, which gene ally ha e a la ge c oss-sec ion dimension
bu a limi ed ime se ies dimension (la ge N, small T). These p ope ies also ypically ma ch
he dimensions o ou da a based on i e-yea a e age (N a ound 43, T a ound 7) well. So, in
second ins ance, we es ima e a dynamic panel da a model, and, apply a wo-s ep GMM
es ima o .20 The second ou columns o able 4 epo he esul s o his sensi i i y exe cise.
Th oughou his sensi i i y analysis, h ee main esul s eme ge clea ly. Fi s ly, he
inclusion o new a iables is no esponsible o he s ong iscal e ec s iden i ied ea lie ; he
signi icance o he iscal a iables is no sensi i e o he inclusion o hese mac o a iables.
Secondly, wo o he new a iables conside ed a e signi ican and wi h he expec ed sign
showing ha an inc ease in in e na ional ade aises economic g ow h while an inc ease in
in la ion educes i ; bo h esul s we e also ound by Mendoza e al. (1997), Ba o (1990), and
Cas elló-Climen (2010); espec i ely. Finally, i we look a he esul s in Table 4 o he
dynamic models wi h IV, we obse e ha he coe icien s, signs and signi icance o inequali y
and all he ele an iscal a iables emain unchanged indica ing ha he e ec s iden i ied
ea lie a e no simply he esul o endogenei y. The e o e, he in e p e a ion o he ole o key
19 Ob ious candida es in his case a e p i a e in es men and iscal a iables.
20 In s udying economic g ow h, his p ocedu e has impo an ad an ages o e simple c oss-sec ion eg essions and o he
me hods o dynamic panel da a models, such as he one-s ep GMM p ocedu e. Fi s ly, es ima es will no longe be biased by
any omi ed a iables ha a e cons an o e he ime (“ ixed” e ec s). In condi ional con e gence eg essions, his a oids he
p oblem aised by he omission o ini ial e iciency. Secondly, he use o ins umen al a iables allows pa ame e s o be
es ima ed consis en ly in models which include endogenous igh -hand-side a iables such as p i a e in es men and iscal
a iables. Finally, he use o ins umen s po en ially allows consis en es ima ion e en in he p esence o measu emen e o .
19

iscal a iables on g ow h is subs an ially una ec ed: inc easing he size o go e nmen
( h ough cu en expendi u e and di ec axes) educes economic g ow h while an inc ease in
inequali y educes economic g ow h.
5.2 Sensi i i y Analysis II: The inequali y equa ions
In he case o he inequali y equa ions, we p oceed wi h a simila s a egy o ha
employed in he g ow h equa ions. In i s ins ance, we conduc a s epwise eg ession analysis
by adding o he con ol a iables discussed in he inequali y li e a u e. Thus, he selec ion o
he addi ional a iables o include (in la ion, p i a e in es men , ade and ini ial GDP), is
based on he con ibu ions o de Li, Squi e and Zou (1998) and Li and Zou (2002).
In u n, while he da a panel on g ow h and i s de e minan s is big enough o in oduce
dynamics o he model, he da a panel on income inequali y is a he limi ed ( o some
coun ies he e a e only h ee a ailable obse a ions on an ex ended sample o se en i e-yea
pe iods). In addi ion, he mos common app oach o es ima e dynamic panel da a models is
he i s -di e ence Gene alised Me hod o Momen s (GMM) es ima o . This es ima o akes
i s di e ences in o de o elimina e he sou ce o inconsis ency and uses he le els o he
lagged explana o y a iables as ins umen s. Howe e , by aking i s di e ences, mos o he
a ia ion o he inequali y da a, which comes basically om a iabili y ac oss coun ies,
disappea s. Hampe ed by hese econome ic di icul ies and da a limi a ions, he es ima ion o
a dynamic panel da a equa ion wi h ins umen s is no a iable op ion. Since ou main conce n
is endogenei y, which is cons an ly aised in income dis ibu ion li e a u e (see o example,
Li and Zou, 1998), we apply he ins umen al a iables me hod (IV) ia wo-s age leas
squa es o co ec o he endogenei y o he iscal a iables.
Table 5 epo s he esul s o his wo sensi i i y analysis applied o he inequali y
equa ions. In columns 1 o 4, we show he esul s adding he ade a iable o he o hogonal
and s uc u al inequali y equa ions co esponding o Model 6 and 8 ( ade is he only
20
addi ional a iable which u ns ou o be signi ican a leas in some o he new eg essions).21
In columns 5 o 8, we p esen he ins umen al a iables (IV) es ima ion esul s o bo h
inequali y equa ions when he i e-yea lagged alues o he independen a iables a e used as
a se o ins umen s, since he e is usually no co ela ion be ween he dis u bance and he
lagged alues (see Iimi, 2005; and, Enikolopo and Zhu a skaya, 2007).
Two main esul s eme ge clea ly om Table 5. Fi s ly, h oughou he sensi i i y
analysis, public cu en expendi u e and di ec axes emain signi ican and he es ima ed
coe icien s a e simila o hose o he o iginal eg essions on Table 3; he e o e hese esul s
appea o be qui e obus o he inclusion o new a iables and o ins umen al a iables (IV)
es ima ion. Secondly, we also obse e ha public in es men has a signi ican impac on
inequali y. Hence, we con i m he esul ob ained in he benchma k model using he s uc u al
equa ions in he sense ha public in es men educes he le els o economic inequali y.
Taking in o accoun ha his in es men is cons i u ed mainly by in as uc u es, one would
expec ha i educes income inequali y indi ec ly, as explained be o e.
6. Conclusions
Recen imes ha e seen go e nmen spending, axa ion, and de ici inancing mo e o
he o e on o policy analysis. Fiscal policy a ec s agg ega e demand, he dis ibu ion o
weal h, and he economy’s capaci y o p oduce goods and se ices. Howe e , he majo i y o
exis ing empi ical s udies ha e ocused on he e ec s o iscal policy on economic ac i i y
wi hou conside ing he edis ibu i e e ec s and, no o e ing, in u n, an analysis o he
impac o di e en iscal policy ins umen s.
In his pape , we conside an unbalanced panel o 43 uppe -middle and high income
coun ies o he pe iod 1972-2006 o show ha bo h he magni ude and he composi ion o
he iscal policy ha e signi ican impac s no only on economic g ow h bu also on economic
21 The inclusion o addi ional no signi ican a iables did no change he signi icance and sign o iscal a iables.
21
inequali y. The e o e, di e en iscal policies could be used o a ec bo h g ow h and income
dis ibu ion.
Rega ding he mac oeconomic e ec s o iscal policies, ou empi ical esul s a e in
ag eemen wi h hose ound in o he empi ical s udies and sugges ha cu s o di ec axes
inc ease GDP, whe eas inc eases o public cu en expendi u es diminish i . Beyond ha , and
unlike o he empi ical wo k, ou esul s also show ha di e en iscal policies ha e
signi ican edis ibu i e e ec s: an inc ease o public expendi u e (cu en o in public
in es men ) p oduces signi ican educ ions in income inequali y, as does inc easing di ec
axes. Mo eo e , he es ima ion o s uc u al equa ions, whe e a ela ion o mu ual in luence
be ween g ow h and inequali y is allowed o , shows ha a educ ion in income inequali y
s imula es economic g ow h, whe eas he p ocess o economic g ow h educes economic
inequali y. Consequen ly, hese esul s sugges ha some p e iously omi ed cha ac e is ics o
g ow h a e ela ed o inequali y, and ice e sa; he e o e we could a gue ha he o hogonal
equa ions we e p obably su e ing om omi ed a iables bias. In any case, he esul s o bo h
ypes o equa ions conside ed a e e y consis en showing ha di e en iscal policies ha e
signi ican g ow h and dis ibu i e e ec s in he analysed economies. On he o he hand,
sensi i i y analyses indica e ha ou mac oeconomic esul s a e obus o he inclusion o
o he con ol a iables and also o di e en es ima ion echniques conside ing endogenei y
p oblems.
F om a policy pe spec i e, ou esul s ha e clea implica ions. Acco ding o ou
es ima es, inc easing he size o he public sec o ( h ough cu en expendi u es and di ec
axes) imp o es he dis ibu ion o income a he expense o economic g ow h. The e ec s o
indi ec axes on bo h ou pu and inequali y a e ound o be s a is ically insigni ican . The
only iscal policy ha may b eak his ade-o be ween e iciency and equi y is public
in es men , since inc eases in his i em educes inequali y wi hou ha ming ou pu . The la e
22
esul s poin on he same di ec ion o Ga cía-Peñalosa and Tu no sky (2007) p o iding
empi ical suppo o he possibili y o iscal policies inc easing he size o go e nmen -
h oughou indi ec axes o inance public in es men - ha inc ease economic g ow h
educing simul aneously inequali y.
Acknowledgemen s: We would like o hank And eas Ky iacou and Hèc o Sala o aluable
commen s and sugges ions. The au ho s acknowledge inancial suppo o p ojec s, SEJ2007-
67911-C03-02/ECON (Minis e io de Ciencia y Tecnología), 2009SGR-2013 and XREPP
(Di ecció Gene al de Rece ca).
23
30
Tables
Table 1 – Theo e ical agg ega ion o iscal policy
Theo e ical classi ica ion Go e nmen Finance S a is ics classi ica ion
Di ec axes  Taxes on income, p o i s, and capi al gains
 Taxes on pay oll and wo k o ce
 Taxes on p ope y
Indi ec axes  Taxes on goods and se ices
 Taxes on in e na ional ade and ansac ions
O he e enues  O he axes
 G an s
 O he e enue
Cu en public expendi u e  Compensa ion o employees
 Use o goods and se ices
 Consump ion o ixed capi al
 In e es
 Subsidies
 G an s
 Social bene i s
 O he expense
Public In es men  Ne acquisi ion o non inancial asse s
T ansac ions in inancial asse s and liabili ies  Ne acquisi ion o inancial asse s
 Ne incu ence o liabili ies
Go e nmen Su plus/De ici  To al e enues minus o al ou lays
No es: The classi ica ion is based on he manual GFS 2001 and co esponds o he Gene al Go e nmen .

Table 2 – Economic G ow h. Dependen a iable: Real pe capi a GDP g ow h
ORTHOGONAL EQUATIONS STRUCTURAL EQUATIONS
MODEL 1 MODEL 2 MODEL 3 MODEL 4
Omi ed Fiscal Va iable: Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes
Ini ial GDP p.c. -0.0100*** -0.0103*** -0.0079*** -0.0078*** -0.0086*** -0.0102*** -0.0095*** -0.0106***
Inequali y -- -- -- -- -0.0128*** -0.0099*** -0.0109** -0.0091**
P i a e in es men 0.0053*** 0.0044*** 0.0046*** 0.0032*** 0.0073** 0.0065*** 0.0061*** 0.0046*
Popula ion g ow h -0.0004* -0.0020 -0.0014*** -0.0012*** -0.0016*** -0.0012*** -0.0018*** -0.0016***
Human capi al -- -- 0.0044* 0.0036** -- -- 0.0038** 0.0029
Public in es men 0.0018* 0.0017 0.0015 0.0011 0.0019 0.0020 0.0021 0.0022
Cu en public expendi u e -0.0053*** -0.0090*** -0.0077*** -0.0096*** -0.0073*** -0.0103*** -0.0085*** -0.0110***
Di ec axes -0.0039*** Omi ed -0.0042*** Omi ed -0.0051*** Omi ed -0.0042** Omi ed
Indi ec axes Omi ed 0.0022* Omi ed 0.0001 Omi ed 0.0009 Omi ed 0.0003
Adjus ed 2
R
0.60 0.59 0.57 0.54 0.60 0.55 0.58 0.53
No. o obse a ions 149 146 130 128 117 114 104 102
No. o coun ies 43 43 35 35 41 41 33 33
No es: Es ima ion echnique: 5-yea s a e ages, wo-way ixed e ec s. All explana o y a iables a e exp essed in logs excep popula ion g ow h.
* Signi ican a 10% le el; ** Signi ican a 5%, *** Signi ican a 1%
Table 3 – Income Inequali y. Dependen a iable: log o Gini Index
ORTHOGONAL EQUATIONS STRUCTURAL EQUATIONS
MODEL 5 MODEL 6 MODEL 7 MODEL 8
Omi ed Fiscal Va iable: Indi ec
Taxes Di ec
Taxes Indi ec
Taxes Di ec Taxes Indi ec
Taxes Di ec Taxes Indi ec
Taxes Di ec Taxes
Ci il libe ies 0.0427** 0.0510** 0.0484** 0.0558*** 0.0673*** 0.0562** 0.0756*** 0.0564**
Educa ion inequali y 0.1056** 0.1080** 0.1026** 0.1042** 0.0812* 0.1035* 0.0784 0.1014*
G ow h -- -- -- -- -0.017* -0.0139 -0.0162* -0.0128
Public in es men -0.0411 -0.0634** -0.0525* -0.0692** -0.0450 -0.0680* -0.0604** -0.0778**
Cu en public expendi u e -0.1974*** -0.2948*** -0.2555*** -0.3320*** -0.2318*** -0.3113*** -0.3262*** -0.3811***
Di ec axes -0.0857*** Omi ed -0.0702** Omi ed -0.0680* Omi ed -0.0385* Omi ed
Indi ec axes Omi ed 0.0510 Omi ed 0.0410 Omi ed 0.0672 Omi ed 0.1175
Disposable income dummy -- -- 0.0780* 0.0827* -- -- 0.0935** 0.096**
Adjus ed R2 0.56 0.56 0.62 0.62 0.57 0.54 0.65 0.63
No. o obse a ions 110 109 110 109 101 100 101 100
No. o coun ies 35 35 35 35 34 34 34 34
No es: Es ima ion echnique: 5-yea s a e ages, andom e ec s model wi h empo al dummies. All he a iables a e exp essed in logs.
* Signi ican a 10% le el; ** Signi ican a 5%, *** Signi ican a 1%
31
Table 4 – Sensi i i y Analysis I. Economic G ow h – Dependen a iable: Real pe capi al GDP g ow h
Mo e con ol a iables on Model 2 and 4 Dynamics on Model 2 and 4 wi h IV
O hogonal equa ions S uc u al equa ions O hogonal equa ions S uc u al equa ions
Omi ed Fiscal Va iable: Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes
Ini ial GDP p.c. -0.0072*** -0.0077*** -0.0078** -0.0085** -- -- -- --
(GDP p.c. g ow h -1) -- -- -- -- -0.3548*** -0.5133*** -0.3552*** -0.4968***
Inequali y -- -- -0.0066 -0.0046 -- -- -0.0070* -0.0111**
P i a e in es men 0.0032** 0.0022 0.0042 0.0025 0.0050 0.0073* 0.0030 -0.0036
Popula ion g ow h -0.0015*** -0.0012*** -0.0019*** -0.0014*** -0.0014*** -0.0006 -0.0010 -0.0003
In la ion -0.0018*** -0.0018*** -0.0006* -0.0012*** -- -- -- --
In e na ional ade 0.0057*** 0.0066*** 0.0046* 0.0062* -- -- -- --
Human capi al 0.0040* 0.0030 0.0027 0.0011 0.0082** 0.0094*** 0.0040 0.0049*
Public in es men 0.0014 0.0011 0.0015 0.0013 0.0015 0.0033 -0.0026 0.0010
Cu en public expendi u e -0.0097*** -0.0121*** -0.0092*** -0.0122*** -0.0164*** -0.0116** -0.0268*** -0.0279***
Di ec axes -0.0045*** Omi ed -0.0038** Omi ed -0.0035* Omi ed -0.0068*** Omi ed
Indi ec axes Omi ed 0.0003 Omi ed 0.0006 Omi ed 0.0022 Omi ed 0.0022
Adjus ed 2
R
0.63 0.60 0.55 0.52 -- -- -- --
p- alue o Sa gan es o o e
iden i ying es ic ions -- -- -- -- 0.49 0.80 0.98 0.96
p- alue o he es o no- i s -o de
se ial co ela ion -- -- -- -- 0.02 0.10 0.10 0.10
No. o obse a ions 121 119 97 95 57 55 41 39
No. o coun ies 33 33 31 31 29 29 23 22
No es: Es ima ion echnique: 5-yea s a e ages, wo-way ixed e ec s; Dynamic model - wo-s ep GMM es ima o . All explana o y a iables a e exp essed in
logs excep popula ion g ow h. * Signi ican a 10% le el; ** Signi ican a 5%, *** Signi ican a 1%
32
33
Table 5 – Sensi i i y Analysis II. Income Inequali y. Dependen a iable: log o Gini Index
Mo e con ol a iables on Model 6 and 8 Ins umen al Va iables on Model 6 and 8
O hogonal equa ions S uc u al equa ions O hogonal equa ions S uc u al equa ions
Omi ed Fiscal Va iable: Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes Indi ec Taxes Di ec Taxes
Ci il libe ies 0.0499* 0.0553* 0.0731** 0.0585* 0.0938* 0.0250 0.0644 0.0466
Educa ion inequali y 0.0985** 0.1129** 0.0733* 0.0993* 0.0856 0.1034** 0.1010 0.1568
G ow h -- -- -0.0114 -0.0083 -- -- -0.0090 -0.0106
Public in es men -0.0530 -0.0854*** -0.0647* -0.1043*** -0.3181** -0.4289** -0.2679* -0.4451
Cu en public expendi u e -0.2394*** -0.3465*** -0.3180*** -0.4176*** -0.4571*** -0.5365*** -0.4483*** -0.6204***
Di ec axes -0.0738** Omi ed -0.0452* Omi ed -0.0583*** Omi ed -0.0619** Omi ed
Indi ec axes Omi ed 0.0823** Omi ed 0.0956** Omi ed 0.04385 Omi ed 0.1176
In e na ional ade -0.0658* -0.0953** -0.0441 -0.0668 -- -- -- --
Disposable income dummy 0.1125** 0.1212** 0.1174** 0.1252** 0.1134*** 0.1165*** 0.1232** 0.1256**
Adjus ed R2 0.61 0.63 0.64 0.66 0.34 0.59 0.45 0.39
No. o obse a ions 103 101 95 93 67 67 63 61
No. o coun ies 33 33 32 32 29 29 28 27
No es: Es ima ion echnique: 5-yea s a e ages, andom e ec s model wi h empo al dummies; IV wo-s age leas squa es. All he a iables a e
exp essed in logs excep disposable income dummy a iable. * Signi ican a 10% le el; ** Signi ican a 5%, *** Signi ican a 1%
Appendix 1: Sou ces and De ini ions o Da a Used in Reg essions
In e na ional T ade: Wo ld De elopmen Indica o s o Wo ld Bank (WDI), expo s plus
impo s as a sha e o GDP.
Popula ion g ow h: Wo ld De elopmen Indica o s o Wo ld Bank (WDI), annual
g ow h a e o popula ion.
Ci il libe ies: F eedom House: index on a scale o 1 o 7, wi h 1 ep esen ing he
highe le el and 7 ep esen ing he lowe le el.
Educa ion inequali y: Cas elló and Doménech (2002), Gini index o educa ion.
Inequali y o income: UNU-WIDER e sion 2b, Gini index.
P i a e In es men : Penn Wo ld Tables 6.1 and Go e nmen Finance S a is ics o
In e na ional Mone a y Fund (GFS-IMF), To al in es men (PWT 6.1) minus
public in es men (GFS-IMF) as a sha e o GDP.
Human Capi al: Ba o and Lee (2001), a e age yea s o schooling o he popula ion
aged 25 and o e .
In la ion: Wo ld De elopmen Indica o s o Wo ld Bank (WDI), decembe - o-decembe
change in consume p ice index in logs (CPI).
GDP: Penn Wo ld Table 6.1 da abase, Real GDP pe capi a in logs (RGDPCH, 2002
PPP$).
GDP g ow h - Penn Wo ld Table 6.1 da abase, annual GDP g ow h (GDP – GDP -1)
Cu en public expense: Go e nmen Finance S a is ics o In e na ional Mone a y Fund
(GFS-IMF), cu en expendi u e o gene al go e nmen as a sha e o GDP.
Public In es men : Go e nmen Finance S a is ics o In e na ional Mone a y Fund
(GFS-IMF), public in es men o gene al go e nmen as a sha e o GDP.
Di ec axes: Go e nmen Finance S a is ics o In e na ional Mone a y Fund (GFS-
IMF), e enues o gene al go e nmen due o di ec axes as a sha e o GDP.
Indi ec axes: Go e nmen Finance S a is ics o In e na ional Mone a y Fund (GFS-
IMF), e enues o gene al go e nmen due o indi ec axes as a sha e o GDP.
T ansac ions in inancial asse s and liabili ies: Go e nmen Finance S a is ics o
In e na ional Mone a y Fund (GFS-IMF), e enues minus expenses in inancial
asse s as a sha e o GDP
O he e enues: Go e nmen Finance S a is ics o In e na ional Mone a y Fund (GFS-
IMF), e enues o gene al go e nmen due o o he axes, g an s and o he
e enues as a sha e GDP.
Go e nmen Su plus/De ici : Go e nmen Finance S a is ics o In e na ional Mone a y
Fund (GFS-IMF), o al e enues minus o al ou lays o gene al go e nmen as a
sha e o GDP.
34
Appendix 2: Desc ip i e S a is ics
Mean S anda d
De ia ion Minimum Maximum Obse a ions
GDP g ow h O e all
Be ween
Wi hin
2.22 2.83
1.70
2.52
-21.49
-3.99
-15.44
13.96
5.97
10.65
N=284
n=43
T-ba = 6.60
Log (Ini ial GDP) O e all
Be ween
Wi hin
8.75 2.47
1.97
1.52
0
2.60
3.54
10.80
10.22
15.50
N=301
N=43
T=7
Inequali y o income O e all
Be ween
Wi hin
34.48 9.09
8.67
2.64
18.66
22.29
25.74
58.7
52.77
43.02
N = 235
n = 43
T-ba = 5.46
Cu en public expense O e all
Be ween
Wi hin
17.18 5.64
5.28
2.04
3.99
6.88
10.54
39.34
32.31
24.76
N = 284
n = 43
T-ba = 6.60
Public in es men O e all
Be ween
Wi hin
4.38 1.72
1.40
1.12
1.13
1.90
1.78
9.10
7.35
7.97
N = 178
n = 43
T-ba = 4.14
P i a e in es men O e all
Be ween
Wi hin
15.99 7.18
6.07
3.26
2.22
5.35
1.73
45.91
36.92
26.07
N = 178
n = 43
T-ba = 4.14
Di ec axes O e all
Be ween
Wi hin
11.97 6.53
6.32
1.97
1.67
2.17
3.76
31.90
28.35
20.89
N = 242
n = 43
T-ba = 5.63
Indi ec axes O e all
Be ween
Wi hin
9.73 3.97
3.72
1.90
0.01
2.58
0.98
24.30
18.82
16.84
N = 205
n = 42
T-ba = 4.88
O he e enues O e all
Be ween
Wi hin
27.94 11.70
10.69
5.19
4.43
8.88
14.82
56.83
45.42
44.43
N = 233
n = 43
T-ba = 5.42
T ansac ions in inancial asse s and
liabili ies O e all
Be ween
Wi hin
1.99 4.84
2.76
3.96
-6.69
-2.00
-9.72
44.39
11.88
34.50
N = 236
n = 43
T-ba = 5.49
Go e nmen Su plus/De ici O e all
Be ween
Wi hin
4.20 6.23
4.00
4.78
-16.20
-6.20
-13.68
28.10
14.77
18.66
N = 263
n = 43
T-ba = 6.11
Popula ion g ow h O e all
Be ween
Wi hin
0.89 0.79
0.71
0.37
-1.33
-0.13
-0.70
3.15
2.44
1.96
N = 301
n = 43
T = 7
Human capi al O e all
Be ween
Wi hin
2.26 1.21
1.08
0.57
0.39
0.80
0.77
5.08
4.66
3.55
N = 247
n = 36
T-ba = 6.86
In e na ional ade O e all
Be ween
Wi hin
75.54 42.24
39.97
15.05
13.49
20.17
11.23
231.53
190.47
143.16
N = 269
n = 42
T-ba = 6.40
In la ion O e all
Be ween
Wi hin
14.71 26.30
17.91
21.00
-0.34
2.00
-49.31
198.51
76.66
166.44
N = 272
n = 43
T-ba = 6.32
Educa ion inequali y O e all
Be ween
Wi hin
26.57 11.63
10.82
4.57
9.3
12.7
8.85
66.00
54.5
43.32
N = 252
n = 36
T = 7
Ci il libe ies O e all
Be ween
Wi hin
2.10 1.68
1.29
1.13
1
1
-1.47
7
5.33
6.10
N= 284
n = 43
T-ba = 6.60
Sou ces: Fiscal a iables come om GFS - FMI
The Gini coe icien comes om UNU-WIDER e sion 2b
In es men and GDP come om Penn Wo ld Table 6.1
Educa ion comes om Ba o and Lee (2001)
T ade and in la ion come om Wo ld De elopmen Indica o s o he Wo ld Bank (WDI-WB)
The Gini o educa ion come om Cas elló and Domènech (2002)
The a iable o ci il libe ies comes om F eedom House, 2007
35

36
Appendix 3: Sample o coun ies
High Income (22):
Aus alia, Aus ia, Belgium, Canada, Denma k, Ne he lands, Finland, F ance, Ge many,
Iceland, I eland, Is ael, I aly, Ko ea Republic, Luxembou g, No way, Po ugal, Spain,
Sweden, Swi ze land, Uni ed Kingdom and Uni ed S a es.
Uppe middle income (21):
A gen ina, Bela us, Chile, Colombia, C oa ia, Czech Republic, Dominican Republic,
G eece, Hunga y, Jamaica, La ia, Malaysia, Mal a, Mau i ius, Pe u, Poland, Singapo e,
Slo ak Republic, Sou h A ica, Tu key and U uguay.
TÍTOL
NUM AUTOR DATA
Feb e 2011Economic G ow h and Inequali y: The Role o Fiscal
Policies
11.05
Leonel Muinelo,
O iol Roca-Sagalés
Feb e 2011Homogeneización en un Sis ema de ipo Leon ie (o
Leon ie -S a a).
11.04
Xose Luis Quiñoa,
Laia Pié Dols
Feb e 2011Ciudades que con ibuyen a la Sos enibilidad Global11.03
I an Muñiz Oli e a,
Rose Masjuan,
Pau Mo e a,
Mi
q
uel-An
g
el Ga cia Lo
p
ez
Feb e 2011Medición del pode de me cado en la indus ia del
cob e de Es ados Unidos: Una ap oximación desde la
pe spec i a de la Nue a O ganización Indus ial
11.02
And és E. Luengo
Gene 2011
Mone a y Policy Rules and Financial S ess: Does
Financial Ins abili y Ma e o Mone a y Policy?
11.01
Ja omí Baxa,
Roman Ho á h,
Bo ek Vašícek
Desemb e
2010
Is Mone a y Policy in New Membe s S a es
Asymme ic?
10.10
Bo ek Vasicek
Desemb e
2010
CO2 emissions and economic ac i i y: he e ogenei y
ac oss coun ies and non s a iona y se ies
10.09
Ma ías Piaggio,
Emilio Padilla
Desemb e
2010
Inequali y ac oss coun ies in ene gy in ensi ies: an
analysis o he ole o ene gy ans o ma ion and inal
ene gy consump ion
10.08
Juan An onio Du o,
Emilio Padilla
Se emb e
2010
How Does Mone a y Policy Change? E idence on
In la ion Ta ge ing Coun ies
10.07
Ja omí Baxa,
Roman Ho á h,
Bo ek Vasícek
Juliol 2010The Wage-P oduc i i y Gap Re isi ed: Is he Labou
Sha e Neu al o Employmen ?
10.06
Ma ika Ka anassou,
Hec o Sala
Juliol 2010Oil p ice shocks and labo ma ke luc ua ions10.05
Ja ie O doñez,
Hec o Sala,
Jose I. Sil a
Juliol 2010Vulne abili y o Po e y: A Mic oeconome ic App oach
and Applica ion o he Republic o Hai i
10.04
E ans Jado e
Maig 2010Nue os y iejos c i e ios de en abilidad que
concue dan con el c i e io del Valo Ac ual Ne o.
10.03
Joan Pasqual,
Emilio Padilla
Ma ç 2010Memo y in Con ac s:
The Expe ience o he EBRD (1991-2003)
10.02
Lionel A ige,
Rosella Nicolini
Feb e 2010Language knowledge and ea nings in Ca alonia10.01
An onio Di Paolo,
Josep Lluís Raymond-Ba a