623
HOW DOES ECONOMIC FREEDOM INFLUENCE THE RELATIONSHIP BETWEEN
GOVERNMENT SIZE AND CONVERGENCE?
Le en e Nádasi
Uni e si y o Deb ecen, Facul y o Economics, Ins i u e o Economics, Deb ecen, Hunga y
nadasi.le en [email protected]
Abs ac : In his pape we p esen some empi ical esul s abou absolu e and condi ional
con e gence o eal GDP wi hin 121 coun ies using c oss-coun y da a. We assume ha
he e is an in e ed U-shaped cu e, which desc ibes he ela ionship be ween economic
g ow h and go e nmen spending. I is mainly because he ins i u ional condi ions o
p oduc i i y do no exis a lowe le els o go e nmen spending. A highe le els, he
go e nmen needs o le y highe axes o inance i s expendi u es, which hinde s g ow h.
So he e can be somewhe e an op imal go e nmen edis ibu ion le el ha maximizes
g ow h. This op imal le el depends on ins i u ional ac o s ha can be g abbed by ce ain
Economic F eedom and Wo ldwide Go e nance Indica o s. I was no ou aim necessa ily
o de e mine exac ly he le el o op imal go e nmen edis ibu ion, i would be di icul
because o he he e ogenei y o he coun ies, only o make a compa ison be ween ee
and less ee coun ies, and d aw some conclusions abou how his le el depends on hese
ins i u ional a iables. Summing up we can say ha ou aim was o compa e ee and less
ee, legally “good” and “bad”, as well as co up and less co up coun ies om he aspec
o go e nmen edis ibu ion le el. I we di ide coun ies in o ee and less ee coun ies
and assume ha bo h g oups ha e an in e ed U-shaped cu e, he op imal le el o
go e nmen spending sha e is la ge in he iche coun ies because o hei be e
ins i u ional sys em. These esul s do no con adic hose indings ha decla e posi i e o
nega i e ela ionship be ween go e nmen spending and economic g ow h. One pa o he
li e a u e p esumes ha he e is only one op imal le el o go e nmen spending, we poin
ou ha he e can be a leas wo op imal le els, and hey depend on he ins i u ional
quali y.
Keywo ds: Economic g ow h; con e gence; go e nmen spending; c oss-coun y analysis
JEL classi ica ion: O40; O47; H50
1. In oduc ion
Va ious au ho s ha e examined he ela ionship be ween go e nmen size and economic
g ow h. Some o he esea che s ound a posi i e link, like My dal (1960) and Eas e lin
(1974), bu some o hem ound i nega i e, o example Rubinson (1977), Came on
(1982), Landau (1983) and Ma low (1986). The mos ecen ones also ound i nega i e,
like Rome o-A ila and S auch (2008), A onso and Fu ce i (2010) among many o he
au ho s. They used mainly c oss-coun y and panel da a o illus a e i . All in all, he e is no
gene al con incing empi ical e idence abou he sign o his ela ionship, bu some o he
a iables will help o unde s and he p oblem.
A nai e conclusion om he gene al amewo k o Solow (1956) model leads o he inding
ha poo e coun ies a e g owing wi h a as e speed han he iche ones because o he
dec easing ma ginal p oduc o capi al. So heo e ically, poo e coun ies can ca ch up
(Solow, 1956, 1957).
On he con a y, acco ding o Ba o (1990), we know ha he e is no absolu e con e gence
wi hin he coun ies o he wo ld. Absolu e con e gence occu s only wi hin mo e
homogeneous g oups like Eu opean Union o Eu ope and Asia oge he . A ican coun ies
o example seem o lag behind and i is mainly due o ins i u ional di e ences. I he less
624
de eloped coun ies had he same ins i u ion quali y, hey would be able o con e ge o he
de eloped coun ies. Summing up, absolu e con e gence is no ue due o ins i u ional
di e ences, bu he e can be a condi ional con e gence. I he go e nmen spending is
nea he op imal le el in a gi en coun y, his coun y can g ow as e . As we men ioned,
oo high and oo low edis ibu ion hinde s g ow h. We assume ha iche economies ha e
a highe op imal le el, and he ins i u ionally unde de eloped coun ies ha e a lowe
op imal go e nmen size. Acco ding o Tanzi (2005) he op imal le el o go e nmen
spending is 30 pe cen . We do no suppose ha he e is only one op imal le el. We
assume ha he e is a leas wo op imal le els, and i depends on ins i u ional a iables.
These a iables a e measu ed by Economic F eedom, Legal Sys em & P ope y Righ s,
and Con ol o Co up ion.
In Chap e 2 we examine 121 coun ies. Mos o he da a come om he Wo ld
De elopmen Indica o s (WDI, 2014). A i s we p o ide some desc ip i e s a is ics o
asce ain whe he i is an absolu e con e gence among all o he coun ies, han we
di ided he sample in o g oups acco ding o he abo emen ioned he e ins i u ional
a iables: Economic F eedom, Legal Sys em and P ope y Righ s, and inally Con ol o
Co up ion. I ee economies a e iche , bu g ow slowe han he o he s, less ee ones
can ca ch up. We will see ha i is no ue.
F ee coun ies a e much weal hie and g ow as e a he same ime, so less de eloped
coun ies do ha e a chance unde hese ci cums ances o app oach hem. The esul is
simila in he o he wo si ua ions. So one di iding a iable is enough o un eg essions. In
Chap e 3 we un h ee eg ession, he i s is o all coun ies, bu hese coun ies a e e y
he e ogeneous, so he i is no so good. A e ha we un he same on only he ee
coun ies, han only on less ee coun ies. O cou se, we know ha ee economies a e
also weal hie , and ha e be e legal ins i u ions. Acco ding o ou esul s ee coun ies
ha e a highe op imal go e nmen edis ibu ion le el han less ee, less de eloped, so
hose coun ies ha ha e wo se legal sys em.
2. Va iables, sample and da a
In he nex Chap e , om he ables we can see h ee ca ego iza ions o 121 coun ies. We
examined an 18-yea pe iod be ween 1995 and 2012. We ook in o accoun he ini ial eal
GDP pe capi a RGDP1995, he a e age eal g ow h a e o he pe iod, indica ed by g ,
and he s uc u e o go e nmen spending. G o a iable means he o al go e nmen
spending as a pe cen age o GDP. This expense means cash paymen s o ope a ing
ac i i ies o he go e nmen in p o iding goods and se ices. I includes compensa ion o
employees (such as wages and sala ies), in e es and subsidies, g an s, social bene i s,
and o he expenses such as en s and di idends (WDI, 2014). Gcon a iable means he
go e nmen consump ion also as a pe cen age o GDP, mo e exac ly he inal go e nmen
expendi u e including all go e nmen cu en expendi u es o pu chases o goods and
se ices and compensa ion o employees (WDI, 2014). Ghea, Gedu and Gmil a iables
simply mean he go e nmen expendi u e spen on Heal h, Educa ion and Mili a y as a
pe cen age o go e nmen expendi u e.
I also used some impo an a iables o di ide he da a in o wo g oups: The Economic
F eedom o he Wo ld Index, indica ed by EFW (EFW, 2012), Legal Sys em & P ope y
Righ s, indica ed by Legal (EFW, 2012), which is a componen o Economic F eedom, and
inally he Con ol o Co up ion, indica ed by CC, also in 2012, which is a e y impo an
componen o The Wo ldwide Go e nmen Indica o s (WGI 2014).
625
3. Desc ip i e s a is ics
Le ’s see a i s some desc ip i e s a is ics by di iding he sample in o g oups. The i s
ca ego iza ion is acco ding o Economic F eedom o he Wo ld Index, which is annually
published by he Canadian F ase Ins i u e. Table 1 shows ha he e a e 64 uni s in he
sample, whe e EFW>7, and 57, whe e EFW<7. Coun ies abo e 7 a e supposed o be
ee, below 7 a e less ee. We can see om Table 1 ha ee economies had mo e han
i e imes la ge eal GDP in 1995, and since hen ee economies a e g owing wi h as e
pace han he o he s. So less ee coun ies will ne e ca ch up, because hei ins i u ional
sys em does no allow g owing mo e apidly. F ee coun ies ha e la ge go e nmen
edis ibu ion, because hey a e iche and hey can a o d i . Acco ding o Wagne ’s law,
he demand on public goods is g owing in he de eloped coun ies (Ram, 1987). I we
ca e ully examine he di e ences in he s uc u e o he expendi u es, we can see ha
money spen on heal h is la ge , bu spen on educa ion and mili a y a e lowe in he
de eloped coun ies han in he less de eloped ones on he pe cen age o o al
go e nmen expendi u es.
Table 1: Di e gence among ee and less ee coun ies
EFW
N
Mean
S d. E o Mean
RGDP1995
>= 7,00
64
14747,84
1892,79
< 7,00
57
2737,82
627,83
g
>= 7,00
64
2,66
0,21
< 7,00
57
2,36
0,25
G o
>= 7,00
64
29,17
1,45
< 7,00
57
21,12
1,13
Gcon
>= 7,00
64
16,52
0,64
< 7,00
57
14,39
0,68
Ghea
>= 7,00
64
12,50
0,44
< 7,00
57
9,91
0,40
Gedu
>= 7,00
64
13,32
0,43
< 7,00
57
15,73
0,61
Gmil
>= 7,00
64
8,01
0,85
< 7,00
57
10,14
0,80
Sou ce o da a: EFW (2014), WDI (2014)
The second di ision is acco ding o he quali y o Legal Sys em & P ope y Righ s (EFW,
2014). I he quali y measu e o he legal sys em and p ope y igh s (Legal) is mo e han
5.5, he gi en coun y is conside ed o be legally “good”, o i i is unde 5.5, we call i
legally “bad”. We can clea ly see om Table 2 ha legally “good” coun ies we e mo e han
i e old iche han legally “bad” ones, and he g ow h a e (g ) was also signi ican ly
la ge , so his di e ence is con inuously inc easing. We can also clea ly see ha legally
“good” go e nmen s spend a la ge sha e o hei GDP han he o he s. O cou se we do
no know he causal ela ionship, bu we can guess ha he legally “good” economies a e
also iche , so hey can a o d o ha e a bigge edis ibu ion.
626
Table 2: Di e gence among legally “good” and “bad” coun ies
Legal
N
Mean
S d. E o Mean
RGDP1995
>= 5.5
60
15506,13
1968,87
< 5.5
61
2779,51
625,12
g
>= 5.5
60
2,76
0,25
< 5.5
61
2,27
0,21
G o
>= 5.5
60
30,38
1,46
< 5.5
61
20,45
1,04
Gcon
>= 5.5
60
17,57
0,59
< 5.5
61
13,50
0,64
Ghea
>= 5.5
60
12,31
0,44
< 5.5
61
10,26
0,44
Gedu
>= 5.5
60
13,84
0,47
< 5.5
61
15,06
0,59
Gmil
>= 5.5
60
7,96
0,92
< 5.5
61
10,05
0,73
Sou ce o da a: EFW (2014), WDI (2014)
I also di ided he sample in o wo g oups wi h espec o he con ol o co up ion index
om 2012. I ound ha he median alue was abou a -0.25. The di e ence is bigge han
by he ea lie wo cases. Coun ies ha a e less co up , had 9 imes la ge eal GDP in
1995 han he co up ones, and he g ow h a e o hese wo g oups we e oughly he
same, so he e is only a sligh ca ch up be ween hese wo g oups.
Table 3: No signi ican con e gence among “co up ” and “less co up ” coun ies
CC
N
Mean
S d. E o Mean
RGDP1995
>= -.2500
59
16678,46
1950,81
< -.2500
62
1869,18
324,75
g
>= -.2500
59
2,44
0,19
< -.2500
62
2,59
0,26
G o
>= -.2500
59
31,52
1,43
< -.2500
62
19,53
0,92
Gcon
>= -.2500
59
18,48
0,57
< -.2500
62
12,70
0,54
Ghea
>= -.2500
59
12,18
0,48
< -.2500
62
10,42
0,41
Gedu
>= -.2500
59
14,08
0,56
< -.2500
62
14,81
0,52
Gmil
>= -.2500
59
7,29
0,89
< -.2500
62
10,65
0,74
Sou ce o da a: WGI (2014), WDI (2014)
627
These desc ip i e s a is ics do no con ain new esul s, bu hese indings coincide wi h
some o well-known ac s (Ba o 1990, Czeglédi 2007). I p o es ha his da ase is
eliable, so we can un u he es s wi h hem. I is clea ha e e y ca ego iza ion p o ided
almos he same esul , so he ea e only one a iable will be enough o di ide he sample.
4. Resul s o he eg essions
As we assumed, he e mus be an in e ed U-shaped ela ionship be ween go e nmen
spending and economic g ow h, because oo high and oo low go e nmen spending does
no os e g ow h. So a simple eg ession model was used, in which we pu he G o and
also he squa e o he G o as explaining a iables. Ou eg ession model was he
ollowing:
(1)
This amewo k is acco ding o Mankiw a al. (1992). We omi ed some a iables, which
can cause a li le bias, bu he o de o hese le els a e good. In his case we can es ima e
he G o le el ha maximizes g ow h acco ding o he ollowing o mula:
(2)
O cou se, we know ha hese es ima ions a e no exac , because he e can be some
omi ed a iable dis o ions, we only wan o use i in compa ing hese le els among ee
and less ee coun ies.
Table 4 shows he coe icien s o he whole sample, i con ains 121 coun ies. O cou se I
ook he ini ial lnRGDP1995 as well, whe e ln means he na u al loga i hm. Theo e ically i
we ge nega i e coe icien o he G o 2 and posi i e o G o , i e e s o an in e ed U-
shape as we assumed, whe e G o 2 means he squa e o a iable G o . In addi ion, i we
ge nega i e coe icien o he ini ial GDP, i e e s o con e gence. The op imal alue o
G o can be calcula ed om he coe icien s, which maximizes g ow h. Fo he whole
sample, i is 35 pe cen . O cou se we canno ake i se iously, because o he omi ed
a iables, bu I will use his numbe only o compa e i o he o he cases. The Adjus ed R
Squa e is e y low, i means ha he model i lea es much o be desi ed. We do no wan
o d aw conclusions om his case. As we will see la e (Table 5 and Table 6), he op imal
size is 49 pe cen o he de eloped coun ies and 27 o he less de eloped coun ies.
Sadly hese numbe s a e no exac because o he dis o ions, hey do no show
magni udes, we can only compa e hem, o which i is la ge .
Table 4: Coe icien s o he i s eg ession (Adjus ed R Squa e=0,04)
Model
Uns anda dized
Coe icien s
Sig.
B
S d.
E o
(Cons an )
3,898
,873
4,468
,000
lnGDP1995
-,447
,122
-3,657
,000
G o
,140
,053
2,622
,010
G o 2
-,002
,001
-2,439
,016
Sou ce o da a: WGI (2014), WDI (2014)
2
3210 1995ln G o G o RGDPg
bbbb
+++=
3
2
*
2
b
b
-=G o

628
Table 5 con ains only hose coun ies whe e he EFW index is la ge han 7. As we saw
om desc ip i e s a is ics, hese a e mainly he iche , he de eloped coun ies. We can
conclude con e gence om he coe icien o he ini ial GDP, and in e ed U-shape om
he G o 2. The op imal G o le el would be 49 pe cen , so bigge han in he o he s. In his
case, he Adjus ed R Squa e is e y much highe han in he p ecious case, pa ly because
he sample is mo e he e ogeneous.
Table 5: Coe icien s o he second eg ession (Adjus ed R Squa e=0,429)
Model (EFW>7)
Uns anda dized
Coe icien s
Sig.
B
S d.
E o
(Cons an )
8,431
1,048
8,044
,000
lnGDP1995
-,843
,137
-6,161
,000
G o
,098
,056
1,756
,084
G o 2
-,001
,001
-1,748
,085
Sou ce o da a: WGI (2014), WDI (2014)
F om Table 6 i is clea ha he e is also a con e gence wi hin less ee coun ies, bu he e
he nega i e coe icien is s a is ically no signi ican . Despi e his, i seems ha he e is an
in e ed U-shaped cu e. The op imal G o le el o hese p obably less de eloped
coun ies ha maximizes g ow h is 27 pe cen . O cou se, his eg ession esul does no
mean ha i is exac ly he op imal le el, bu i is qui e su e ha lowe han in he iche
economies. Acco ding o some esea che s, he ela ionship be ween g ow h and
go e nmen size is in e ed U-shaped (Tanzi 2005, A mey 1995, Peden 2005, Sheehy
1995). In all hese cases, he coe icien o he G o 2 is nega i e, which e e s o he
in e ed U-shape. The coe icien o lnGDP1995 is nega i e o he de eloped coun ies, i
means he e is a signi ican con e gence among hem. I is a li le bi su p ising ha he
coe icien o lnGDP1995 is nega i e o he whole sample, which a i s glance seems o
con adic o Ba o (1990), bu we mus ema k ha he whole sample is no he whole
wo ld. Finally, he e is no con e gence be ween he less de eloped coun ies, because he
coe icien o lnGDP1995 is no signi ican , bu i seems ha he e is a U-shape ela ionship
be ween he in es iga ed a iables, he coe icien o G o 2 is e en signi ican .
Table 6: Coe icien s o he hi d eg ession (Adjus ed R Squa e=0,053)
Model (EFW<7)
Uns anda dized
Coe icien s
Sig.
B
S d. E o
(Cons an )
,477
1,966
,243
,809
lnGDP1995
-,295
,223
-
1,327
,190
G o
,324
,143
2,262
,028
G o 2
-,006
,003
-
2,010
,050
Sou ce o da a: WGI (2014), WDI (2014)
629
5. Concluding ema ks
F om he desc ip i e s a is ics we can see he well-known ac , ha he ee , “legally”
be e and he less co up coun ies a e also weal hie . And he e is a di e gence among
hese coun ies, bu a leas does no seem o be any signi ican con e gence be ween
hem. So i seems ha he ee , legally be e and less co up coun ies a e no only
iche , bu g ow as e . Bu why do he poo e coun ies lag behind, and do no ca ch up?
We can explain i wi h ins i u ional di e ences. I he poo coun ies had he same quali y o
ins i u ional sys em, hey would be able o ake he ad an age o highe ma ginal p oduc o
capi al and a ac much mo e in es men . Bu hese coun ies a e no so ee, he quali y
o legal sys em is wo se, and he co up ion is bigge .
We assumed acco ding o A mey (1995) and Tanzi (2005), ha he e is an in e ed U-
shaped ela ionship be ween go e nmen spending and economic g ow h. I we ake he
ins i u ional quali y in o accoun , we can assume ha he de eloped coun ies ha e a U-
shaped cu e, and he less de eloped ha e ano he . By he de eloped coun ies, he
go e nmen spending le el ha maximizes g ow h is la ge han by he less de eloped
ones. We can see i om Table 5 and Table 6, ha ee coun ies ha e a la ge op imal
le el. I was no ou aim necessa ily o de e mine he exac le el o go e nmen spending
le el, his eg ession me hod does no make i possible because o he omi ed a iable
dis o ions, bu we can conclude ha in he ee coun ies he op imal edis ibu ion is
highe . The go e nmen spending sha e is highe in he de eloped coun ies, because hey
ha e a highe quali y o legal sys em, which is mo e e icien and enables a highe
p oduc i i y.
Re e ences
A onso, A. – Fu ce i D. (2010) Go e nmen Size, Composi ion, Vola ili y and Economic
G ow h. Eu opean Jou nal o Poli ical Economy 26 (4): 517–532.
A mey, D. (1995) The F eedom Re olu ion. Washing on, DC: Regne y Publishing.
Ba o, R. J. (1990) Go e nmen Spending in a Simple Model o Endogenous G ow h.
Jou nal o Poli ical Economy 98 (5): 103–125.
Ba o, R. J. (1997) De e minan s o Economic G ow h: A C oss-Coun y Empi ical S udy.
Camb idge, MA: MIT P ess.
Came on, D. (1982) On he Limi s o he Public Economy. Annals o he Academy o
Poli ical and Social Science 459 (1): 46–62.
Czeglédi Pál (2007) Piaci in ézmények és gazdasági nö ekedés – a mode n osz ák iskola
nézőpon ja, Philosophiae Doc o es, Akadémiai Kiadó, Budapes .
Eas e lin, R.A. (1974): Does economic g ow h imp o e he human lo ? Some empi ical
e idence. In P. A. Da id and M.W. Rede (Eds.), Na ions and households in economic
g ow h. New Yo k: Academic P ess.
Landau, D. (1986) Go e nmen and economic g ow h in he less de eloped coun ies: an
empi ical s udy o 1960–1980. Economic De elopmen and Cul u al Change 35: 35–75.
Ma low, M. L. (1986) P i a e Sec o Sh inkage and he G ow h o Indus ialized
Economies. Public Choice 49 (2): 143–154.
Mankiw, N.G. – Rome D. – Weil, D. N. (1992) “A Con ibu ion o he Empi ics o
Economic G ow h”, The Qua e ly Jou nal o Economics 107, 2:407-437.
My dal, G. (1960) Beyond he wel a e s a e. New Ha en, CN: Yale Uni e si y P ess.
Peden, E. A. (1991) “P oduc i i y in he Uni ed S a es and I s Rela ionship o Go e nmen
Ac i i y: An Analysis o 57 Yea s, 1929–86,” Public Choice 86, 153–73.
Ram, R. (1987) Wagne ’s hypo hesis in ime-se ies and c oss-sec ion pe spec i es:
E idence
om“ eal” da a o 115 coun ies. The Re iew o Economics and S a is ics, 194–204.
Rome o-A ila, D. – S auch, R. (2008): Public Finances and Long-Te m G ow h in Eu ope:
E idence om a Panel Da a Analysis. Eu opean Jou nal o Poli ical Economy 24 (1): 172–
191.
630
Rubinson, R. (1977) Dependency, Go e nmen Re enue, and Economic G ow h, 1955-70.
S udies in Compa a i e In e na ional De elopmen , 12, 3-28.
Sheehey, E. (1993) The e ec o go e nmen size on economic g ow h. Eas e n Economic
Jou nal,19(3), 321-328.
Solow, R. (1956) A Con ibu ion o he Theo y o Economic G ow h, Qua e ly Jou nal o
Economics 70, 65-94.
Solow, R. (1957) Technical Change and he Agg ega e P oduc ion Func ion. Re iew o
Economics and S a is ics, 39, 3: 321-320.
Tanzi, V. (2005) The Economic Role o he S a e in he 21s Cen u y, Ca o Jou nal o
Economics, 25, 3: 617-638.
WDI (2014) Wo ld De elopmen Indica o s, 2014. h p://da a.wo ldbank.o g/da a-
ca alog/wo ld-de elopmen -indica o s, [12. Ma 2015.]
WGI (2014) The Wo ldwide Go e nmen Indica o s, 2014. www.go indica o s.o g [10 Ma
2015.]