In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
517
In es iga ing Capi al Mobili y and Sa ing-
In es men Rela ionship: Case S udy o MENA
Coun ies
Mansoo Mahini Zadeh
Assis an P o esso o Economics, Yazd Uni e si y, Yazd, I an
Elahe Mahmoodi
1
Ph.D. S uden in Economics, Yazd Uni e si y, Yazd, I an
Majid Mahmoodi
Facul y Membe , Velaya Uni e si y, I anshah , I an
Abs ac
This pape de e mines he deg ee o capi al mobili y and sa ing- in es men
associa ion by using he Felds ein- Ho ioka (1980) app oach among Middle
Eas and No h A ican (MENA) coun ies du ing 1990-2011. Acco ding o he
Felds ein- Ho ioka hypo hesis, in a coun y wi h high deg ee o capi al
mobili y he e should be no ela ionship be ween domes ic sa ing and domes ic
in es men and in e sely, in a coun y wi h low deg ee o capi al mobili y he e
is a high co ela ion be ween domes ic sa ing and domes ic in es men . Using
panel da a and employing Random e ec Model o es ima e he model, he
ob ained esul s show ha capi al is highly mobile in hese coun ies.
Keywo ds: Capi al mobili y, Feldes ein and Ho ioka app oach, MENA
egion, Panel Da a
Ci e his a icle: Mahini Zadeh, M., Mahmoodi, E., & Mahmoodi, M. (2015). In es iga ing
Capi al Mobili y and Sa ing- In es men Rela ionship: Case S udy o MENA Coun ies.
In e na ional Jou nal o Managemen , Accoun ing and Economics, 2(6), 517-526.
In oduc ion
Rapid economic libe aliza ion in eme ging ma ke s in he la e 1980s and he ea ly
1990s h ough echnological p og esses in anspo a ion sys ems and communica ion,
1
Co esponding au ho ’s email: mahmoodi.elah[email p o ec ed]
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
518
has led o apid inc easing in ade, expanding o in e na ional capi al lows and mo e
in eg a ion o coun ies in ecen yea s. (Eicheng een, 2003)
De e mining he deg ee o capi al mobili y is a s iking issue in mac oeconomics
analysis because i is impo an o de e mining he op imal iscal and mone a y policy,
managing he exchange a e, se ing he ax induced by in la ion and o many o he
pu poses. (Jamilo , 2013)
Coun ies can gain om inc easing capi al mobili y. Fo ins ance capi al mobili y can
imp o e alloca ion o global esou ces. In addi ion, by s imula ing in es men , i can
imp o e g ow h beyond he limi s o domes ic sa ing (Slimane, e al., 2013). Wi h his
iew i is necessa y o de ec he ex en o capi al mobili y and in his pape we p obe
his issue in MENA coun ies.
The es o his pape is o ganized as ollows: sec ion 2 in oduce he heo e ical
amewo k, sec ion 3 desc ibes a summa y o empi ical s udies conce ning wi h his
issue, sec ion 3 p esen s he model and da a, sec ion 4 discusses he me hodology and
inally sec ion 5 includes summa y and conclusion.
Theo e ical amewo k
In he economics li e a u e, he e a e se e al ways o examine he deg ee o
in e na ional capi al mobili y. One es p oposed by Felds ein and Ho ioka (1980) o
measu ing capi al mobili y is o in es iga e he ela ionship be ween sa ing and
in es men . As an indica o o capi al mobili y he sa ing-in es men co ela ion o
Felds ein and Ho ioka has some ad an ages o e o he indica o s in he li e a u e. Fo
example, i does no ha e o deal wi h he p oblem o he e ogenei y o asse s, as occu s
wi h he es s o pa i y condi ions, and does no inco po a e mul iple hypo heses as he
Eule equa ion es s o he consump ion-smoo hing app oach. (Rocha, 2007)
So, in his pape we a emp o su ey he deg ee o capi al mobili y in MENA
coun ies.
Using c oss-sec ion eg essions ac oss 16 OECD coun ies o he pe iod om 1960
o 1974, Felds ein and Ho ioka (1980) ound ha he es ima ion o he co ela ion
be ween sa ing and in es men anges om 0.87 o 0.91. This esul implies a low
deg ee o capi al mobili y among indus ial coun ies, in con adic ion wi h he belie
ha he indus ial coun ies had ew ba ie s o capi al mo emen s.
They s a ed ha sa ing esponded o in e na ional oppo uni ies o in es men and
in es men in one coun y can be inanced by domes ic and o eign sa ing.
(Eslamloueyan & Ja a i, 2010). They a gued ha “Wi h pe ec wo ld capi al mobili y,
he e should be no ela ion be ween domes ic sa ing and domes ic in es men ”
(Felds ein and Ho ioka, 1980, p.317). I means ha unde capi al immobili y sa ing
equals in es men .
So, b ie ly, he elds ein- Ho ioka app oach (he ea e is no ed by F-H) implies ha
in a coun y wi h high deg ee o capi al mobili y he e should be no ela ionship
be ween domes ic sa ing and domes ic in es men and in e sely, in a coun y wi h low
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
519
deg ee o capi al mobili y he e is a high co ela ion be ween domes ic sa ing and
domes ic in es men .
Li e a u e e iew
In ecen yea s, a lo o s udies ha e been conduc ed on he deg ee o in e na ional
capi al mobili y based on he ela ion be ween domes ic in es men and domes ic
sa ing.
K ol (1996) belie ed ha using ime- a e aged da a in c oss- sec ional sa ing-
in es men eg essions can bias he esul s agains capi al mobili y. So, in o de o sol e
his p oblem he es ima ed he model by using annual da a o he pe iod 1962 o 1990
o 21 OECD coun ies. The esul s indica ed ha capi al is in e na ionally mobile.
Ano he suppo o his esul in his pape is he exis ence o a signi ican impac o
bo h sa ing and in es men on he cu en accoun .
Adebola and Dahalan (2012) by applying ARDL bound es o es ima e F-H (1980)
model, examined he deg ee o capi al mobili y in Tunisia du ing 1970- 2009 and inally
hey showed ha he e is low capi al mobili y in Tunisia.
Meu onen (2012) in es iga ed he ela ion be ween domes ic sa ing and domes ic
in es men in Finland du ing he pe iod 1960Q1-1986Q2 and 1986Q3-2011Q2. In he
ea lie pe iod s udied, he a iables we e ound o be co-in eg a ed wi h a ec o
sugges ing low capi al mobili y. Howe e , in he la e pe iod he ela ionship
disappea ed.
Cheng Li (2010) showed ha he sa ings and in es men (bo h exp essed as a ios o
GDP) a e posi i ely co ela ed o a sample o 28 Chinese p o inces du ing 1978-2006.
He a gued ha acco ding o he F-H hypo hesis such a co ela ion can be in e p e ed as
e idence o low capi al mobili y wi hin china.
Jain and Sami (2011) a emp ed o in es iga e he exis ence o F-H (1980) hypo hesis
in Small Island S a es Employing h ee co-in eg a ion p ocedu es. They indica ed ha
o e whelming e idence ha sa ings and in es men a e co-in eg a ed in hese small
economies.
Bankage and Eggoh (2011) in es iga ed he ela ionship be ween sa ing and
in es men a es o 37 A ican coun ies du ing 1970–2006. They used he ecen ly
de eloped Pooled Mean G oup co-in eg a ion echnique and ound ha in he long- un,
capi al was ela i ely mobile in A ican coun ies, while, in he sho - un, coe icien s
we e no signi ican .
slimane e al (2013) es ed he hypo hesis o pe ec capi al mobili y in Tunisia and
Mo oco using F-H me hodology. In his s udy hey in oduced addi ional a iables o
explain he deg ee o inancial openness mo e e ec i ely. They concluded ha he
deg ee o capi al mobili y was ela i ely high in hese coun ies. Following Felds ein
and Ho ioka, Dooley e al. (1987) es ima ed he F-H (1980) eg ession wi h c oss-
sec ion da a on de eloping coun ies. Thei e idence indica es a close ela ion be ween
sa ing and in es men , sugges ing a low deg ee o capi al mobili y. Rocha (2000) ied
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
520
o p o ide mo e e idence on he deg ee o capi al mobili y in de eloping coun ies
using annual da a o he pe iod 1996-1960 o 36 de eloping coun ies. The esul s o
his s udy sugges ha he deg ee o capi al mobili y in de eloping coun ies is highe
han usually belie ed. In addi ion, capi al mobili y inc eased conside ably a e 1975.
Model speci ica ion and da a
Model speci ica ion
Following Felds ein and Ho ioka (1980) and many o he esea ches, we es ima e he
ela ionship be ween domes ic sa ing and domes ic in es men in o de o de e mine he
deg ee o capi al mobili y. They es ima e he ollowing c oss-sec ion eg ession
(I/Y)i = α + β (S/Y)i + εi (1)
Whe e (I/Y) and (S/Y) a e espec i ely he sa ing and in es men a es o coun y i ,
is he sa ings- e en ion coe icien and is he e o e m. Acco ding o he F-H
inding o a small-open economy whe e capi al is pe ec ly mobile in e na ionally,
should be close o ze o. I equals ze o, hen he e is no ela ionship be ween sa ing
and in es men . On he o he hand, i is la ge, capi al is conside ed immobile
in e na ionally. Fo example, i equals one, hen all addi ional sa ing goes o inance
domes ic in es men (Felds ein and Ho ioka, 1980).
Following he o iginal s udy o F-H (1980), sa ings is de ined as g oss domes ic
sa ings as a pe cen age o GDP, and in es men is measu ed by g oss ixed capi al
o ma ion di ided by GDP.
Da a
Because using ime-a e aged da a on sa ing and in es men , like ha F-H used in
hei esea ch can bias he esul s owa d ejec ing he idea ha capi al is highly mobile,
so, in his pape we use panel da a o es ima e he model. Gene ally, using o panel da a
has nume ous bene i s in con as wi h ime se ies da a: con olling o indi idual
he e ogeni y and gi e mo e in o ma i e da a, mo e a iabili y, less co-linea i y among
he a iables, and mo e e iciency (Bal agi, 2005). The e o e, in his pape we ha e
applied panel da a o 14 MENA coun ies o e he 1990 o 2011 yea s. Annual da a
ha e been aken om he Wo ld De elopmen Indica o s (WDI). Sample o coun ies
includes: Alge ia, Bah ain, Egyp , I an, Jo dan, Kuwai , Lebanon, Libya, Mo occo,
Qa a , Saudi A abia, Sy ia, Tunisia, Yemen.
Me hodology
LR Tes
A he i s s ep, in o de o selec be ween pooled da a and panel da a, we ha e used
LR es . The null hypo hesis o his es implies ha he da a ha e pooled s uc u e. In
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
521
con as , he al e na i e hypo hesis implies ha we should es ima e he model based on
panel da a.
Panel Uni Roo Tes
In he nex s ep, in o de o iden i y whe he he da a a e s a iona y o no , panel uni
oo es s should be employed.
Se e al Panel uni oo es s ha e been p esen ed o in es iga e he s a iona y
p ope ies o panel da a. This pape applied ou es s p oposed by Le in e al. (LLC,
2002), Im e al. (IPS, 2003), B ei ung (2000) and Fishe - ype es p oposed by Maddala
and Wu (1999) and Choi (2001) o es he null hypo hesis o ha ing uni oo .
Following Dickey and Fulle (1979, 1981), Le in and Lin (1993), and Le in, Lin and
Chu (2002), conside a panel ex ension o he null hypo hesis ha each indi idual ime
se ies in he panel con ains a uni oo agains he al e na i e hypo hesis ha all
indi idual se ies a e s a iona y (Hsiao, 2003).
The adjus ed -s a is ic o LLC can be w i en as ollows:
2*
*
*
ˆˆ
ˆˆ
()
NmT
mT
NTS
(2)
Whe e 𝜇𝑚𝑇
∗ and 𝜎𝑚𝑇
∗ a e he mean and s anda d de ia ion adjus men s p o ided by
able 2 o LLC. Le in, Lin and Chu show ha 𝑡𝜌
∗ is asymp o ically dis ibu ed as N (0,
1).
The es o Im, Pesa an and Shin (IPS, 2003) allows o a he e ogeneous coe icien
o yi -1 and p opose an al e na i e es ing p ocedu e based on a e aging indi idual uni
oo es s a is ics. IPS sugges s an a e age o he ADF es s when ui is se ially
co ela ed wi h di e en se ial co ela ion p ope ies ac oss c oss-sec ional uni s.
The -s a is ic o IPS can be exp essed as ollows:
1
1
1
( [ | 0])
(0,1)
1 a [ 0]
N
iT i
i
IPS
N
iT i
i
N E
N
N
N
(3)
Values o E [ iT | ρi = 0] and a [ iT | ρi = 0] ob ained om he esul s o Mon e Ca lo
simula ions ca ied ou by IPS.
As men ioned in Bal agi (2005), LLC and IPS es s may no keep nominal size well
when ei he N is small o N is la ge ela i e o T. B ei ung (2000) ound ha he LLC
and IPS es s su e om a d ama ic loss o powe i indi idual-speci ic ends a e
included. B ei ung sugges s a es s a is ic ha does no employ a bias adjus men whose
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
522
powe is subs an ially highe han LLC o he IPS es s using Mon e Ca lo expe imen s.
The es s a is ic o B ei ung (2000) panel uni oo es has he ollowing o m:
2 * *
1
1
2 * *
1
1
N
ii
i
BN
ii
i
yX
X A A
(4)
Maddala and Wu (1999) and Choi (2001) p oposed a Fishe - ype es o uni oo ,
which combines he p- alues om uni oo es s o each c oss-sec ion i o es o uni
oo in panel da a. The Fishe es is nonpa ame ic and dis ibu ed as chi-squa e wi h
wo deg ees o eedom:
𝑝𝜆 = −2 ∑𝑙𝑜𝑔𝑒𝜋𝑖 (5)
Panel Co- in eg a ion Tes
Se e al es s we e p esen ed o su ey he exis ence o co-in eg a ion in panel da a
model. This pape applied panel co-in eg a ion es o Ped oni (1999, 2004) and Kao
(1999).
Ped oni has p esen ed se en s a is ics o es ing he null hypo hesis o no co-
in eg a ion in panel da a. Fou s a is ics called panel co-in eg a ion s a is ics and based
on pooling along wha is commonly e e ed o he “wi hin-dimension” and o he h ee
s a is ics de eloped by Ped oni called g oup-mean panel co in eg a ion s a is ics, a e
based on pooling along wha is commonly e e ed o “be ween-dimension”.
(Dahma deh and Mahmoodi, 2012)
Hausman Tes
I should be no ed, in panel da a models, o es ima e he model we should selec one
o he Fixed E ec s Model (FEM) o andom E ec s Model (REM). Hausman (1978)
ha e p esen ed a es o his aim. Acco ding o Hausman es unde he null hypo hesis
and assump ion o he lack o co ela ion be ween c oss-sec ional da a and o he
explana o y a iables, bo h es ima o (LSDV and REM GLS) a e inconsis en bu he
LSDV es ima o is also ine icien . Bu , in con as in e ms o co ela ion be ween
c oss-sec ional da a and o he explana o y a iables (FEM), he LSDV es ima o is
consis en bu GLS is inconsis en . (G eene, 2003, 301)
B ie ly, his es in oduces i s assump ions as ollows; H0: The wo es ima o s should
no be signi ican ly di e en om each o he howe e , he andom e ec s model is
p e e ed, and H1 implies he exis ence o ixed e ec s model and ejec ion o andom
e ec s model. (Shahiki ash & Ghod a , 2012)
Empi ical esul s
LR Tes
As s a ed be o e, in o de o iden i y he s uc u e o da a we can use he LR es . The
esul s o his es ha e been shown in able 1.
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
523
Table1. LR Tes
E ec s Tes
S a is ic
d. .
P ob.
C oss-Sec ion F
19.802
(12,267)
0.000
C oss-Sec ion Chi-Squa e
178.877
12
0.000
Panel Uni Roo Tes
The esul s o Im e al. (IPS, 2003), Le in e al. (LLC, 2002), B ei ung (2000) and
Fishe - ype panel uni oo es s a e epo ed in able 1.
Table 1.Panel Uni Roo Tes s
Tes
Va iable
LLC (2002)
IPS (2003)
B ei ung (2000)
ADF-Fishe
PP-Fishe
(I/Y)
-1.616
(0.052)
-2.315
(0.010)
0.893
(0.814)
47.717
(0.005)
40.563
(0.034)
(S/Y)
-3.978
(0.000)
-3.451
(0.000)
1.053
(0.854)
53.736
(0.001)
50.966
(0.002)
No e: P obabili y alues ha e been epo ed in pa en hesis.
As seen in able 1, in gene al, he esul s o di e en panel uni oo es s indica e ha
all a iables a e s a iona y in le els.
Panel Co-in eg a ion Tes
The esul s o Ped oni panel co- in eg a ion es s ha e been p esen ed in able 3.
Table 3.Ped oni Panel Co-in eg a ion Tes
S a is ics
Panel -s a is ic
2
0.855 ***
Panel ρ-s a is ic
-
1.622**
Panel non-pa ame ic (PP) -s a is ic
-
1.920 **
Panel pa ame ic (ADF) -s a is ic
-
2.046 **
G oup ρ-s a is ic
0
0.132
G oup non-pa ame ic -s a is ic
-
1.608 **
G oup pa ame ic -s a is ic
-
1.934 **
No e: *** and ** deno e s a is ical signi icance a he 1and 5% le els, espec i ely.
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
524
Acco ding o he esul s o able3.Excep G oup ρ-s a is ic, o he s a is ics show ha
he hypo hesis o no co-in eg a ion is s ongly ejec ed. So we can de ec he e is a long
un ela ionship in he model.
Model Es ima ion
Finally, a he las s age, we should es ima e he model. As men ioned be o e, in
o de o es ima e he eg ession, we ha e o selec one o he ixed e ec s model (FEM)
o Random E ec s Model (REM).
Table4. Resul s o hausman es
Tes summa y
Chi- sq. S a is ic
d. .
P ob.
C oss sec ion andom
1.962
1
0.161
The esul s o Hausman Tes ha e been shown in able 4. Based on he esul s o
able 4 he null Hypo hesis which implies using Random E ec Model (REM) is
accep ed. So, we employed REM o es ima e he eg ession. The esul o model
es ima ion has been epo ed in able 5.
Table5. Resul s o model es ima ion
Va iable
Coe icien
P ob.
(S/Y)
( -s a is ic)
0.035
(1.880)
0.061
R2
0.48
As seen in able 5 he (S/Y) coe icien is equal o 0.035 and acco ding o he F-H
hypo hesis, i implies ha he capi al is highly mobile in MENA coun ies du ing he
1990-2011 pe iod.
Conclusion
This pape aims o in es iga e he ex en o capi al mobili y in MENA coun ies o
he pe iod 1990-2011. To he bes o ou knowledge, he e isn’ any pape which p obes
he deg ee o capi al mobili y in MENA coun ies. To achie e his pu pose, we used he
F-H app oach and in es iga ed he Sa ing-in es men ela ion. Se e al panel uni oo
es s indica ed ha sa ing and in es men a e s a iona y in le els. Also, acco ding o he
esul s o Ped oni (1999) co-in eg a ion es s he null hypo hesis o no co-in eg a ion
was ejec ed and exis ence o long un ela ion be ween sa ing and in es men was
con i med. Based on he esul s o es ima ing he model, he sa ing- in es men
coe icien is equal o 0.035 which implies pe ec capi al mobili y in MENA coun ies.
So, on he con a y o common belie ha he majo i y o de eloping coun ies
including MENA coun ies cons ain ba ie s and es ic ions o in e na ional ade and
capi al lows, he esul s demons a e ha his belie should be modi ied. Also his esul
e eals ha all he na ional sa es in hese coun ies don’ in es in domes ic coun y
necessa ily because capi al is qui e mobile.
In e na ional Jou nal o Managemen , Accoun ing and Economics
Vol. 2, No. 6, June, 2015
ISSN 2383-2126 (Online)
© IJMAE, All Righ s Rese ed www.ijmae.com
525
Re e ences
Alam, M and Islam, R., (2010). Re isi ing he Felds ein-Ho ioka Hypo hesis o
sa ings, in es men and capi al mobili y: e idence om 27 EU coun ies, In e na ional
Jou nal o Economics,.4 (1): 71-90.
Bagnai, A. and Manzocchi, S. (1996). Uni oo es s o capi al mobili y in he less
de eloped coun ies, Wel wi scha liches A chi e, 132 (3): 545-557.
Bangake, C., Eggoh, J.C., (2012) Pooled Mean G oup es ima ion on in e na ional
capi al mobili y in A ican coun ies, Resea ch in economics, 66: 7-17.
Dahma de, N and Mahmoodi, E (2012).The e ec o economic globaliza ion on
in la ion in 15 OECD coun ies, in e na ional jou nal o basic science and applied
esea ch, 2(5): 545-550.
De We , A.H., Van Eyden, R., (2005). Capi al mobili y in sub-Saha an A ica: a
panel da a app oach. Sou h A ican Jou nal o Economics, 73: 1–22
Dooley, M., J. F ankel & Ma hieson, D.,(1987). In e na ional Capi al mobili y in
de eloping coun ies e sus indus ialized coun ies: Wha do sa ings-in es men
co ela ions ell us? IMF S a Pape s, 34: 503-529.
Eicheng een, B (2003). Capi al lows and c isis, MIT p ess.
Eslamlouiyan, K and Ja a i, M (2010). Capi al mobili y, Openness, Sa ing-
in es men ela ionship in Asia, Economic Modeling, 27: 1246-1252.
Felds ein, M., and Ho ioka,. C (1980). Domes ic sa ing and in e na ional capi al
low. The Economic Jou nal, 90: 314-329.
Holmes, M.J., (2005).Wha do Sa ings-In es men Co ela ions ell us abou he
In e na ional Capi al Mobili y o Less De eloped Coun ies? Jou nal o Economic
In eg a ion, 20(5):590-603.
Im KS, Pesa an MH, Shin Y, (2003). Tes ing o uni oo s in he e ogeneous panels,
Jou nal o Econome ics. 115(1): 53-74.
Kayıkçı, F (2012) Sa ing In es men Co ela ion and Capi al Mobili y: E idence
om T ansi ion Economies, In e na ional Jou nal o Social Sciences and Humani y
S udies, 4(2):191- 198.
K ol, R., (1996). In e na ional Capi al Mobili y: e idence om panel da a, Jou nal o
in e na ional Money and Finance,15(3): 467-474.
Li, C., (2010) Sa ings, In es men , and Capi al Mobili y wi hin China, China
Economic Re iew, 21(1): 14-23
