You are on page 1of 19

Journal of Economics and Sustainable Development www.iiste.

org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*1
Economic Growth and Wagners Hypothesis: The Nigerian
Experience

+niefio, -do
1
( ./arles Effiong
2

1. Department of Economics( -niversit0 of .alabar( .alabar( Nigeria
2. Department of Economics( .ross 1iver State .ollege of Education( +,am,pa
2Email of t/e .orresponding aut/or3 aniefio,.benedict4gmail.com

Abstract
5agner6s law viewed t/at public e7penditure is a conse8uence rat/er t/an cause of national income /ence9 it
pla0s no role in generating national income. 5/ile :e0nes viewed t/at public e7penditure is a cause rat/er t/an
effect of national income t/erefore can be used to /eig/ten economic activities. In t/e developing econom0 li,e
Nigeria( w/ic/ of t/ese sc/ools prevails( 5agner or :e0nes; </is stud0 see,s to answer t/is 8uestion b0
determining t/e nature and direction of causalit0 between government spending and t/e economic growt/ as well
as t/e relations/ip between t/ese macroeconomic variables. </e stud0 emplo0s t/e =ranger causalit0 and
ordinar0 least s8uare !&>S# tec/ni8ue to evaluate t/e empirical evidence of t/e relations/ip between fiscal
polic0 and economic growt/ in Nigeria b0 using an econometric tec/ni8ue t/roug/ multiple regression models
t/at was derived from t/e Solow growt/ model. +fter testing for granger causalit0( t/e result reveals t/at t/ere is
a bidirectional relations/ip between government spending and economic growt/ in Nigeria( t/us we find support
of 5agner6s and :e0nesian /0pot/eses. +lso( t/e anal0sis s/owed t/at government e7penditure in our Nigerian
econom0 /ad direct effect on economic growt/9 t/erefore( t/ere is need for appropriate policies wit/ respect to
government spending ,nowing t/at it affects t/e level of growt/. <o ac/ieve sustainable economic growt/(
=overnment e7penditure s/ould be increased in t/e econom0.
Keywords: =overnment E7penditure( Economic =rowt/( 5agner law and =ranger causalit0.

1. NT!"#$%T"N
"ublic e7penditure and economic growt/ /ave been at t/e focus of public finance( since t/e magnitude
of public e7penditure /as been increasing over time in almost all t/e countries of t/e world. It is t/erefore
necessar0 for governments to ,now t/e causal relations/ip between t/e two. </eoreticall0( t/ere are two
competing sc/ool of t/oug/t defining t/is causal relations/ip. ?irst( 5agner !1$$@# postulated t/at public
e7penditure is an endogenous variable and t/at t/ere e7ist long-run tendencies for public e7penditure to grow
relativel0 to some national income aggregates suc/ as t/e gross domestic product !=D"#. Aoreover( public
e7penditure is a conse8uence rat/er t/an cause of national income. In ot/er words( t/e causalit0 between public
e7penditure and national income runs from national income to public e7penditure. </erefore( 5agner6s law
viewed t/at public e7penditure pla0s no role in generating national income.
Bowever( :e0nes !1C@)# argued t/at public e7penditure is an e7ogenous variable and can be used to
generate national income. ?or t/is reason( public e7penditure is a cause rat/er t/an effect of national income
w/ic/ is in contrast wit/ 5agner6s law. Be raised t/e idea t/at during economic depression government
e7penditure can be used to /eig/ten economic activities. </erefore( t/e causal relations/ip s/ould run from
public e7penditure to national income !<ang 2C#.
</e term fiscal polic0 /as conventionall0 been associated wit/ t/e use of ta7ation and public
e7penditure to influence t/e level of economic activities. It /as to do wit/ two maDor activities9 ta7ation on one
side9 t/en government e7penditure on t/e ot/er side. </is stud0 will onl0 concentrate on t/e government
e7penditure side.
In Nigeria( government e7penditure /as been on t/e rise owing to t/e /uge receipts from production and
sales of crude oil( and t/e increased demand for public goods li,e roads( power( education( communication( and
/ealt/. Aoreover( t/ere is increasing need to provide bot/ internal and e7ternal securit0 for t/e people and t/e
nation. -nfortunatel0( t/is rising government e7penditure /as not translated into meaningful growt/ and
development( as Nigeria ran,s among t/e poorest countries in t/e world. </e result of government role in
economic activities and t/e ac/ievements in economic performance /ave been mi7ed. ?or instance( t/e econom0
will e7perience growt/ in real output in some 0ears and declines in ot/ers. Aeanw/ile( t/e econom0 is mostl0
dominated b0 t/e public sector e7cept recentl0 t/at t/e government is tr0ing to adopt privatiEation polic0. Fut
t/e overall picture is low scoring for t/e countr06s developmental efforts. </e obDectives of monetar0 and fiscal
policies in Nigeria are wide-ranging( involving =ross Domestic "roduct growt/ rate( reduction in t/e rates of
inflation and unemplo0ment( improvement in t/e balance of pa0ments( accumulation of financial savings and
e7ternal reserves as well as stabilit0 in Naira e7c/ange rate. </e guiding principle as well as instruments applied
to attain t/ese obDectives( /owever( /ave until recentl0 been far from ade8uate. "er/aps( t/is could be attributed
to inconsistenc0 in t/e formulation and implementation of vibrant policies.
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*2
'arious empirical studies on t/e relations/ip between government e7penditure and economic growt/
arrived at different and even conflicting results. Some studies suggest t/at increase in government e7penditure
on socio-economic and p/0sical infrastructures impact on long run growt/ rate. ?or instance( government
e7penditure on /ealt/ and education raises t/e productivit0 of labour and increase t/e growt/ of national output.
E8uall0( e7penditure on infrastructure suc/ as road( power etc. reduces production costs( increase
private sector investment and profitabilit0 of firms( t/us ensuring increase in economic activities and economic
growt/ !see9 Farro( 1CC9 Farro G Sali -i-Aartin( 1CC29 1ou7( 1CC*9 &,oDie( 1CC%#. &n t/e ot/er /and(
observations t/at growt/ in government spending( mainl0 based on non-productive spending is accompanied b0
a reduction in income growt/ /as given rise to t/e /0pot/esis t/at t/e greater t/e siEe of government intervention
t/e more negative is its impact on economic growt/ !=lomm and 1avi,umar( 1CC79 +bu and +bdulla/( 21#.
=overnment e7penditure is considered an important variable w/ic/ ma0 determine c/anges in national
income in developing countries li,e Nigeria. In ot/er words( fiscal polic0 is a maDor economic stabilisation
weapon t/at involves measure ta,en to regulate and control t/e volume( cost and availabilit0 as well as direction
of mone0 in an econom0 to ac/ieve some specified macroeconomic polic0 obDective and to counteract
undesirable trends in t/e Nigerian econom0 !=bosi( 1CC$#. <o stimulate t/e economic growt/ b0 means of fiscal
polic0( t/e countr0 must adopt more instruments. </ese according to Ebimobowei !21# include9 t/e financing
of direct investments w/ic/ t/e private sector would not provide an ade8uate 8uantities9 t/e efficient suppl0 of
certain public services w/ic/ are necessar0 to ensure t/e basic conditions to displa0 t/e economic activit0 and
long term investments9 and t/e financing of public activities so as to minimiEe t/e distortions to come up wit/
t/e decisions to spend and invest proper in t/e private sector. </ese instruments can be gotten t/roug/ t/e nature
and level of government spending in t/e econom0. </oug/ it can also be ac/ieve eit/er b0 an increase or a
decrease in ta7es( government e7penditures constitute t/e bedroc, of fiscal polic0 but in realit0( government
polic0 re8uires a mi7ture of bot/ fiscal and monetar0 polic0 instruments to stabiliEe an econom0 because none of
t/ese single instruments can cure all t/e problems in an econom0 !Ndi0o and -da/( 2@#.
Despite several fiscal measures introduced since 1C$)( and given t/e prominence of fiscal polic0 in
macroeconomic management in Nigeria( growt/ /as not accelerated as e7pected and as suc/ povert0 remains
widespread and pervasive( particularl0 in t/e rural areas. &ne could as,(
5/at is t/e role of fiscal polic0 in inducing economic growt/ in an econom0( redistributing income and
reducing povert0 in Nigeria; .ould fiscal polic0 be designed so as to ensure economic growt/ and reduce
povert0 w/ile maintaining macroeconomic stabilit0; ?urt/ermore( does government spending in Nigeria
contribute to economic growt/ and development; </ese are crucial 8uestions to as, given t/e renewed interest
of t/e current democratic structure in povert0 alleviation and given t/at fiscal polic0 is t/e arrow/ead of t/e
polic0 pac,age of t/e current polic0 framewor, in Nigeria. </is stud0 intend to focus specificall0 on one side
!government e7penditure# in ac/ieving t/e following obDectives9 1. <o determine t/e nature and direction of
causalit0 between government spending and economic growt/ in Nigeria( b0 testing for t/e 5agner6s /0pot/esis
and its reverse !:e0nesian approac/#. 2. Determining t/e relations/ip between governments spending and
economic growt/ wit/ ot/er control variables li,e mone0 suppl0( domestic investment and labour. </is will /elp
to decide if t/e current pace of public spending in our econom0 is productive and s/ould be encouraged or not.
</e paper /as five sections9 section one is t/e Introduction( section two contains t/e >iterature review( section
t/ree is t/e Aet/odolog0( section four is Empirical results and discussion w/ile section five is conclusion and
polic0 recommendations.

&. 'TE!AT$!E !E(EW
Aan0 studies s/ow t/at government e7penditure is positivel0 related wit/ economic growt/ and
povert0 reduction but due to /ig/ e7penditure most of t/e developing countries are facing t/e problem of fiscal
deficit. ?iscal deficit leads to inflation in t/e econom0. +ccording to Ae/mood and Sadi8 !21#( in man0
developing countries /ig/ fiscal deficit crowding out t/e private investment in t/e long run and decreases t/e
emplo0ment and output w/ic/ adversel0 affects t/e povert0.
+s economic growt/ ma0 increases t/roug/ government spending. Jams/aid et al !21# e7amined t/e
relations/ip between economic growt/ and government e7penditure( bot/ at bivariate !aggregate# and
multivariate !disaggregate# s0stems and concluded t/at economic growt/ causes government e7penditure at
bivariate level and also supported t/at increase in =D" causes growt/ in government e7penditure - 5agner6s
/0pot/esis.
Sing/ and Sa/ni !1C$*# investigated t/e relations/ip between national income and government e7penditures in
India and discovered no causal relations/ip among t/e variables indicating t/e failure of bot/ 5agner6s law and
:e0nes /0pot/esis. In Au/lis and Ba,an !2@# wor,( an investigation of t/e long-run relations/ip between
public e7penditure and =D" for t/e <ur,is/ econom0 was studied using time series annual data. </e0 emplo0ed
co-integration and =ranger .ausalit0 tests and discovered t/at neit/er 5agner6s >aw nor :e0nes6 /0pot/esis
was valid in <ur,e0.
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*@
+lso( Ergun and <uc, !2)# in stud0ing t/e direction of causalit0 between national income and government
e7penditures for Indonesia( Aala0sia( "/ilippines( Singapore( and </ailand using =ranger causalit0 test(
discovered no Support for t/e /0pot/esis t/at causalit0 runs from government e7penditures to national income.
</is was found onl0 in t/e case of "/ilippines. </ere was no evidence for t/is /0pot/esis and its reverse for t/e
ot/er countries.
Similarl0( &lugbenga and &wo0e !27# studied t/e relations/ips between government e7penditure and
economic growt/ for a group of @ &E.D countries( using annual data during t/e period 1C7-2%. </e
variables of interest were total government e7penditure !<=E# and gross domestic product !=D"# wit/ t/e use of
co-integration and =ranger causalit0 tests. </e results s/owed t/e e7istence of a long-run relations/ip between
government e7penditure and economic growt/. Aore so( t/e aut/ors observed a unidirectional causalit0 from
government e7penditure to growt/ for onl0 1) countries( /ence supporting t/e :e0nesian /0pot/esis.
Nevert/eless( causalit0 runs from economic growt/ to government e7penditure in 1 among t/e @ countries(
confirming 5agner6s law as 8uoted in Seviten0i !212#( w/ile a bi-causal relations/ip between government
e7penditure and economic growt/( for four countries was discovered.
&mo,e !2C# investigated t/e direction of causalit0 between =overnment e7penditure !=E# and National
Income !NI# in Nigeria using co-integration and =ranger .ausalit0 tests for annual time series data. In /is result(
/e discovered t/at t/ere was no long-run relations/ip e7isted between government e7penditure and national
income in Nigeria between 1C7 and 2%. +lso( t/e =ranger causalit0 test revealed t/at causalit0 ran from
government e7penditure to national income t/us concluding t/at government e7penditure pla0s a significant role
in promoting economic growt/ in Nigeria. Jams/aid et al. !21# e7amined t/e nature and t/e direction of
causalit0 in "a,istan between public e7penditure and national income. +ppl0ing t/e <oda-Hamamoto causalit0
test for annual data( t/e0 concluded t/at t/ere was a unidirectional causalit0 running from =D" to government
e7penditure( w/ic/ supports 5agner6s >aw. Interest for t/e 5agner /0pot/esis attracted t/e attention of man0
economists after t/e translation of t/e original wor, of 5agner b0 .oo,e !1C%$#( /owever t/e interest /ad
declined at t/e end of 1C7s. +lt/oug/( t/e increased public spending in most countries( new development of
econometric tec/ni8ues( and t/e last translation of 5agner6s wor, b0 Fie/l !1CC$# attracted again t/e interest of
man0 polic0 ma,ers and economists.
+s indicated b0 1ic/ter and Dimitrios !212#( t/ere are si7 !)# different versions of 5agner6s law3
"eacoc, and 5iseman !1C)1#( =upta !1C)7#( =offman !1C)$# ( "r0or !1C)C#( Ausgrave !1C)C#( =offman and
Aa/ar !1C71# and Aann !1C$#. </ese are listed below9

1. )eacoc*+Wiseman ,ersion
L6tIaJ a1LYtJet a1K1 -1.
Notes: >= is t/e log of real government e7penditures( >=. is t/e log of real government consumption
e7penditure( >" is log of population( >!=LH# is t/e log of t/e s/are of government spending in total output(
>!HL"# is t/e log of t/e per capita real output( >!=L"# is t/e log of t/e per capita real government e7penditures (>
H is t/e log of real =D".
&. )eacoc*+Wiseman share ,ersion -/ann ,ersion.

!6LY#IJ 1LYtJet 1K -&.

0. /1sgra,e ,ersion

!CLY#tIyJy1 !YLP# MJet y1K -0.

*. G1pta ,ersion

!CLP#tI6J6!YLP#t MJet 61K1 -2.

3. Go44man ,ersion

L6tIIJI1 !YLP# MJe t I1K1 -3.

5. )ryor ,ersion

L6Ct66J 61LYtJet 01K1 -5.

Derimbas !1CCC# stated t/at N"ublic finance studies( following 5agner( /ave considered public e7penditure as a
be/avioural variable( similar to private consumption e7penditure. F0 contrast( macroeconometric models(
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

**
essentiall0 following :e0nes( /ave treated public e7penditure as an e7ogenous polic0 instrument designed to
correct s/ort-term c0clical fluctuations in aggregate e7pendituresO!Demirbas 1CCC as 8uoted in 1ic/er G
Dimitrios 212#.
In Nigeria( t/e trend of government e7penditure during 1C7 to 212 s/ows t/at government
e7penditure in t/e earl0 7s was mainl0 deficits /as s/own in figure 1 below.
?rom t/e figure we /ave it t/at government e7penditure /ave t/e same pattern of movement wit/
national income e7cept in 1C77 to 1C7$ w/en t/ere was a decrease in government e7penditure.



7ig1re 1: Trend o4 Go,ernment Expendit1re and Nationa8 ncome in Nigeria -19:;+&;1&.

0. /ETH"#"'"G< AN# #ATA
</is stud0 adopts a 8uantitative met/od to evaluate t/e empirical evidence of t/e relations/ip between
government e7penditure and economic growt/ in Nigeria. </e met/od of anal0sis /as been an econometric
tec/ni8ue using multiple regression models t/at is derived from t/e Solow growt/ model. </e data used in t/is
stud0 is secondar0 annual time series covering 1C7 P 212. </e basic data for t/is anal0sis are rate of9 =ross
Domestic "roduct !=D"#( government total e7penditure( labour force( !pro7ied b0 population# gross fi7ed capital
formation !=?.?# used as pro70 for domestic investment and mone0 suppl0. </ese data were collected from t/e
statistical bulletin P a publication of t/e .entral Fan, of Nigeria.
Fased on t/e specific obDectives of t/is stud0( we approac/ed t/e met/odolog0 t/us3
&bDective 1 was anal0sed b0 using t/e =ranger causalit0 test to ascertain t/e causal relations/ip between
government spending and economic growt/ in Nigeria. </e =ranger causalit0 test is a statistical /0pot/esis test
for determining w/et/er one time series is useful in forecasting anot/er. In ot/er word it is a test to c/ec, if t/e
action or performance of one variable /as an effect or causes t/e e7istence of anot/er.
&bDective 2 was anal0sed b0 using t/e ordinar0 least s8uare !&>S# regression tec/ni8ue to determine
t/e relations/ip between fiscal polic0 and economic growt/ in Nigeria. In statistics( ordinar0 least s8uares !&>S#
or linear least s8uares is a met/od for estimating t/e un,nown parameters in a linear regression model. </is
met/od minimiEes t/e sum of s8uared vertical distances between t/e observed responses in t/e dataset and t/e
responses predicted b0 t/e linear appro7imation. </e resulting estimator can be e7pressed b0 a simple formula(
especiall0 in t/e case of a single regressor on t/e rig/t-/and side.
</e &>S estimator is consistent w/en t/e regressors are e7ogenous and t/ere is no perfect multicollinearit0(
and optimal in t/e class of linear unbiased estimators w/en t/e errors are /omoscedastic and seriall0
uncorrelated. -nder t/ese conditions( t/e met/od of &>S provides minimum-variance mean-unbiased estimation
w/en t/e errors /ave finite variances. -nder t/e additional assumption t/at t/e errors be normall0 distributed(
&>S is t/e ma7imum li,eli/ood estimator and /ave t/e F>-E properties w/ic/ are Fest >inear -nbiased
Estimator.
5e e7perimented wit/ t/e different functional forms of t/e e8uation relating fiscal and economic
growt/ viE3 linear( semi-log( double-log and e7ponential forms.

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*%
0.1 /"#E' =)E%7%AT"N
In t/is section( we postulate a model t/at see,s to e7amine t/e effects of some selected fiscal polic0 variables on
economic growt/ in Nigeria. &ur specification of a growt/ model is based on t/e Solow growt/ model t/at
emp/asiEed t/e significance of investment !i.e. capital# and labour effectiveness in promoting growt/. </e
Solow growt/ model is s0mbolicall0 represented below3
Q I f !:( ># RRRRRRRRRRRRRRRR !1#
5/ere Q is t/e national output( : represents capital resources emplo0ed and > for unit of labour emplo0ed in t/e
production process. Bowever( since our focus is on t/e public sector influence( t/e model includes =overnment
spending as one of t/e factors t/at e7plain growt/.
</e output !growt/# model specified for t/e purpose of t/is stud0 is presented t/us3
=D"1
t
I f !<=E1
t
( =?.?1
t
( "&"1
t(
AS1
t
# RRRRRRRRRRRRRRR. !2#
!J# !J# !J# !J#
5/ere3
=D"
t
I =rowt/ rate of real =D"
<=E1
t
I growt/ rate of total =overnment e7penditure
=?.?1
t
I rate of Investment !pro7ied b0 =ross ?i7ed .apital formation or Domestic Investment#
"&"1
t
I growt/ rate of t/e population
AS1
t
I rate of Aone0 suppl0 at time t
</e figures in t/e parent/eses represent t/e a priori e7pectations about t/e signs of t/e coefficients.
</e stud0 also emplo0ed t/e following Diagnostic tests3
1. $nit root tests o4 stationarity: In statistics( a unit root test tests w/et/er a time series variable is non-
stationar0 using an autoregressive model. + well-,nown test t/at is valid in large samples is t/e augmented
Dic,e0P?uller test. </e optimal finite sample tests for a unit root in autoregressive models were developed
b0 Denis Sargan and +lo, F/argava. +not/er test is t/e "/illipsP"erron test. </ese tests use t/e e7istence
of a unit root as t/e null /0pot/esis. </e stationarit0 status of t/e variables in t/is stud0 is establis/ed b0
considering t/e order of integration of eac/ variable in t/e model using t/e +ugmented Dic,e0-?uller
!+D?# and "/illips-"erron !""# classes of unit root tests. 5e use t/e "" approac/ to test for stationar0 of
t/e variables because "" test statistic( w/ic/ is a modification of t/e +D?( ta,es into account t/e less
restrictive nature of t/e error process. Aoreover( t/is replaces t/e use of lag in t/e +D? test.
2. Norma8ity test: normalit0 tests are used to determine if a data set is well-modelled b0 a normal distribution
and to compute /ow li,el0 it is for a random variable underl0ing t/e data set to be normall0 distributed.
&ne of t/e normalit0 tests is t/e Jar8uePFera test. </is is a goodness-of-fit test of w/et/er sample data /ave
t/e s,ewness and ,urtosis matc/ing a normal distribution. </e test is named after .arlos Jar8ue and +nil :.
Fera.
@. =eria8 corre8ation: t/is is t/e relations/ip between a given variable and itself over various time intervals.
Serial correlations are often found in repeating patterns w/en t/e level of a variable affects its future level.
E7ample of t/is test is t/e Freusc/P=odfre0-Fertolo test w/ic/ is used to assess t/e validit0 of some of t/e
modelling assumptions in/erent in appl0ing regression-li,e models to observed data series. In particular( it
tests for t/e presence of serial dependence t/at /as not been included in a proposed model structure and
w/ic/( if present( would mean t/at incorrect conclusions would be drawn from ot/er tests( or t/at sub-
optimal estimates of model parameters are obtained if it is not ta,en into account.
*. A1to regressi,e conditiona8 heteroscedasticity -arch '/ test.: In econometrics( +uto 1egressive
.onditional Beteros,edasticit0 !+1.B# models are used to c/aracteriEe and model observed time series.
</e0 are used w/enever t/ere is reason to believe t/at( at an0 point in a series( t/e error terms will /ave a
c/aracteristic siEe( or variance. In particular +1.B models assume t/e variance of t/e current error term or
innovation to be a function of t/e actual siEes of t/e previous time periodsS error terms3 often t/e variance is
related to t/e s8uares of t/e previous innovations. Suc/ models are often called +1.B models !Engle(
1C$2#( alt/oug/ a variet0 of ot/er acron0ms are applied to particular structures of model w/ic/ /ave a
similar basis. +1.B models are emplo0ed commonl0 in modelling financial time series t/at e7/ibit time-
var0ing volatilit0 clustering( i.e. periods of swings followed b0 periods of relative calm.
%. =peci4ication error: In regression anal0sis specification is t/e process of developing a regression model.
</is process consists of selecting an appropriate functional form for t/e model and c/oosing w/ic/
variables to include. +s a first step of regression anal0sis( a person specifies t/e model. If an estimated
model is misspecified( it will be biased and inconsistent. +n e7ample of t/is test is t/e 1amse0 1egression
E8uation Specification Error <est !1ESE<# test !1amse0( 1C)C# is a general specification test for t/e linear
regression model. Aore specificall0( it tests w/et/er non-linear combinations of t/e fitted values /elp
e7plain t/e response variable. </e intuition be/ind t/e test is t/at if non-linear combinations of t/e
e7planator0 variables /ave an0 power in e7plaining t/e response variable( t/e model is mis-specified.

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*)
2. E/)!%A' ANA'<== AN# #=%$=="N "7 7N#NG=
2.1 $NT !""T Tests !es18ts
</e result of unit root test of stationarit0 is s/own in table 1a for t/e +ugmented Dic,e0-?uller !+D?# test and
table 1b for t/e "/illip "erron !""# test below. In t/e results( we discover t/at all t/e variables in t/e estimated
model are integrated in t/e same level I!#( bot/ in t/e +ugmented Dic,e0-?uller !+D?# and "/illip-"erron !""#
test for unit root( /ence t/e variables are stable and t/e model can be estimated using &rdinar0 >east S8uare
!&>S# met/od.

TA>'E 1a: A#7 TE=T !E=$'T
(ariab8e A#7 statistic 1? 8e,e8 3? 8e,e8 1;? 8e,e8 "rder o4
integration
=D"1 ).%71%%@ -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
"&"1 @.2$7222 -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
=?.?1 ).@72* -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
<=E1 %.721C* -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
AS1 -%.$)1 -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
Source: computed by the authors
TA>'E 1b: )H'') )E!!"N TE=T !E=$'T
(ariab8e )) statistic 1? 8e,e8 3? 8e,e8 1;? 8e,e8 "rder o4
integration
=D"1 -@$.1C1%$ -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
"&"1 ).7721% -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
=?.?1 -)2C*%1 -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
<=E1 -).7$$ -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
AS -%.$)1 -@.)*)@*2 -2.C%*21 -2.)1%$17 I!#
Source: computed by the authors

2.& G!ANGE! %A$=A'T< TE=T !E=$'T
TA>'E &: )A!W=E G!ANGE! %A$=A'T< TE=T !E=$'T
"airwise =ranger .ausalit0 <ests
Date3 %L2@L1* <ime3 $31$
Sample3 1C7 212
>ags3 1
Null B0pot/esis3 &bs ?-Statistic "robabilit0
<=E1 does not =ranger .ause =D"1 *2 17.)* .1%
=D"1 does not =ranger .ause <=E1 1%21*.2 .
=?.?1 does not =ranger .ause =D"1 *2 17.121% .1$
=D"1 does not =ranger .ause =?.?1 %$%2*.7 .
"&"1 does not =ranger .ause =D"1 *2 .@1 .C$%C%
=D"1 does not =ranger .ause "&"1 .2*C* .$7%@2
=?.?1 does not =ranger .ause <=E1 *2 2.)$@) .1C)*
<=E1 does not =ranger .ause =?.?1 2.C7%2 .C)12
"&"1 does not =ranger .ause <=E1 *2 ).*%7$@ .1%1*
<=E1 does not =ranger .ause "&"1 .$$%% .7)7)1
"&"1 does not =ranger .ause =?.?1 *2 %.C$2*C .1C)
=?.?1 does not =ranger .ause "&"1 .7*%7 .7$)2@
Source: computed by the authors base on CBN data
</e result of =ranger causalit0 test as s/own in table 2 above reveals t/at t/ere is a significant relations/ip
between government spending and economic growt/ in Nigeria. </e direction of causalit0 flows from
government spending to economic growt/ !:e0nesian /0pot/esis# and also from economic growt/ to
government spending !5agner /0pot/esis#( in ot/er word( t/ere is a bidirectional relations/ip between
government spending and economic growt/. &t/er pair wise results s/ow t/at labour !population# does not
granger cause economic growt/ and vice versa( t/ere is a bidirectional effect between domestic capital and
economic growt/ in t/e Nigerian econom0 as investment !gross fi7ed capital formation /ave a bidirectional
effect wit/ government e7penditure also. >astl0( t/e result indicates t/at t/ere is a unidirectional effect between
population growt/ and growt/ in government e7penditure( flowing from population growt/ to government
e7penditure. ?rom t/is( it is pertinent to be careful of t/e spending made b0 t/e central government in Nigeria as
it will affect t/e level of growt/ eit/er negativel0 or positivel0.
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*7
5e also went a/ead to conduct t/e normalit0 test for t/e distribution using Jar8ue-Fera test. </is is to determine
if our data set is well-modelled b0 a normal distribution and to c/ec, it6s goodness-of-fit. </e result of t/is test
s/ows t/at t/ere is a problem wit/ normalit0 and t/erefore we will use log to transform t/e variables in t/e
model. </is e7plains t/e use of log in t/e estimated model. +not/er test was t/e serial correlation test.
-sing Freusc/ =odfre0 >A we tested for serial correlation of t/e variables in t/e model. </e result
of t/is test is s/own in appendi7 @. In t/e result( t/e top part of t/e output presents t/e test statistics and
associated probabilit0 values. </e test regression used to carr0 out t/e test is reported below t/e statistics. </e
statistic labelled N&bs21-s8uaredO is t/e >A test statistic for t/e null /0pot/esis of no serial correlation. </e
appro7imatel0 one probabilit0 value in t/e auto regressive conditional /eteroscedasticit0 >A test strongl0
indicates t/at t/ere is no presence of /eteros,edasticit0 in t/e residuals.

2.0 "'= !E=$'T "7 /A%!"E%"N"/% (A!A>'E= N7'$EN%NG THE "$T)$T
TA>'E 0: "'= !E=$'T "7 /A%!"E%"N"/% (A!A>'E=
(A!A>'E #"$>'E '"G E@)"NENTA' =E/'"G 'NEA!
. -7.2)C$!.%@C# .*$2@$1!.%72C# -))@!.1C2# -7!.1@)#
<=E1 .**21!.%*C1# 7.@C7!.%*C1# -1$7!.C1*7# 2$.1$!.@CCC#
=?.?1 .$C*%C!.1C)*# -1.)%)!.1*%# -2$%)!.1@7# -@.*@!.#
"&"1 @.1*@C!.*)C@# *.2C)!.# *$!.1$C# @1.1*!.$@#
AS1 .*%@1@2!.$)C# -1.$)!.22@7# -172C!.212C# -1$.C!.1C@#
1
2
.$@7@ .$)$)71 .2*1$)C .%7$)12
+DJ 1
2
.$*** .$%@))2 .112 .%@@%7
?-S<+< 2%.)7%*2 %7.$7)7 1.$@***2 12.712C
".!?-Stat# . . .1%)%7% (1
D5 2.@C%%1 1.CC)** 2.C$2%$C 1.C*)$
+I. -1.2@1CCC @.2CC@2 @C.@$$7 @7.C272
SI. -.C$$22* @.%1*12 @C. 27)7) @$.1@*)
Norm. test .%2*1$!.7)C*# 2.$$7%C!.@%1C# @2.)1%@1!.# 11$7.%%2!.#
5/ite <est 1@.%$)1)!.C@1# 12.)%@7!.12*%# 21.)C@*!.%%# @2.7C$!.#
Source: computed by the authors
Note:
1# Details of regression result is found at t/e appendi7
2# 'ariables are defined as in e8uation 2 in section t/ree( . I intercept9 1
2
I coefficient of determination9
D5 I Durbin-5atson statistic9 +I. I +,ai,e information criteria9 SI. I Sc/warE information criteria
@# 'alues in parent/esis are t/e p-values

</e regression anal0sis was conducted to c/ec, t/e effect of government spending on t/e level of economic
growt/ in Nigeria. 5e emplo0 Eviews %. to anal0se t/e empirical evidence for Nigeria using ordinar0 least
s8uare !&>S# met/od. Specificall0 we e7perimented wit/ different functional form9 double log9 e7ponential9
semi log and linear e8uation. ?rom t/e results we c/ose e7ponential e8uation since t/is result /as t/e /ig/est
e7planator0 power !coefficient of determination# wit/ minimum values of +,ai,e and Sc/warE information
criteria !according to =uDarati and "orter 2C t/e e8uation wit/ t/e least +I. and SI. s/ould be accepted#. In
t/e e7ponential e8uation result( using t/e p-values( t/e t/ree variables t/at are statisticall0 significant are
domestic investment( population and mone0 suppl0. Investment and mone0 suppl0 /as a negative effect on
economic growt/ in Nigeria. </is is contrar0 to t/e apriori e7pectation w/ile increase in population positivel0
affects t/e economic growt/ as e7plained in t/e t/eor0. </e main variable of focus !growt/ of =overnment
e7penditure# positivel0 influences t/e rate of economic growt/ in t/e econom0 and conforms to t/e apriori
e7pectation.
</e coefficient of determination s/ows t/at $%.@T of t/e c/anges in t/e economic s0stem is caused b0
t/e model w/ile onl0 1*.7T of t/is variation is cause b0 variables outside t/e model !error term#. </e p-value
also s/ows t/at t/e entire model is statisticall0 significant at %T level of significance as t/e value is Eero. +lso(
from t/e Durbin-5atson result( t/e result is accepted wit/ t/e result of 1.CC w/ic/ is appro7imatel0 2( s/owing
t/at t/ere is no auto correlation problem in t/e model.
-sing t/e estimated residual from t/e regression( t/e probabilit0 distribution of t/e error terms s/ows t/at t/e
residuals from t/e economic growt/ regression seem to be s0mmetricall0 distributed. +pplication of t/e Jar8ue-
Fera test s/ows t/at t/e JF statistic is about 2.$$7 and t/e probabilit0 of obtaining suc/ a statistic is about @%T.
</erefore( we do not reDect t/e /0pot/esis t/at t/e error terms are normall0 distributed. Fut ,eep in mind t/at t/e
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*$
sample siEe of *2 observations ma0 not be large enoug/. </e w/ite test s/ows t/at t/ere is no /eteroscedasticit0
on t/e basis t/at t/e product of observation and 1
2
of 12.)%@ is /ig/er t/an t/e critical value for @ df.
Aoreover( from t/e adDusted 1-s8uared( we observed t/at $.*T of variation in t/e econom0 is e7plained b0 t/e
model w/ile onl0 1C.)T of c/anges in t/e econom0 is e7plained b0 variables outside t/e model !error term#. </e
?-statistic result indicates t/at t/e entire model is statisticall0 significant given t/at its value is greater t/an t/e
tabulated ?-ratio at %T level of significance.
2.2 )"'%< /)'%AT"N "7 7N#NG=
Fased on t/e empirical findings( we /ave t/e following polic0 implications9
</e granger causalit0 result s/ows t/at t/ere is bidirectional causalit0 between economic growt/ and growt/ of
government e7penditure. </is implies t/at t/e level and nature of government spending will affect t/e rate of
economic growt/ and t/e rate of growt/ too will affect t/e level of government spending. ?or t/is reason(
government s/ould be prudent in t/eir spending ,nowing t/at t/is will /ave effect on t/e level of growt/ in t/e
econom0.
</e result s/ows t/at t/e growt/ of mone0 suppl0 does not promote economic growt/ in our econom0 and is
significant in t/e estimated model. </is to some e7tent could be as a result of balance of trade dise8uilibrium in
t/e international mar,et( as an increase in t/e mone0 suppl0 in our econom0 onl0 leads to increase in t/e
importation of foreign goods and services rat/er t/an promoting t/e domestic product in our econom0 for
increase in economic activities and growt/ in t/e econom0. </is is also t/e true picture of our econom0 since an
average Nigerian prefers foreign goods and services to t/e domestic ones. <o t/is e7tent it is pertinent for
government to control t/e level of mone0 suppl0 to ac/ieve economic growt/ in t/e econom0.
Since labour in t/e result promotes economic growt/ in t/e Nigerian econom0( policies on t/e economic sector
t/at will be labour intensive s/ould be promoted to ma,e use of t/e abundant labour force t/at Nigeria is ric/l0
endowed wit/. </is will in turn promote our economic growt/.
=overnment e7penditure s/ould be increased in t/e econom0 since t/is macroeconomic variable directl0
influences t/e econom0 to promote economic growt/.

3. %"N%'$="N
</is stud0 soug/t to appraise t/e nature and direction of causalit0 as well as t/e relations/ip between
government spending and economic growt/ b0 testing for t/e 5agner6s /0pot/esis and its reverse !:e0nesian
approac/# for of Nigeria spanning from 1C7-212. In order to give t/is stud0 a direction( two null researc/
/0pot/eses were formulated. 1elevant literature was reviewed based on t/e t/eories e7plaining t/e subDect
matter and determining t/e maDor variables of t/e stud0. </e researc/ design used for t/is stud0 was t/e
e7plorator0 and 8uantitative researc/ design. Econometric tec/ni8ues were t/e 8uantitative tools used to conduct
t/e empirical anal0sis in t/e Nigeria6s conte7t. +ccordingl0( starting from t/e nature and direction of causation(
=ranger pair wise causalit0 model was used w/ile a multiple regression model was formulated based on t/e
t/eoretical bac,ground of t/e stud0. &rdinar0 least s8uares !&>S# met/od was used to estimate t/e e8uation( to
evaluate t/e in/erent connectivit0 between government spending and economic growt/. In particular it
underta,es and approac/es to identif0 t/e determinant of Nigeria6s economic growt/( t/e influence of
macroeconomic and investment as well as t/e population effect on growt/.
?irstl0( t/ere is a bidirectional effect or relations/ip between government spending and economic
growt/ in Nigeria. Secondl0( government spending /as positive effect on economic growt/ t/e
coefficient of determination s/ows t/at $%.@T of c/anges in t/e Nigeria6s econom0 is e7plained in t/e
model w/ile onl0 1*.7T of t/e variation in t/e econom0 is cause b0 variables outside t/e model !error
term#. </is proportion s/ows t/at t/e model /as goodness-of-fit and a strong e7planator0 power.
</irdl0( domestic investment /as positive effect on economic growt/ and is statisticall0 significant at
five per cent !%T# level of significance. +n increase in t/e level of population will cause a @.CT
increase in t/e growt/ of Nigeria6s econom0 and t/is is /ig/l0 significant in t/e estimated model. +lso
t/e anal0sis s/owed t/at government e7penditure in our Nigerian econom0 /ad direct effect on
economic growt/( t/erefore( t/ere is need for appropriate policies wit/ respect to government
spending variables for sustainable economic growt/.

=iven t/e outcome of our regression result( we came up wit/ t/e following recommendations for polic0
reforms
1. =overnment e7penditure s/ould be increased to promote economic growt/ in t/e countr0.
2. </e ?ederal =overnment s/ould ensure t/at t/e level of mone0 suppl0 in t/e econom0 is
controlled b0 t/e monetar0 aut/orities to ac/ieve economic growt/ in t/e econom0 since
mone0 suppl0 /as adverse effect on economic growt/ in Nigeria.

@. </e ?ederal =overnment s/ould also lin, /er e7penditure to domestic investment in ot/er to
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

*C
boost investment for economic growt/ in t/e countr0.

!E7E!EN%E
+bu( N and +bdulla/i( -. !21#( =overnment E7penditure and Economic =rowt/ in Nigeria( 1C7-2$3 +
Disaggregated +nal0sis( Business and Economics Journal( 'olume 213 FEJ-*3 1-11.

Farro( 1. !1CC#. =overnment Spending in a Simple model of Endogenous =rowt/. Journal of Political
Economy, C$( S1@ - S12%.
Farro( 1. and Sala-Aartin( U. !1CC2#. "ublic ?inance in Aodels of Economic =rowt/. 1eview of Economic
Studies( %C( )*% - ))1
Farro( 1. J. !1CC1#. Economic =rowt/ in a .ross section of countries. Quarterly Journal of Economics, 1)( *7
- ***.
Fie/l( D. !1CC$#( V5agnerSs >aw3 +n Introduction to and a <ranslation of t/e >ast 'ersion of +dolp/ 5agnerSs
<e7t of 1C11V( "ublic ?inance( vol. %@( pp. 12-111.
.oo,e( N. 1C%$( </ree E7tracts on "ublic ?inance b0 +dolp/ 5agner <ranslated from =erman( Aacmillan G
.o. >td( >ondon.
Dic,e0( D. and ?uller 5. !1C7C#. Distribution of t/e Estimators for +utoregressive <ime Series wit/ a -nit 1oot.
Journal of American Statistical Association, 7*( *27 - *@1.
Ebimobowei( +. !21#. </e 1elations/ip between ?iscal "olic0 and Economic =rowt/ in Nigeria !1CC1 - 2%#.
nternational Journal of Economic !e"elopment #esearch and n"estment, 1!2G@#
Engle( 1.?.( !1C$2#( +utoregressive conditional /eteros,edasticit0 wit/ estimates of t/e variance of -.:.
inflation( Econometrica %( C$7-1$.
Ergun( D and <uc,( .. !2)#( =overnment E7penditure and National Income3 .ausalit0 <ests for ?ive Sout/
East +sian .ountries( nternational Business and #esearch Journal 'ol.%( No( 13 *C-%$.
=arba( +. =. !1CCC#. 1egime ./ange( E7ternal S/oc, and ?ederal ?iscal &perations in Nigeria3 1C7%-C@.
National Economic Societ0 +nnual .onference. pp @%@ - @77.
=bosi( +. N. !2$#. .ontemporar0 Aacroeconomic "roblems and StabiliEation "olicies. "ort Barcourt3
+utomatic 'entures.
=bosi( +.N. !1CC$#. Ban$s, %inancial Crisis and the Ni&erian Economy 'oday( .orporate Impression "ublis/ers(
&werri.
=lomm( J.J. and 1avi,umar( D.J. !1CC7#. N</e =rowt/ of "ublic E7penditure in Selected Developing Nations3
Si7 .aribbean .ountriesO( Public %inance(%inances Publi)ue, @!@#3 %7 P 7*
=uDarati( D. N. and D. .. "oter !2C#. Basic Econometric !%
t/
ed.#. New Hor,3 Ac=raw-Bill
Jams/aid( 1( I8bal( +. and Siddi8i( A. !21#( .ointegration-.ausalit0 +nal0sis between "ublic E7penditures
and Economic =rowt/ in "a,istan( European Journal of Social Sciences( 'ol. 1@( No *3 %%)-%)%.
:e0nes( J. A. !1C@)#. </e =eneral </eor0 of Emplo0ment( Interest and Aone0. New Hor,3 Barcourt Frace
"ublis/ers.
Ae/mood( 1. and S. Sadi8( !21#( </e relations/ip between =overnment E7penditure and povert03 +
.oitegration +nal0sis( Journal of %iscal policy vol.1 !iss.1# pp. 2C-@7
Au/lis( F and Ba,an .. !2@#( .ausalit0 between "ublic E7penditure and Economic =rowt/3 </e <ur,is/
.ase( Journal of Economic and Social #esearch ) !1#3 %@-72.
Ndi0o( N.+. and E.F. -da/( !2@#. D0namics of monetar0 polic0 and povert0 in a small open econom03 </e
Nigerian e7perience( Ni&erian Journal of Economics and !e"elopment *atters( 2!*#( *-)$.
&,oDie( ..E.E( !1CC%#. VBuman .apital ?ormation for "roductivit0 =rowt/ in NigeriaO( Nigerian Economic and
?inancial 1eview( June( pp. **-%.
&lugbenga and &wo0e !27#( "ublic E7penditure and Economic =rowt/3 New Evidence from &E.D .ountries
W/ttp3LLiaes.confe7.comLiaesL1omeX)7Ltec/programLS1$$$.B<AY
&mo,e ". !2C#( =overnment E7penditure and National Income3 + .ausalit0 <est for Nigeria( European
Journal of Economic and Political Studies, 'o. 23 1-11.
1amse0( J.F. !1C)C# V<ests for Specification Errors in .lassical >inear >east S8uares 1egression +nal0sisV(
Journal of the #oyal Statistical Society, Series B.( @1!2#( @%P@71.
1ic/ter( .. and Dimitrios( ".!212#( </e 'alidit0 of 5agner6s >aw in -nited :ingdom. nternational Net+or$
for Economic #esearch wor,ing paper.
1ou7( +( !1CC*#. NDefence( Buman .apital and Economic Development in Sout/ +fricaO( +frican Defence
1eview( No 1C.
Seviten0i >. N. !212#( =overnment E7penditure and Economic =rowt/ in Nigeria3 +n Empirical Investigation
!1C)1-2C#( ,ournal of Economic Analysis,vol.@ !issue 1# pp @$-%1
Sing/ and Sa/ni !1C$*#( .ausalit0 between "ublic E7penditure and National Income( 'he #e"ie+ of Economics
and Statistics ))( )@-)**.
Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%
<ang( .. ?. !2C#. +n E7amination of t/e =overnment Spending and Economic =rowt/ Ne7us for Aala0sia
-sing t/e >everaged Footstrap Simulation +pproac/( =lobal Economic 1eview( @$!2#( 21%-227.
5agner +. !1$$@#( </ree E7tracts on "ublic ?inance( translated and reprinted in 1.+. Ausgrave and +.<.
"eacoc, !Eds#( Classics in the 'heory of Public %inance( >ondon3 Aacmillan W1C%$Y.

A))EN#@
E=T/ATE# !E=$'T=
!E=$'T "7 =E!A' %"!!E'AT"N
Freusc/-=odfre0 Serial .orrelation >A <est3
?-statistic .%@2 "robabilit0 .C*2@@$
&bs21-s8uared .%CC$ "robabilit0 .C@$2)7

<est E8uation3
Dependent 'ariable3 1ESID
Aet/od3 >east S8uares
Date3 %L7L1* <ime3 213%)
"resample missing value lagged residuals set to Eero.
'ariable .oefficient Std. Error t-Statistic "rob.
. -2$2C$). **)C$$)$ -.)@2C2 .C*CC
"&"1 1.2C)$*7 1C.7C*%* .)%%1% .C*$1
<=E1 -1.@%))C *1.71* -.@2@$2 .C7*@
=?.?1 .2)*1$@ 1%.12%* .17*)7 .C$)2
1ESID!-1# .$7*@* 1.2$1$ .72$12 .C*2@
1-s8uared .1@C Aean dependent var $.))E-C
+dDusted 1-s8uared -.1%1C S.D. dependent var @$22%@@$
S.E. of regression *1$*)) +,ai,e info criterion @7.C)*7$
Sum s8uared resid ).1*EJ1) Sc/warE criterion @$.1)C%7
>og li,eli/ood -$11.2*2$ ?-statistic .1@2%
Durbin-5atson stat 1.C))%7$ "rob!?-statistic# .CCCCC)
Source: computed by the author
A!%H !E=$'T
+1.B <est3
?-statistic .1)@17 "robabilit0 .$C$CC%
&bs21-s8uared .1712) "robabilit0 .$C%$$

<est E8uation3
Dependent 'ariable3 1ESIDZ2
Aet/od3 >east S8uares
Date3 %L12L1* <ime3 231
Sample!adDusted#3 1C71 212
Included observations3 *2 after adDusting endpoints
'ariable .oefficient Std. Error t-Statistic "rob.
. 1.*CEJ1% 1.@EJ1% 1.1*C$C .2)7
1ESIDZ2!-1# -.21$$ .1%$@C -.1277* .$CC
1-s8uared .*$ Aean dependent var 1.*)EJ1%
+dDusted 1-s8uared -.2*%$2 S.D. dependent var $.21EJ1%
S.E. of regression $.@1EJ1% +,ai,e info criterion 7).1C)2%
Sum s8uared resid 2.7)EJ@@ Sc/warE criterion 7).27$CC
>og li,eli/ood -1%C$.121 ?-statistic .1)@17
Durbin-5atson stat 1.CCC)$ "rob!?-statistic# .$C$CC%





Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%1
'NEA! !E=$'T
Dependent 'ariable3 =D"1
Aet/od3 >east S8uares
Date3 )L1L1* <ime3 113*
Sample!adDusted#3 1C7 211
Included observations3 *2 after adDusting endpoints
'ariable .oefficient Std. Error t-Statistic "rob.
. -7C*$C) 27)**C1* -2.%@%%** .1%)
<=E1 2$.1$7)% @@.1C@ .$%1%)7 .@CCC
=?.?1 -@.*@)72C .%%%7$ -).7C7)2 .
"&"1 @1.1*)) 11.1)$* 2.7$$$%7 .$@
AS1 -1$.C21) 1@.))@** -1.@2*12C .1C@)
1-s8uared .%7$)12 Aean dependent var $C@7C2.1
+dDusted 1-s8uared .%@@%7 S.D. dependent var %7)*1CC1
S.E. of regression @C@$$)@ +,ai,e info criterion @7.C272
Sum s8uared resid %.7*EJ1) Sc/warE criterion @$.1@*)
>og li,eli/ood -7C1.*711 ?-statistic 12.712C
Durbin-5atson stat 1.C*)$ "rob!?-statistic# .1

E@)"NENTA' !E=$'T
Dependent 'ariable3 >&=!=D"1#
Aet/od3 >east S8uares
Date3 )L@L1* <ime3 213*)
Sample3 1C7 212
Included observations3 *
E7cluded observations3 @
'ariable .oefficient Std. Error t-Statistic "rob.
. .*$2@$1 .$*7%*$ .%)C1*$ .%72C
"&"1 *.2CE-) @.)E-7 11.C1C7 .
<=E1 7.@CE-7 1.22E-) .)%@2 .%*C1
=?.?1 -1.)%E-) ).*2E-7 -2.%71*)1 .1*%
AS1 -1.$E-) *.%1E-7 -2.@C@2 .22@
1-s8uared .$)$)71 Aean dependent var 11.1)*@7
+dDusted 1-s8uared .$%@))2 S.D. dependent var @.1)1C
S.E. of regression 1.1$$2*) +,ai,e info criterion @.2CC@2
Sum s8uared resid *C.*17*7 Sc/warE criterion @.%1*12
>og li,eli/ood -).C$)@ ?-statistic %7.$7)7
Durbin-5atson stat 1.CC)** "rob!?-statistic# .


Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%2
#"$>'E '"G !E=$'T
Dependent 'ariable3 >&=!=D"1#
Aet/od3 >east S8uares
Date3 )L@L1* <ime3 2@321
Sample3 1C7 212
Included observations3 2%
E7cluded observations3 1$
'ariable .oefficient Std. Error t-Statistic "rob.
. -7.2)C$2 1.)$%$ -.)$)%) .%@C
>&=!"&"1# @.1*@C1 *.2$$7C .7@7%$1 .*)C@
>&=!<=E1# .**21 .11$%21 .@*1*7 .7@))
>&=!=?.?1# .$C*%C .))C@C 1.@@)*1% .1C)*
>&=!AS1# .*%@1@2 .2%1)C2 1.$@@C .$)C
1-s8uared .$@7@ Aean dependent var 2.*@)1%%
+dDusted 1-s8uared .$*** S.D. dependent var .27*C1
S.E. of regression .11C)2$ +,ai,e info criterion -1.2@1CCC
Sum s8uared resid .2$)21$ Sc/warE criterion -.C$$22*
>og li,eli/ood 2.@CCCC ?-statistic 2%.)7%*2
Durbin-5atson stat 2.@C%%1 "rob!?-statistic# .


N"!/A'T< TE=T 7"! E@)"NENTA' EA$AT"N
0
1
2
3
4
5
6
7
8
9
-3 -2 -1 0 1 2
Series: Residuals
Sample 1970 2010
Observations 39
ean -3!99"-16
edian 0!010924
a#imum 2!044345
inimum -2!854762
Std! $ev! 1!130605
S%e&ness -0!489740
'urtosis 3!570970
(ar)ue-*era 2!088759
+robabilit, 0!351910

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%@
N"!/A'T< TE=T 7"! #"$>'E '"G EA$AT"N
0
1
2
3
4
5
6
7
8
-0!3 -0!2 -0!1 0!0 0!1 0!2 0!3
Series: Residuals
Sample 1970 2010
Observations 26
ean 0!001343
edian 0!011723
a#imum 0!255995
inimum -0!257768
Std! $ev! 0!106562
S%e&ness -0!104623
'urtosis 3!663331
(ar)ue-*era 0!524108
+robabilit, 0!769469


WHTE HETE!"=KE#A=T%T< TE=T 7"! E@)"NENTA' EA$AT"N
5/ite Beteros,edasticit0 <est3
?-statistic 1*.%2@$* "robabilit0 .
&bs21-s8uared @2.7C$$ "robabilit0 .)C

<est E8uation3
Dependent 'ariable3 1ESIDZ2
Aet/od3 >east S8uares
Date3 )L%L1* <ime3 73%2
Sample3 1C7 211
Included observations3 *2
'ariable .oefficient Std. Error t-Statistic "rob.
. *.CEJ1) $.C*EJ1% *.%77%7@ .1
<=E1 -$.C$EJC 1.)@EJ1 -.%%1%)) .%$%
<=E1Z2 @$1@.$* 12@%*.*7 .@$71 .7%C%
=?.?1 1.1%EJ1 %.27EJC 2.1$*$27 .@)1
=?.?1Z2 -1@).2$2% ).$%777 -2.2@C@) .@2
"&"1 -*.%EJ1 7.@CEJC -%.*7%*C% .
"&"1Z2 C$%$.@2) 1%7%.1%1 ).2%$)%* .
AS1 -@.$%EJ1 %.CCEJC -).*@712@ .
AS1Z2 $1@@.$$ 1)1).12$ %.@2*%@ .
1-s8uared .77$$7 Aean dependent var 1.@7EJ1%
+dDusted 1-s8uared .72%1$* S.D. dependent var 7.1EJ1%
S.E. of regression @.72EJ1% +,ai,e info criterion 7*.7@1C*
Sum s8uared resid *.%7EJ@2 Sc/warE criterion 7%.1*@
>og li,eli/ood -1%).@71 ?-statistic 1*.%2@$*
Durbin-5atson stat 2.1*%CC1 "rob!?-statistic# .

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%*

N"!/A'T< TE=T 7"! 'NEA! EA$AT"N
0
4
8
12
16
20
24
0!00000 1!0"-08 2!0"-08
Series: Residuals
Sample 1970 2011
Observations 42
ean 6!46"-09
edian -5213443!
a#imum 2!15"-08
inimum -32863893
Std! $ev! 37417934
S%e&ness 4!617888
'urtosis 27!35778
(ar)ue-*era 1187!552
+robabilit, 0!000000

=E/ '"G !E=$'T
Dependent 'ariable3 =D"1
Aet/od3 >east S8uares
Date3 )L%L1* <ime3 C3@@
Sample!adDusted#3 1C7 211
Included observations3 2$
E7cluded observations3 1* after adDusting endpoints
'ariable .oefficient Std. Error t-Statistic "rob.
. -).)@EJC 2.)@EJC -2.%17@1 .1C2
>&=!<=E1# -1$7)*$. 1727*7) -.1$@1% .C1*7
>&=!=?.?1# -2$%)%27 1)C$%7@ -2.))CC$% .1@7
>&=!"&"1# *.$EJ$ 1.CEJ$ 2.%2%C) .1$C
>&=!AS1# -172C%%1 1@*C)$C$ -1.2$17) .212C
1-s8uared .2*1$)C Aean dependent var 1177%).
+dDusted 1-s8uared .112 S.D. dependent var 712)7
S.E. of regression )7%1 +,ai,e info criterion @C.@$$7
Sum s8uared resid 1.@EJ17 Sc/warE criterion @C.27)7)
>og li,eli/ood -%*1.%**2 ?-statistic 1.$@***2
Durbin-5atson stat 2.C$2%$C "rob!?-statistic# .1%)%7%




Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%%
N"!/A'T< TE=T 7"! =E/'"G EA$AT"N
0
2
4
6
8
10
12
14
-2!0"-08 0!00000 2!0"-08
Series: Residuals
Sample 1970 2011
Observations 28
ean -1!89"-06
edian -1147539!
a#imum 2!06"-08
inimum -1!89"-08
Std! $ev! 61842907
S%e&ness 0!283374
'urtosis 8!256879
(ar)ue-*era 32!61531
+robabilit, 0!000000

WHTE HETE!"=KE#A=T%T< TE=T 7"! =E/'"G EA$AT"N
5/ite Beteros,edasticit0 <est3
?-statistic $.1)C))% "robabilit0 .C
&bs21-s8uared 21.)C@*C "robabilit0 .%%1)

<est E8uation3
Dependent 'ariable3 1ESIDZ2
Aet/od3 >east S8uares
Date3 )L%L1* <ime3 C3*1
Sample3 1C7 211
Included observations3 2$
E7cluded observations3 1*
'ariable .oefficient Std. Error t-Statistic "rob.
. 7.7%EJ1$ %.C@EJ1$ 1.@)$7% .2)C
>&=!<=E1# -@.%%EJ1) @.CEJ1) -1.1*C2$@ .2)*7
!>&=!<=E1##Z2 1.%)EJ1% 1.@2EJ1% 1.1$2C1C .2%1*
>&=!=?.?1# -).)@EJ1% 2.7@EJ1% -2.*@27@ .2%1
!>&=!=?.?1##Z2 2.)$EJ1* 1.1%EJ1* 2.@2$1)% .@11
>&=!"&"1# -1.1@EJ1$ $.1@EJ17 -1.@$71@@ .1$1%
!>&=!"&"1##Z2 *.1)EJ1) 2.7*EJ1) 1.%17%1% .1*%)
>&=!AS1# 1.C@EJ1) ).7@EJ1% 2.$)C))C .C$
!>&=!AS1##Z2 -1.2*EJ1% @.*%EJ1* -@.%C2CC .2
1-s8uared .77*7)$ Aean dependent var @.)CEJ1%
+dDusted 1-s8uared .)7CC@@ S.D. dependent var 1.1EJ1)
S.E. of regression %.72EJ1% +,ai,e info criterion 7%.)%C7)
Sum s8uared resid ).22EJ@2 Sc/warE criterion 7).$7C7
>og li,eli/ood -1%.2@7 ?-statistic $.1)C))%
Durbin-5atson stat @.C%1%2 "rob!?-statistic# .C

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%)
WHTE HETE!"=KE#A=T%T< TE=T 7"! E@)"NENTA' EA$AT"N
5/ite Beteros,edasticit0 <est3
?-statistic 1.$@)2 "robabilit0 .11)@%1
&bs21-s8uared 12.)%@7 "robabilit0 .12**%C

<est E8uation3
Dependent 'ariable3 1ESIDZ2
Aet/od3 >east S8uares
Date3 )L%L1* <ime3 C3%@
Sample3 1C7 21
Included observations3 @C
E7cluded observations3 2
'ariable .oefficient Std. Error t-Statistic "rob.
. -2.@7)C)@ %.@)*7% -.*71C% .)**
"&"1 2.CCE-) *.21E-) .71%C$ .*$2$
"&"1Z2 -).*@E-1@ C.E-1@ -.71*1C2 .*$)
<=E1 *.1CE-) C.1%E-) .*%7%%* .)%)
<=E1Z2 -@.%@E-12 ).$$E-12 -.%1@71 .)117
=?.?1 -).$*E-) ).)@E-) -1.@2$$$ .@CC
=?.?1Z2 1.@)E-11 ).$E-12 2.7)2% .%@$
AS1 -@.7)E-7 @.@7E-) -.111)C1 .C11$
AS1Z2 2.)2E-1@ $.))E-1@ .@@) .7)@C
1-s8uared .@2*@)$ Aean dependent var 1.2*%*C2
+dDusted 1-s8uared .1**2 S.D. dependent var 2.2@1%C
S.E. of regression 1.$71)1* +,ai,e info criterion *.2C)%@
Sum s8uared resid 1%.$$1 Sc/warE criterion *.)7*%%2
>og li,eli/ood -7*.))77* ?-statistic 1.$@)2
Durbin-5atson stat 2.27%$7 "rob!?-statistic# .11)@%1



Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%7
/A%!"E%"N"/% (A!A>'E=
HE+1 =D"1 <=E1 =?.?1 "&"1 AS1
1C7 $C*.1 2@$ 2.@ 1272)*@ 1C$.@
1C71 1@)C.$ C7**1 .) 1@21$C 1C@1.C
1C72 %@).) -C%%2$ -1.1 1@7%%$) 27$%.%
1C7@ 1**@ 1%) 2.1 1*%)1@* 2*C.C
1C7* 11C2.) %@*2 -*.2 1%71%7 2C7.)
1C7% 2)%2.1* )*%2 $.* 17$%7) 1112.2
1C7) %1$.%* )2)$ -1).$ 1$)1@7$ )277.)
1C77 *$)*.%) 2%%1 @@.) 1CC$*) 1177C.%
1C7$ @1C.7) -2C@$ -)7.2 2$)7C7 1)$.1
1C7C 7*@*.) 2$7 1@*.* 21)C$@ 21)7.%
1C$ 7)%7.)2 *71 -2)$.$ 27$$$ 1$$@7.@
1C$1 -212.)) )$C$.7 %@7.2C 2@1*7% 1C$.@
1C$2 1**C.)2 %C.% -17*.77 22@1 1C@1.C
1C$@ *@$.1 -22$).7 -@$1.*C 1CCC%$ 27$%.%
1C$* )%1%.1% 2C1.1 -*1$%.%7 2*%C* 2*C.C
1C$% $2$).2 @11@.% -@%.2$ 212)@%% 2C7.)
1C$) 12@$.** @1$2.) 2%%1.C$ 221)*7* 1112.2
1C$7 @)7%.$% %7C% @$77.12 22C*$7* )277.)
1C$$ @@$)2.*) %7@.$ 2@@@.)@ 2@))C7 1177C.%
1C$C 77712.2* 1@27$.$ C2)@.@ 2*)1*@ 1)$.1
1CC %7%2.*% 1C2@C.C 1@2C%.$ 2*@7%C 21)7.%
1CC1 **%$C.7% )@1).2 %)$.C2 2*)$2@ 1$$@7.@
1CC2 22*7*.C 2)21@ 2%)1$.C@ 2%)$)C *1%$%.7
1CC@ 1%12%%.C) C$*@1.% 2)1).@% 2%%2%7 )C@C@.7
1CC* 21%CC@.*@ -@@@%.7 $)%C.C$ 2)$*7 )$*)%.7
1CC% 1@@@*$.@@ $7$%2.C @)@**.7% 2)72@1 %1$1$.)
1CC) 7)C%7.%$ $$*71.% )2127.2 27*1@$@ %1%7
1CC7 CC2%@.*% CCC7.) @$$%2.@% 2$1@271 %C@C7.$
1CC$ -C@%*1.72 %$$C$.2 -)*@.%@ 2$$7$C C%C).%
1CCC *$%%$*.11 *)%7).) -1%C*.) 2C)*%17 17*C%.C
2 1@$$112.@2 -2*))@.) CC@C%.* @**$@C @@)@*%.$
21 1*2C%$.71 @1)C)).2 *17$.C2 @12$2)% 27C7$C.)
22 21$72C%.2% 1@.2 127%*%.* @21C)*C 2$@)2%.%
2@ 1%7*)%.@2 27$1.1 @))1C%.% @@2%%% @$%)C7.2
2* 2C2*@%.@* 22@*.1 -2$@.$* @**C1* 27$@C).1
2% @1)1172.21 @C%C -%$)71.$ @%$))*1 %%12%$.2
2) @CC2@%%.)1 11%C2.% 7*212*.$ @72C1$ 121@%)
27 2C2722.C@7 @C%)%7.1 @$$%1*.% @$72*** 1$*%$7
2$ @)@C11.)2 $%C7$.* 117722.% *2727 @@7%C7*
2C *1)@*.%C$* 2%C%%.$ CC%2).$ *172C* 1%721)%
21 2)72C@11@.1 7*1227.$ C%C$C *@2)7) 7**C@.2
211 2%**)212$.@ @@C%)%)7.@ $%)C2C$$ **$%1*% 17C@*)
212 1%CC1$.@ 1)1*@*1.%$ 2$$$)% *)*$%1 2%

Journal of Economics and Sustainable Development www.iiste.org
ISSN 2222-17 !"aper# ISSN 2222-2$%% !&nline#
'ol.%( No.1)( 21*

%$
DES.1I"<I'E S<+<IS<I.
Date3 )L)L1* <ime3
$3@)

Sample3 1C7 212
=D"1 <=E1 =?.?1 "&"1
Aean $C@7C2.1 1)7$C.2 21@%@C. 2)@1$*.
Aedian *@@2.$ 7C@)1. 2**2.$% 2*%2$).
Aa7imum 2.)7EJ$ 1272)*@. $%)C2C$$ **$%1*%.
Ainimum -2.%*EJ$ -*7*%. -%$)71.$ 1272)*@.
Std. Dev. %7)*1CC1 1C1%*1.* 1@2CC27 $*C%.1
S,ewness .2C)7@ %.%221@ ).2*@)% .%2C$)@
:urtosis 21.217% @*.@2@72 @C.CC7** 2.%22*1

Jar8ue-Fera %)$.C)2% 1C@.%) 2))$.@ 2.@)***@
"robabilit0 . . . .@)%C7

Sum @7%@C2)$ **$%1*%. $C)$2C) 1.CEJ$
Sum S8. Dev. 1.@)EJ17 1.%EJ12 7.1%EJ1% 2.CEJ1@

&bservations *2 *2 *2 *2



The IISTE is a pioneer in the Open-Access hosting service and academic event
management. The aim of the firm is Accelerating Global Knowledge Sharing.

More information about the firm can be found on the homepage:
http://www.iiste.org

CALL FOR JOURNAL PAPERS
There are more than 30 peer-reviewed academic journals hosted under the hosting
platform.
Prospective authors of journals can find the submission instruction on the
following page: http://www.iiste.org/journals/ All the journals articles are available
online to the readers all over the world without financial, legal, or technical barriers
other than those inseparable from gaining access to the internet itself. Paper version
of the journals is also available upon request of readers and authors.

MORE RESOURCES
Book publication information: http://www.iiste.org/book/

IISTE Knowledge Sharing Partners
EBSCO, Index Copernicus, Ulrich's Periodicals Directory, JournalTOCS, PKP Open
Archives Harvester, Bielefeld Academic Search Engine, Elektronische
Zeitschriftenbibliothek EZB, Open J-Gate, OCLC WorldCat, Universe Digtial
Library , NewJour, Google Scholar

You might also like