= + +
(4)
where is the first difference operation,
t
is stationary random error and
t
y is LGDP
t
, LCE
t
, LGE
t
,
LX
t
, LER
t
or LFDI
t
. The null hypothesis for this test is 0 = . If the null hypothesis 0 = cannot be
rejected, then the data set for
t
y contains unit root (not stationary).
0 1 1
2
p
t t i t i t
i
y y y
+
=
= + + +
(5)
where p is autoregressive level that is AR(p) which is suitably selected based on the smallest Schwarz
Information Criterion (SIC) value. The hypothesis 0 = will be rejected if is significant (negative).
While the equation for SIC is shown as follows:
log( )
2
l T
SIC k
T T
= +
(6)
where l is the value of the log of the likelihood function with the k parameters estimated using T
observations. The various information criteria are all based on -2 times the average log likelihood
function, adjusted by a penalty function.
An alternative approach is the Phillips-Perron (PP) method (Phillips and Perron, 1988). Rather
than taking account of the extra terms in the data-generating process (d.g.p) by adding them to the
regression model (as in ADF test), a non-parametric correction to the t-test statistic is under-taken to
account for autocorrelation that will be present when the underlying d.g.p. is not autoregressive at first
level, AR(1). Phillips and Perron (1988) propose an alternative (non-parametric) method of controlling
for serial correlation when testing for a unit root. The PP method estimates the non-augmented Dickey-
Fuller (DF) test equation (4), and modified the t-ratio of the coefficient so that serial correlation does
not affect the asymptotic distribution of the test statistic. The PP test procedure (a generalization of the
ADF test procedure) is based on the following regression
1 0 1 t t t
y y
= + +
(7)
and the statistic:
1 2
0 0 0
1 2
0 0
( )( ( ))
2
T f se
t t
f f s
=
%
(8)
where is the estimate, and t
(10)
where
0
f is an estimator of the residual spectrum at frequency zero and where ( ) S t is a cumulative
residual function:
1
( )
t
r
r
S t u
=
=
(11)
129 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
based on the residual
' (0)
t t t
u y x =
.
Although, the ADF and PP tests are routine and widely used in testing for unit roots in most
time series macroeconomic data, these tests are considered to have several problems, particularly in
terms of their low power and size distortions. Schwert (1989) finds that if y
t
has an ARMA
representation with a large and negative MA component, then the ADF and PP tests are severely size
distorted (reject I(1) null much too often when it is true) and that the PP tests are more size distorted
than the ADF tests. Caner and Killian (2001) have found similar problems with the KPSS test.
Recently, Perron and Ng (1996) have suggested useful modifications to the PP tests to mitigate this
size distortion.
The more robust and efficient unit root test, NP test was proposed by Ng and Perron in 2001.
Ng and Perron (2001) use the GLS detrending procedure of ERS (Elliot, Rothenberg and Stock, 1996)
to create efficient versions of the modified PP tests of Perron and Ng (1996). These efficient modified
PP tests do not exhibit the severe size distortions of the PP tests for errors with large negative MA or
AR roots, and they can have substantially higher power than the PP tests especially when is close to
unity. Ng and Perron (2001) constructed test statistics that are based upon the Generalized Least
Squared (GLS) detrended data y
t
.
Using the GLS detrended data
d
t
y , the efficient modified PP tests are defined as
1
1 2 2
1
1
( ) 2
T
d d
t t
t
MZ T y T y
=
=
(12)
1 2
2 2
1
1
/
T
d
t
t
MSB T y
=
=
(13)
t
MZ MZ MSB
=
(14)
The statistics MZ
and MZ
t
are efficient versions of the PP Z
and Z
t
tests that have much
smaller size distortions in the presence of negative moving average errors. Ng and Perron (2001) also
stress that good size and power properties of the all the efficient unit root tests rely on the proper
choice of the lag length, p. They argue, however, that traditional model selection criteria such as AIC
are not well suited for determining p with integrated data. Ng and Perron (2001) suggested the
Modified Akaike Information Criteria (MAIC) instead of AIC in determining the lag length p by
minimizing the value of MAIC.
This study then used the Johansen approach to test the movement of the variables in the long-
run. Only variables of the same integration level are taken into account in testing for the presence of
cointegration. Johansens cointegration test is based on the maximum likelihood estimation towards k-
dimensional vector autoregression (VAR) at level p,
1 1 2 2 1 1 1
...
t t t k t p t t
Z Z Z Z Z
+ +
= + + + + +
(15)
where Z
t
is k x 1 stochastic variable vector, is k x 1 constant vector,
t
is k x 1 random vector,
and
1 1
.......
+
k
is k x k parameter matrix. Meanwhile, if the coefficient matrix has reduced level, r <
k, therefore the matrix can be simplified to '. = The Johansen test involves level testing for matrix
by examining whether the eigen value for is significantly different from zero. There are some
conditions where it might which are: (i) r = k meaning
t
Z is stationary at the level, (ii) r = 0 meaning
t
Z is the autoregressive vector at first difference and (iii) 0 < r < k which means there is r linear
combination towards
t
Z which is stationary or cointegrated. This study also made use of the Trace (Tr)
eigen value statistic and Maximum (L-max) eigen value statistic (Johansen, 1988; Johansen and
Juselius, 1990). The Trace likelihood ratio test statistic is:
2
1
1
(1 )
p
r
Tr T In
+
=
)
(16)
130 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
where
p r
) )
,.......,
1 +
is the smallest eigen value on the estimated r p . The null hypothesis for the Trace
eigen value test is where there are most r cointegrating vectors. Meanwhile L-max can be calculated
as:
1
max (1 )
r
L TIn
+
=
)
(17)
The null hypothesis for Maximum eigen value that the r cointegrating vector is tested compared
to alternative hypothesis on cointegrating vector 1 + r . If the Trace eigen value and Maximum eigen
value tests result in different decisions, then the decision result of Maximum eigen value test should be
used because the force of Maximum eigen value test is bigger when compared to Trace eigen value test
(Johansen and Juselius 1990).
The causal relationship issue in this research is tested using the Error Correction Model (Engle
and Granger, 1987). In association with ECM, a principal feature of cointegrated variables is that their
time paths are influenced by the extent of any deviation from long-run equilibrium. If the system is to
return to the long-run equilibrium, the movements of at least some of the variables must respond to the
magnitude of the disequilibrium. The dynamic model implied by this study is the one of error
correction. In an error correction model, the short term dynamics of the variables in the system are
influenced by the deviation from equilibrium.
Generally, time series variables which are not stationary should not be applied in the regression
model to avoid problems of spurious regression. There are some exceptions for this case where y
t
and
x
t
are non-stationary variables I(1), we expect that at the discriminating level or any linear combination
amongst those variables for example
0 1 t t t
y x = would also integrate at I(1). Even so, there are
more important cases that exist where
0 1 t t t
y x = are the stationary process of I(0). In such case,
y
t
and x
t
are also said to be cointegrated. Based on the cointegration test, if both
t
y and
t
x cointegrate,
then the said cointegration vector must be used as the error correction element (error correction term)
in modelling a short run relationship.
0 1 2
1 1
n n
t i t i j t j t
i j
y y x u
= =
= + + +
(18)
0 1 2
1 1
m m
t i t i j t j t
i j
x y x v
= =
= + + +
(19)
Generally, in the case where
t
y and
t
x are stationary variables I(0), equation (18) and (19) can
be estimated using the least squares method. However, if y
t
and x
t
are non-stationary variables, I(1) and
do not cointegrate, the VAR model such as in equation (20) and (21) in the first difference form can be
used. Whereas equation (22) and (23) can be used in cases where
t
y and
t
x are I(1) and cointegrated.
0 1 2
1 1
n n
t i t i j t j t
i j
y y x u
= =
= + + +
(20)
0 1 2
1 1
m m
t i t i j t j t
i j
x y x v
= =
= + + +
(21)
where
t
u and
t
v are stochastic errors. n and m are the selected lag length based on SIC. x
t
is Granger
causes y
t
if and only if the total of
2 j
in equation (20) is significant and the total of
1i
in equation
(21) is not significant. On the other hand, y
t
is Granger causes to x
t
if the total of
2 j
in equation (20)
is not significant and the total of
1i
in equation (21) is significant. If both the totals of
2 j
and
1i
are
significant, then there exists a bilateral causal relationship between y
t
and x
t
.
Meanwhile, if we assume all variables are I(1), in the case where y
t
and x
t
are cointegrated, the
following equation shall be used:
0 1 2 3 1
1 1
n n
t i t i j t j t t
i j
y y x z u
= =
= + + + +
(22)
131 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
0 1 2 3 1
1 1
m m
t i t i j t j t t
i j
x y x z v
= =
= + + + +
(23)
where
1 t
z
> is due to
1 0 1 1
( ).
t t
y x
> + A
negative error correction coefficient value
3
( ) in the first equation which is equation (22) is to ensure
a decline in the variable y
t
(y
t
), while a positive error correction coefficient
3
( ) in the second
equation which is equation (23) is to ensure a rise in the variable x
t
(x
t
). The values of both error
correction coefficients are less than 1 in absolute value to ensure that the said equation system does not
explode.
4. Empirical Results and Findings
All variable trends tend to increase over time from the year 1970 to 2007, except for the exchange rate
and foreign direct investment (see Figure 1). The results of the unit root tests based on ADF, PP, KPSS
and NP are shown in Table 1. All variables are stationary at first difference (both constant, and
constant & trend) excluding the FDI which is stationary at level (constant & trend), but stationary at
first difference if only constant taken into account (based on ADF and PP tests).
KPSS statistics under the null hypothesis of stationary also confirm that all variables have the
tendency to be stationary at first difference if only constant taken into account, while for constant and
trend, only the GDP, CE and GE showed stationary at first difference. Both the X and FDI on the other
hand are stationary at level, and for the ER is non stationary at both level and first difference.
NP test (only MZ
and MZ
t
statistics are shown in Table 1) is more robust and produced
interesting results. Although both statistics MZ
and MZ
t
produced mix results, both statistics have
tendency to reject the null hypothesis of a unit root at first difference (without concerned about the
assumption of constant, and constant & trend). The results using MZ
and MZ
t
tend to support
stationary for all variables at first difference (constant and trend), except for GE which is stationary
under the assumption of constant without trend.
Multiple cointegration tests using the Johansen approach is shown in Table 2. The results show
that there is more than one cointegrating vectors whereby there are four cointegrating vectors between
the GDP and other variables. In this study, we are not discuss more on the interpretation of these
number of cointegrating vectors due to the problems of interpretation when dealing more than one
cointegrating vector (see Maddala and Kim, 1998: 233-242). Nevertheless, this empirical decision
proposes that GDP and other variables could be cointegrated in the long run. Both statistic tests which
are Trace-eigen and Maximum-eigen statistic produced similar results. This gives the conclusion that
in the long-term GDP moves together with other variables towards equilibrium.
For causality issues, only one way direction is presented here from the determinant factors (as a
whole or individually) to the economic growth. The existence of a long-run cointegration relationship
needs equation (22) and (23) to be used as the cointegration vector or error correction elements need to
be inserted into the model. Equations (22) and (23) are Error Correction Model (ECM) which are said
132 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
to be based on the VARs framework. The VAR lag length order selections based on SIC suggest that
the optimum lag length is 1. The results of causality such as in Table 3 is quite interesting whereby
there is causal relationship running from the CE to the GDP as well as from X to GDP whereby all
variables as a whole are Granger caused to the GDP especially in the short-run. Meanwhile, the error
correction coefficient is also significant at the 10 per cent level showing that the GDP are responsive to
the discrepancy in error correction from long-run equilibrium (see Table 3). The negative sign of the
speed of adjustment coefficient are in accord with convergence toward the long-run equilibrium. The
imbalance between the GDP and all the determinant factors in this study is corrected or adjusted at a
fast pace of 0.93 or 93 per cent. In addition, the diagnostic tests (see Table 3) show that the residual
follow the normal distribution, no serial correlation, no heteroskedasticity, lastly there is no
autoregressive conditional heteroskedasticity.
These results lead to the conclusion that consumption expenditure and export play important
roles as determinant factors to a countrys economic growth especially in the short-run. In addition,
government expenditure, exchange rate and foreign direct investment are not the main important
determinant factors to the economic growth in Malaysia. As a whole, all the variables (combined
determinant factors) tend to cause the economic growth.
Figure 1: All Variables Trends
0
100000
200000
300000
400000
500000
600000
700000
800000
1970 1975 1980 1985 1990 1995 2000 2005
GDP
CE
GE
X
ER
FDI
M
Y
R
M
i
l
l
i
o
n
&
M
Y
R
/
U
S
D
Year
133 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
Table 1: Unit Root Tests
Test Type Variable
Level First Difference
Constant Constant & Trend Constant Constant & Trend
ADF
LGDP -1.59 (0) -2.18 (0) -4.99***(0) -5.19***(0)
LCE -0.97 (1) -2.85(1) -4.34***(0) -4.35***(0)
LGE -1.09 (0) -1.99(0) -5.88***(0) -6.03***(0)
LX -0.98 (0) -2.11(0) -6.11***(0) -6.26***(0)
LER -1.23 (0) -2.85(0) -5.98***(0) -5.95***(0)
LFDI -1.63 (1) -3.49(0) -8.16***(0) -8.09***(0)
PP
LGDP -1.55 [1] -2.22[1] -4.96***[2] -5.15***[2]
LCE -1.13 [1] -2.31[1] -4.26***[3] -4.27***[3]
LGE -1.10 [1] -1.99[1] -5.90***[2] -6.04***[2]
LX -1.14 [6] -2.11[0] -6.11***[0] -6.44***[2]
LER -1.23 [0] -2.83[3] -5.99***[4] -5.97***[5]
LFDI -2.20 [2] -3.54*[3] -8.18***[1] -8.10***[1]
KPSS
LGDP 0.75*** [5] 0.14*[4] 0.24[2] 0.07[1]
LCE 0.75*** [5] 0.14*[4] 0.19[0] 0.07[1]
LGE 0.74*** [5] 0.13*[4] 0.17[1] 0.10[1]
LX 0.74*** [5] 0.09[4] 0.14[3] 0.07[5]
LER 0.50** [5] 0.16**[4] 0.20[3] 0.12*[4]
LFDI 0.70** [5] 0.10[4] 0.13[2] 0.06[2]
NP
MZ
MZ
t
MZ
MZ
t
MZ
MZ
t
MZ
MZ
t
LGDP
0.19
(3)
0.10
(3)
-3.96
(0)
-1.37
(0)
-18.07**
(0)
-2.98**
(0)
-18.01**
(0)
-2.97**
(0)
LCE
-0.02
(3)
-0.00
(3)
-4.50
(0)
-1.49
(0)
-16.91**
(0)
-2.90**
(0)
-17.21*
(0)
-2.92**
(0)
LGE
0.13
(2)
0.07
(2)
-4.31
(0)
-1.47
(0)
-8.25**
(1)
-2.01**
(1)
-10.42
(1)
-2.28
(1)
LX
0.46
(4)
0.28
(4)
-8.53
(0)
-1.95
(0)
-3.06
(3)
-1.24
(3)
-16.64*
(0)
-2.87*
(0)
LER
-3.05
(0)
-1.20
(0)
-5.44
(0)
-1.65
(0)
-5.68
(6)
-1.67*
(6)
-19.33**
(1)
-3.07**
(1)
LFDI
0.34
(1)
0.21
(1)
-7.29
(1)
-1.90
(1)
-10.57**
(6)
-2.29**
(6)
-42.36**
(7)
-4.60**
(7)
Notes: Figures in () and [ ] indicates number of lag and bandwidth structures respectively. *, **, *** indicates
significance at 10%, 5% and 1% levels respectively.
Table 2: Johansen Multivariate Cointegration Test
H
0
H
1
Test Statistic:
Trace Statistic:
trace
r = 0 r > 0 267.9052***
r 1 r > 1 168.5012***
r 2 r > 2 90.9044***
r 3 r > 3 40.7252***
r 4 r > 4 9.8093
r 5 r > 5 1.1328
Max-Eigen Statistic:
max
r = 0 r = 1 99.4040***
r = 1 r = 2 77.5968***
r = 2 r = 3 50.1792***
r = 3 r = 4 30.9159***
r = 4 r = 5 8.6765
r = 5 r = 6 1.1328
Notes: *** denote significant at the 1% level.
134 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
Table 3: Error Correction Model: Multivariate and Bivariate Causality Test
Variable VAR(k) Wald Statistic
Dependent Variable: LGDP
All LGDP
1
14.8737**
LCE LGDP
8.6597***
LGE LGDP
1.7954
LX LGDP
5.2877**
LER LGDP
0.0696
LFDI LGDP
0.3869
Variable VAR (k) t-Statistic
1 t
z
1 -0.9720*
Diagnostic Tests
R
2
= 0.3801 SIC = -2.1592 JB = 0.6320
White Statistic = 17.6841 Q(1) = 0.0018 LM(1) = 0.0062
Sum of Squared Residual = 0.1097 Q
2
(1) = 2.1550 ARCH(1) = 2.6847
Notes: ***, **, * denote significant and rejected at the 1%, 5% and 10% levels respectively. Q = Q-statistic for
correlogram of residuals, Q
2
= Q-statistic for correlogram of squared residuals, JB = Jarque-Bera statistic, LM =
Breusch-Godfrey Serial Correlation LM statistic, and ARCH = Autoregressive Conditional Heteroskedasticity
statistic. Figures in () indicates number of lag structures selected based on the SIC. The results for multi-ways
causal relationship between GDP and other variables (determinant factors) are available from the authors upon
request
5. Discussion and Conclusion
Economic stability is given special attention by many countries. This is because a stable economy
portrays a positive image and good economic positioning. Strong economic stability also becomes an
attraction for other countries to invest and also act as a guarantee on investments made. It can therefore
be said that the deciding factor on whether a country is successful or not can be seen from its economic
stability. Hence, factors determining economic stability must be given due attention. The determinant
factors such as exports, foreign direct investment, government expenditure, domestic consumption, and
exchange rate can be a threat to the stability of country if they are not managed well.
One might consider that by improving institutions or the way of how economy operates, we can
change our economic outcomes for the better. When institutions are weak, even places with abundant
natural resources or other inputs will not promise high and sustain economic growth. Good governance
and well managed economic resources are also important in order to sustain economic growth.
Economic growth is the result of a variety of influencing factors, which can only be
approximate by growth theory. Historically, the simple growth models were extended over time by
relaxing the model restrictions and supplementing new variables over time (as in this study) to give a
better explanation on economic growth. This study also confirmed empirically the success of the
export-led growth strategy.
Our research showed that consumption expenditure and exports play an important role in
boosting economic growth in Malaysia. More emphasis should be accorded on these determinant
factors when drafting related economic policies of a country. Our study also found that the effect and
role of government expenditure, exchange rate and foreign direct investment on economic growth may
be less important. Nevertheless, this does not mean that the importance of these factors in spurring
continuous economic growth should be ignored. Rather, these variables may be viewed as a catalyst
and complement factors of economic growth.
Our findings have important implications on policy reform. Policies to promote economic
growth should be based on evidence of what has, and what has not, worked for other countries.
Evidences presented here compared to previous studies (in the literature review) are mixed and that
economic growth is significantly influenced by various factors. In addition, while other reasons have
135 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
been suggested in the literature, but the questions that one might concern with are: how can we
replicate this in Malaysia? Can we uncover which are the best policies tending to promote economic
growth? Should we suggest that the consumption expenditure and the export are the best determinant
factors of economic growth? These questions are still pending and need further investigation.
References
1] Ahmad, J. and Harnhirun, S., 1996, Cointegration and causality between exports and economic
growth: evidence from the ASEAN countries. Canadian Journal of Economics, XXIX, pp.
S413-6.
2] Agu, C.C. and Chukwu, J.O., 2008, Toda and Yamamoto causality tests between bank based
financial deepening and economic growth in Nigeria. European Journal of Social Science, 7(2),
pp. 189-98.
3] Alfaro, L., Chanda, A., Kalemli-Ozcan, S. and Sayek, S., 2004, FDI and economic growth: the
role of local financial markets. Journal of International Economics, 64, pp. 89-112.
4] Ang, J.B., 2008, Determinants of foreign direct investment in Malaysia. Journal of Policy
Modeling, 30, 185-89.
5] Awokuse, T.O. (2002). Is the export-lead growth hypothesis valid for Canada? FREC Staff Paper
02-01, March. Department of Food and Resource Economics, University of Delaware.
6] Baharumshah, A.Z. and Thanoon, M.A-M., 2006, Foreign capital flows and economic growth in
East Asian countries. China Economic Review, 17, pp. 70-83.
7] Balaguer, J. and Jorda, M. C., 2001, Examining the export-led growth hypothesis for Spain in
the last century. Applied Economics Letters, 8, pp. 681-85.
8] Barro, R. and Sala-i-Martin, X., 1995, Economic Growth. McGraw-Hill: New York.
9] Barro, R.J., 1996a, Determinants of Economic Growth: A Cross-Country Empirical Study.
National Bureau of Economic Research (NBER) Working Paper No. 5968, 1(1). The MIT
Press: 1-79. [Online; cited October 2009] Available http://www.nber.org/papers/w5698.pdf.
10] Barro, R.J., 1996b, Health and Economic Growth. Health, Human Capital and Economic
Growth. Pan American Health Organization (PAHO). [Online; cited October 2009] Available:
http://www.paho.org/English/HDP/HDD/barro.pdf.
11] Bebczuk, R.N., 2000, Productivity and saving channels of economic growth as latent variables:
an application of confirmatory factor analysis. Estudios de Economia, 27(2).
12] Benhabib, J. and Spiegel M.M., 1994, The Role of Human Capital in Economic Development
Evidence From Aggregate Cross-Country Data. Journal of Monetary Economics, 34(2), pp.
14373.
13] Caner, M. and Kilian, L., 2001, Size distortions of tests of the null hypothesis of stationary:
evidence and implications for the PPP debate. Journal of International Money and Finance, 20,
pp. 639-57.
14] Central Bank of Malaysia, 1999, The Central Bank and the Financial System in Malaysia-A
Decade of Change, First Edition, Central Bank of Malaysia: Kuala Lumpur.
15] Chakraborty, C. and Nunnenkamp, P., 2008, Economic reforms, FDI, and economic growth in
India: a sector level analysis. World Development, 36(7), pp. 1192-212.
16] Chen, H., 2009, A literature review on the relationship between foreign trade and economic
growth. International Journal of Economics and Finance, 1(1), pp. 127-30.
17] Chow, P.C.Y., 1987, Causality between export growth and industrial development. Journal of
Development Economics, 26, pp. 55-63.
18] Cortes-Jimenez, I. and Pulina, M., 2006, Tourism and growth: evidence for Spain and Italy.
46th Congress of the Europian Regional Science Association, University of Thessaly (Volos,
Greece), 30 August 3 September.
136 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
19] Dickey, D. A., and Fuller, W. A., 1979, Distribution of the Estimation for Autoregressive
Time Series with a Unit Root, Journal of the American Association, 74, pp. 427-31.
20] Elliot, G., Rothenberg, T.J. and Stock, J.H., 1996, Efficient tests for an autoregressive unit root.
Econometrica, 64, pp. 813-36.
21] Engle, R.F. and Granger, C.W.J., 1987, Co-integration and error correction representation,
estimation and testing. Econometrica, 55(2), pp. 251-76.
22] Easterly, W. and R. Levine., 1999, Its not factor accumulation: stylized facts and growth
models. Mimeo, World Bank.
23] Galbraith, J. K., 1994, A journey through economic time. Houghton Mifflin: Boston, MA.
24] Government of Malaysia, 2007, Malaysia Economic Statistics Time Series. Department of
Statistics, Malaysia, December.
25] Government of Malaysia, 2008, Monthly Statistical Bulletin, Malaysia. Department of
Statistics, Malaysia, Mei.
26] Government of Malaysia, 2008, Monthly Statistical Bulletin. Central Bank of Malaysia,
December-July.
27] Government of Malaysia. Central Bank of Malaysia. [Online, cited January 2010] Available
http://www.bnm.gov.my/.
28] Hendry, D.F., Pagan, A.R. and Sargan, J.D., 1984, Dynamic specification. In: Handbook of
Econometrics II (eds Griliches, Z. and Intriligator, M.D.). North-Holland: Amsterdam.
29] Hsiao, F.S.T. and Hsiao, M.-C. W., 2006, FDI, exports, and GDP in East and Southeast Asia
panel data versus time series causality analyses. Journal of Asian Economics, 17, pp. 1082-106.
30] International Monetary Fund., 2006, CD International Financial Statistics: Database and
Browser, February.
31] Johansen, S., 1988, Statistical Analysis of Cointegration Vectors, Journal of Economic
Dynamics and Control, 12, pp. 231-54.
32] Johansen, S. and Juselius, K., 1990, Some Structural Hypotheses in a Multivariate
Cointegration Analysis of the Purchasing Power Parity and the Uncovered Interest Parity for
UK, Discussion Papers 90-05, Department of Economics, University of Copenhagen.
33] Jung, W.S. and Marshall, P.J., 1985, Exports, growth and causality in developing countries.
Journal of Development Economics, 18, pp. 1-12.
34] Khalafalla, K.Y. and Webb, A.J., 2001, Export-led growth and structural change: evidence
from Malaysia. Applied Economics, 33, pp. 1703-715.
35] Klenow, P. and Rodriguez-Clare, A., 1997, Economic growth: a review essay. Journal of
Monetary Economics, 40.
36] Krugman, P., 1994, The myth of Asias miracle. Foreign Affairs, 73(6).
37] Kwiatkowski, D., Phillips, P.C.B., Schmidt, P. and Shin, Y., 1992, Testing the null hypothesis
of stationarity against the alternative of a unit root. Journal of Econometrics, 54, pp. 159-78.
38] Li, X. and Liu, X., 2005, Foreign direct investment and economic growth: an increasingly
endogenous relationship. World Development, 33(3), pp. 393-407.
39] Maddala, G.S. and Kim, I-M., 1998, Unit roots, cointegration, and structural change.
Cambridge: Cambridge University Press.
40] Mankiw N.G., Weil, D.N. and Romer, D., 1992, A contribution to the empirics of economic
growth. Quarterly Journal of Economics, 107(2), pp. 407-37.
41] Nelson, C.R. and Plosser, C.R., 1982, Trends and random walks in macroeconomic time series:
some evidence and implications. Journal of Monetary Economics, 10(2), 139-62.
42] Ng, S. and Perron, P., 2001, Lag length selection and the construction of unit root tests with
good size and power. Econometrica, 69, pp. 1519-554.
43] Perron, P. and Ng, S., 1996, Useful modifications to some unit root tests with dependent errors
and their local asymptotic properties. Review of Economic Studies, 63, pp. 435-63.
137 European Journal of Economics, Finance and Administrative Sciences - Issue 24 (2010)
44] Phillips, P.C.B. and Perron, P., 1988, Testing for a unit root in times series regression,
Biometrica, 75, pp. 335-446.
45] Ramirez, A., Ranis, G. and Stewart, F., 1998, Economic Growth and Human Capital. QEH
Working Paper Series, No. 18. [Online; cited September 2009] Available:
http://www3.qeh.ox.ac.uk/pdf/qehwp/qehwps18.pdf.
46] Ricardo, D., 1817, The principles of political economy and taxation. J.M. Dent and Sons:
London.
47] Piazolo, M., 1996, Determinants of Indonesian economic growth, 1965-1992. Seoul Journal of
Economics, 9(4), pp. 269-98.
48] Sharma, A. and Panagiotidis, T., 2005, An analysis of exports and growth in India:
cointegration and causality evidence (1971-2001). Review of Development Economics. 9(2), pp.
232-48.
49] Sinha, D. and Sinha, T., 2007, Toda and Yamamoto causality tests between per capita saving
and per capita GDP for India. MRPA Paper No. 2564, 04 January. [Online; cited February
2010] Available http://mpra.ub.uni-muenchen.de/2564/.
50] Schwert, W., 1989, Test for unit roots: a Monte Carlo investigating. Journal of Business and
Economic Statistics, 7, pp. 147-59.
51] Sook-Ching, K., Kogid, M. and Furuoka, F., 2010, Causal relation between life insurance funds
and economic growth: evidence from Malaysia, ASEAN Economic Bulletin, 27(2): 185-99.
52] Tang, C.T., 2003, Bank lending and economic growth in Malaysia: revisited. Analisis, 10(1),
pp. 69-87.
53] Thirlwall, A.P., 1994, Growth and development with special reference to development
economic. Macmillan: Hampshire.
54] United Nations Conference on Trade and Development (UNCTAD). [Online; cited August
2009] Available http://www.unctad.org.
55] Yoo, S.-H., 2006, The causal relationship between electricity consumption and economic
growth in the ASEAN countries. Energy Policy, 34, pp. 3573-82.
56] Young, A., 1995, The tyranny of numbers: confronting the statistical realities of the East Asian
growth experience. Quarterly Journal of Economics, 110(3).