Page 13 - AEI Insights 2018 Vol 4 Issue 1

AEI Insights, Vol 4, Issue 1, 2018

where yt is k-vector of variables, p is lag order, μ is intercept, βi are slope coefficients and εt is
error term. Equation (3) can be transformed into:

p
 y     y t1    i  y t i   (4)
t
t
i1
where Π is coefficient matrix for the lagged dependent variables and Γ is the coefficient matrix
for the lagged differenced dependent variables. Furthermore, the coefficient matrix Π can be
written as:

    (5)

where α is the vector of adjustment parameter and β is the vector of cointegrating equation. In
this study, the maximum eigenvalue method was used to estimate the Johansen test statistic.
Due to the limited number of observations, the lag order was set up as one in the current study.

In the third stage of empirical analysis, the Granger causality test was employed to analyse the
causal relationship between variables. The Granger causality test could be based on the
following equations (Granger, 1969):

p p
i 
1 
GDP     i 1 EXP   i 2 GDP  
t
t
t 1
t
i
i 1 i 1

i 
2 
EXP    p  i 3 GDP  p  i 4 EXP   (6)
t
t
t
t 2
i


i 1 i 1

There were two null hypotheses in the current study. The first null hypothesis of no causality
from exports to economic development can be formulated as:

 11   12  .......... ...  1p  0 (7)

The second null hypothesis of no causality from economic development to exports can be
formulated as:

 21   22  .......... ...  2 p  0 (8)

The Granger causality test was used to test four alternative hypotheses that would explain
causal relations between variables. The first hypothesis is on export-led economic development
(ELED). According to this hypothesis, exports can be considered as a source of economic
development. The ELED hypothesis would be substantiated if the empirical findings could
point to the presence of a unidirectional causality from exports to economic development. The
rejection of the first null hypothesis (Equation 8) and non-rejection of the second hypothesis
(Equation 9) would provide empirical evidence to support this hypothesis.

The second hypothesis concerns development-led export expansion (DLEE). In other words,
economic development will lead to export expansion. The DLEE hypothesis would be
substantiated if the empirical findings could identify the presence of a unidirectional causality
from economic development to export expansion. The rejection of the second hypothesis

13