Page 16 - CITN 2017 Journal

Source: Authors' Computation, 2016

Figure 3: Trend Analysis of each of the variable

4.2.1 Unit Root Test Results
The purpose of the estimation of the unit root test is to determine the stationarity of the time
series data used for this study and check their order of integration.

Table 1: ADF Unit Root Test of Variables used in the Model
VAR ADF TEST ADF TEST MACKINNON CRITICAL ORDER
AT LEVEL AT ONE VALUE OF
DIFF 1% 5% 10% INTEGR
ATION

GEH 2.689555 -4.012163 -3.6576 -2.9591 -2.6181 I (1)
REM 3.674631 10.24628 -3.6576 -2.9591 -2.6181 I (1)
FDI -3.615980 -12.78160 -3.6576 -2.9591 -2.6181 I (1)
EGS 2.513989 -3.962861 -3.6576 -2.9591 -2.6181 I (1)
FX -0.077253 -3.89185 -3.6576 -2.9591 -2.6181 I (1)

Source: Authors' Computation, 2016

The Augmented Dickey Fuller (ADF) test in Table 1 reveals that GEH, REM, FDI EGS and
FX attain their orders of stationary at first difference. Thus, the stationarity tests show that
all the variables are integrated of order one. Hence, based on the empirical analysis of the
unit root test, it is necessary to check if the variables are co-integrated and that there is a
long run relationship between the dependent and the independent variables.

Table 2: Result of Johansen's Co-Integration Test

Likelihood 5 Percent 1 Percent Hypothesized
Eigenvalue Ratio Critical Value Critical Value No. of CE(s)
0.986940 264.3962 68.52 76.07 None **
0.898001 129.9126 47.21 54.46 At most 1 **
0.761097 59.14596 29.68 35.65 At most 2 **
0.367613 14.76332 15.41 20.04 At most 3
0.017822 0.557458 3.76 6.65 At most 4
Source: Authors' Computation, 2016

4.2.2 Johansen Co-Integration Test
Table 2 presents the Johansen co-integration test result. The result indicates the long run
relationship between the variables using maximum eigenvalue and the likelihood ratio
respectively. The study shows that the likelihood ratio is higher than the critical values at 5
and 1 per cent. Hence there are three (3) co-integrating factors in the result. This implies
that the null hypothesis of no co-integration among the variables should be rejected at one
and five percent levels of significance while the alternative hypothesis of co-integration in
9