114
E. N. Udemba
From Eqs. 1 to 4 EFP, GDP, GDP
2 , FDI, AGR, and EU represent per capita
ecological footprint, per capita GDP and squared per capita GDP, foreign direct
investment, agriculture and energy use, respectively. Ecological footprint is a proxy
to measure the environmental dilapidation, GDP and GDP
2 are measures of economic
growth which are utilized for the sake of analyzing EKC hypothesis. Where t and μ it
represent time and error term,b 0 & a 0 represent the long-run and short-run coefficients
to be estimated. From Eq. (4) and ECM t−i denote sign of first difference and
error correction model. To test for cointegration, it is hypothesized that there is no
cointegration with null hypothesis as following:
Null hypothesis (H 0 = b 1 6 = 0) against alternative hypothesis (H 1 = b 1 6 = 0).
The cointegration is identified with comparing F-stat with the critical values of upper
and lower bounds in bound tests. If the F-stat is greater than the critical values at 1.5
or 10% significant level, cointegration is established and vice versa.
4 Empirical Result and Explanations
4.1 Descriptive Statistics
Descriptive analysis of the data was obtained in order to inspect the normality and
stability of data applied in this study. With the outcomes of Kurtosis and JarqueBera, some level of normality was confirmed. Hence, all the statistics under kurtosis
were below 3, and the probability of the Jarque-Bera outcome was all insignificant
except for the value of FDI which is insignificant in affecting the normality of the
entire data. Considering the variability of the variables in use, GDP is the most varied
indices followed by energy use. With the normality of the data, linear analysis will
be effective in this study. The output is shown in Table 1.
4.2 Stationarity Test
Stationarity of the data and order of integration of the variables were confirmed with
unit root test. Time series estimations are known with unstable variables because
of the likelihood of structural changes that may obstruct the stability of the trend in
economic variables. For effective analysis of these estimations, two approaches were
applied, hence, augmented Dickey–Fuller [16] and perron [30]. This will permit for
robust check among the two methods. Unit root was confirmed with mixed order of
integration among the variables. The unit root output paved way for the decision on
the appropriate method to apply further in testing the cointegration. The unit root
output is displayed on Table 2.
E. N. Udemba
From Eqs. 1 to 4 EFP, GDP, GDP
2 , FDI, AGR, and EU represent per capita
ecological footprint, per capita GDP and squared per capita GDP, foreign direct
investment, agriculture and energy use, respectively. Ecological footprint is a proxy
to measure the environmental dilapidation, GDP and GDP
2 are measures of economic
growth which are utilized for the sake of analyzing EKC hypothesis. Where t and μ it
represent time and error term,b 0 & a 0 represent the long-run and short-run coefficients
to be estimated. From Eq. (4) and ECM t−i denote sign of first difference and
error correction model. To test for cointegration, it is hypothesized that there is no
cointegration with null hypothesis as following:
Null hypothesis (H 0 = b 1 6 = 0) against alternative hypothesis (H 1 = b 1 6 = 0).
The cointegration is identified with comparing F-stat with the critical values of upper
and lower bounds in bound tests. If the F-stat is greater than the critical values at 1.5
or 10% significant level, cointegration is established and vice versa.
4 Empirical Result and Explanations
4.1 Descriptive Statistics
Descriptive analysis of the data was obtained in order to inspect the normality and
stability of data applied in this study. With the outcomes of Kurtosis and JarqueBera, some level of normality was confirmed. Hence, all the statistics under kurtosis
were below 3, and the probability of the Jarque-Bera outcome was all insignificant
except for the value of FDI which is insignificant in affecting the normality of the
entire data. Considering the variability of the variables in use, GDP is the most varied
indices followed by energy use. With the normality of the data, linear analysis will
be effective in this study. The output is shown in Table 1.
4.2 Stationarity Test
Stationarity of the data and order of integration of the variables were confirmed with
unit root test. Time series estimations are known with unstable variables because
of the likelihood of structural changes that may obstruct the stability of the trend in
economic variables. For effective analysis of these estimations, two approaches were
applied, hence, augmented Dickey–Fuller [16] and perron [30]. This will permit for
robust check among the two methods. Unit root was confirmed with mixed order of
integration among the variables. The unit root output paved way for the decision on
the appropriate method to apply further in testing the cointegration. The unit root
output is displayed on Table 2.