Obadi and Korček 2014 state that
�
Granger causes
�
if the coefficient values of are significantly different from zero, rejecting the null hypothesis of
= 0. Similarly,
�
Granger causes
�
if the coefficient values of are not zero. If both of these effects occur, then
�
and
�
have a two-way causal relationship.
5 RESULTS
In order to analyse the relationships between world crude oil prices and food commodity prices in Indonesia during the period January 2002 to August 2015 this
study has conducted a series of tests according to the methodology discussed in the previous section. Results of the empirical analysis are reviewed in this section.
5.1 Augmented Dickey-Fuller Unit Root Test
The augmented Dickey-Fuller ADF root test is a standard pre-test to a time series analysis aiming to check the presence of unit roots in the series. The test is
carried out to each logarithm series starting from level to the first difference at the 5 level of significance. The estimated results of the ADF test are presented in
Table 5.1 below.
Table 5.1 Results of Augmented Dickey-Fuller Test
Variable Level
1
st
Difference ADF Statistic
Critical Values ADF Statistic
Critical Values Crude Oil
-2.378 -2.886
-6.537 -2.886
World Rice -1.969
-2.886 -7.265
-2.886 World Maize
-1.951 -2.886
-5.940 -2.886
World Wheat -2.316
-2.886 -6.575
-2.886 World Soybean
-2.474 -2.886
-6.883 -2.886
World Sugar -1.744
-2.886 -6.761
-2.886 World Palm Oil
-2.169 -2.886
-5.948 -2.886
Domestic Rice -0.507
-2.886 -8.442
-2.886 Domestic Maize
-0.588 -2.886
-7.596 -2.886
Domestic Wheat -1.797
-2.886 -4.283
-2.886 Domestic Soybeans
-1.712 -2.886
-6.422 -2.886
Domestic Sugar -1.780
-2.886 -5.873
-2.886 Domestic Cooking Oil
-2.600 -2.886
-6.197 -2.886
Domestic Petroleum -2.180
-2.886 -9.445
-2.886
Table 5.1 shows that the null hypothesis of the existence of unit roots cannot be rejected for each variable at level, meaning that the series are non-stationary at
the level at the 5 level of significance. This implies that these variables at level cannot be used for time series analysis because non-stationary data may produce a
spurious regression. Therefore, the unit root test is carried out in the first differenced logarithm.
As indicated in Table 5.1, results show that the null hypothesis can be rejected for all variables, implying that these variables are stationary and integrated of the
same order, I1. Therefore, the test of cointegration can be carried out in order to detect the existence of long-run relationships between the variables.
5.2 Johansen Cointegration Test
The objective of the Johansen cointegration test is to determine whether a group of variables which are non-stationary at level, is able to fulfill a necessary