Table 5.
Variance inflation factors (VIFs) (Multicollinearity test).
Variable
VIF
y
3.73
y
2
2.62
y
3
4.85
The model stability has also been studied. It is checked on the basis of the CUSUM and CUSUMSQ
tests. As shown in Figure
3
, the cumulative sum of recursive residuals (CUSUM) test and the cumulative
sum squared of recursive residuals (CUSUMSQ) test, show that the estimated coefficients of the error
correction model are very stable throughout the studied period.
Energies 2018, 11, x FOR PEER REVIEW
11 of 18
and gas tariffs for high‐voltage electricity use. Nevertheless, it has not increased tariffs for
low‐voltage electricity, which is used in households [80]. The elimination of these subsides in the
residential sector would make electricity consumption more sensitive to prices, and therefore
responsible electricity consumption could be addressed. Fattouh and El‐Katiri [81] state that
electricity subsidies distort price signals, inducing large energy consumption inefficiencies.
Therefore, they have negative implications on electricity efficiency.
4.4. Multicollinearity and Stability
In order to avoid multicollinearity, all variables in the model were converted to deviation with
respect to their mean, as previously mentioned in the methodology section. Table 5 shows the VIF
values for the GDP per capita and its squared and cubed transformed values. The VIF values are
lower than 5, thereby no multicollinearity is present between y, y
2
and y
3
, in natural logs.
Dostları ilə paylaş: |