Page 90 - Azerbaijan State University of Economics
P. 90
Zahra Huseynova: The Correlation Between Gender Pay GAP AND GDP Growth In Azerbaijan Economy
The one of the main issues for all types of OLS (Ordinary Least Square) estimators is
multicollinearity where independent variables are correlated in high level. This
problem may cause some bias estimation. For checking this problem, we look VIF
(Variable Inflation Factor) score of our variables. For this test, results of our control
variables are less than five which is quite satisfactory (see Table 5).
Table 5: Collinearity analysis
Source: The results are obtained through the author’s analysis
The last test we wanted to check is homoskedasticity. Homoskedasticity means error
term of the regression - u has same variances for all values of independent variables,
the reverse situation is called heteroskedasticity which is violation of assumptions of
Gauss-Markov Theorem-best linear unbiased estimator (Wooldridge, 2012).
However, heteroskedasticity is not an important issue for time series analysis yet we
wanted to check. Therefore, we referred to the White Test where null hypothesis is
homoskedasticity and we failed to reject homoskedasticity so there is no
heteroskedasticity (see Table 6).
Table 6: Heteroskedasticity Test
Source: The results are obtained through the author’s analysis
90
