One
of these methods is the visual inspection of
R2 and significance of t-ratios. When R2 is
high and there are few significant t-ratios,
we suspect the presence of multicollinearity.
Another method of detecting multicollinearity
that we adopted in this study is the
pairwise correlations among regressors (the
result based on this we presented in table 1).
If the pairwise correlations among two regressors
is in excess of 0.8, we suspect that
multicollinearity poses serious challenge
to our estimates (Gujarati, 2003).