Augmented Dickey-Fuller test is based on the regression with the type:
Where α = ρ-1.
The null and alternative hypotheses can be written down as follows:
H0: α = 0.
H1: α < 0.
The test is based on t - statistics:
Where:
- the estimation of the α coefficient.
- a standard coefficient error.
Conclusions concerning hypotheses acceptance are based on unilateral MacKinnon distribution.
See also:
Library of Methods and Models | Tests for Unit Roots | ISmAugmentDickeyFullerTest