Seasonal Augmented Dickey-Fuller Unit Root Test

[andreas_zaras]

Hello,

I have a question about the seasonal unit root test incorporated in proc arima that is named in SAS output as "seasonal augmented dickey fuller unit root test".

I will pose my question by providing an example.

I run the following proc arima procedure:

proc arima data=aaa;

identify var=y stationarity=(adf=(n) dlag=m); */ m, n are integers (e.g. n=4 and m=12);

run

quit;

In the ouput i get a table entitled "seasonal augmented dickey fuller unit root test".

The p-values that i get in the column Pr<Tau are produced by comparing the column tau values of the reported table with the critical values tabulated by Hasza and Fuller (1982) or with the critical values tabulated by Dickey, Hasza and Fuller (1984)?

Alternatively the null hypothesis is that first difference and seasonal difference (s=m) [Hasza and Fuller (1982)] is needed to make the series y stationary or the null hypothesis is that only seasonal difference (s=m) [Dickey, Hasza and Fuller (1984)] is needed to make the series y stationary?

Thanks in advance,

Andreas

[udo_sas · Accepted solution]

Hi Andreas -

You will find details about the augmented dickey fuller unit root test here: http://support.sas.com/documentation/cdl/en/etsug/63939/HTML/default/viewer.htm#etsug_macros_sect007....

Thanks,

Udo

[andreas_zaras]

Hello,

Thank you very much and sorry for the so late reply!

I thought i had replied earlier.

ANdreas