I performed the Dickey Fuller test and I got the following results:

Test statistic:      0.340;
1% critical value:  -3.709;
5% critical value:  -2.983;
10% critical value: -2.623;
MacKinnon approximate p-value for Z(t) = 0.9791.

I know it is a one-tailed (left) test, but can anyone confirm me that it's okay to get a positive value of the test statistic? In all the examples concerning the DF test I saw I found only negative values of the test statistic...

  • $\begingroup$ What language/system/environment you were using? Could you copy the command you have run into the question? $\endgroup$
    – oszkar
    Jun 13, 2019 at 10:30
  • $\begingroup$ I am using Stata and the command I used is the following: dfuller price $\endgroup$
    – Serena
    Jun 13, 2019 at 10:34

1 Answer 1


Confirmed. You can get a positive df test statistic. It will occur when the estimated beta coefficient of your regression of y on y lagged, is greater than one. In this case the df_stat = (beta-1)/se(beta) > 0, since both the numerator and denominator are positive.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.