enter image description here

The attached image is my output for augumented dickey fuller unit root test from stata, I am not sure my data is stationary. if it is please explain based on what rule. because some places i see that test statistic should be less than 5% while other places i see the test statistic should be less than 1.96(Siginificant critical value)


1 Answer 1


The null hypothesis of the (augmented) Dickey Fuller unit root test is that the series being tested has a unit root (that implies nonstationarity).

The basic logic tells you to reject the null hypothesis if the test statistic is sufficiently extreme. The question could be, does it matter whether the test statistic is extremely large and positive or extremely large and negative. Here you are lucky to see which way the extremality of interest goes: 10% critical value is -1.605, 5% critical value (more extreme than 10%) is lower (more negative) at -1.950 and 1% critical value (more extreme than 5%) is even lower at -2.639. So you see that the more negative the test statistic is, the more extreme. In other words, the extremality goes into the negative direction.

Now see what the actual test statistic is: 10.134. That is less extreme than the conventional 5% value (and even the 10% value) because it is less negative than -2.639 (and even -1.605). Therefore, you do not reject the null hypothesis of unit root at any of the conventional significance levels. You data seems to be nonstationary.

This is a generic way of using tests you do not understand well. Of course, it is always better to know more to understand the intuition behind the test, but that is not always the case.


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.