on dummy variable stationarity

[deXter]

Hello guys,

can't seem to find an answer anywhere. I have an LS with an explanatory variables that are all dummy. Now i want to run causality testing.
Should i test my explanatory (dummy) variables for stationarity/unit root? Is it even possible?

Have found on EconometricsBeat this: "you don't need to test dummy variables for stationarity - all dummy variables are stationary by construction".
It would be very helpful if someone could link me to reliable source for this kind of statement (can't seem to find it in my econometrics books too).

Thanks in advance.

[Atalay]

Hello Dexter,

I read your message and I also suffered from the same question that you asked. Could you find any reliable source from econometric literature that tells do we need to check unit root for dummy variables? or it is assumed that all dummies are stationary by its nature?

Take care.

[jmagomez]

Dear Dexter,

I really don't know any reference (book, paper...) that says to test dummy variables for unit root. But I agree with Professor Giles when he says about dummies are stationary by nature.
But you can try an ADF test just to solve your doubts.

Regards.