I am analysing a time series and all the variables are stationary using augmented Dickey–Fuller test (ADF). Does it make sense to test for co-integration in stationary series?
$\begingroup$
$\endgroup$
Add a comment
|
$\begingroup$
$\endgroup$
No, cointegration occurs when two series have the same stochastic trend. In order to have a stochastic trend, they should be nonstationary. Stationary variables do not have a trend. You can perform standard inference.
