# Johansen cointegration test with stationary variables

Journals and books about the Johansen cointegration test suggest that only I(1) variables should be put into a cointegration test. However many journals modelling economic growth and inflation use two I(0) variables in the cointegration test. Is this an acceptable procedure?

What is then the interpretation of the residual if two I(0) variables are used in the cointegrating regression?

-

Could you please suggest us a link to the paper? In case of I(0), there is no rationale for using co-integration.