Hello,

We have difficulty establishing cointegration using the Johansen methodology for a set of macroeconomic variables of a developed country. Similar difficulty appears with the estimation of the ARDL model despite the fact that we have the option to choose variables from a larger set ,that meet the required dynamic stability properties. The left-hand variable is real GDP or its log.

Will it make sense to employ nominal variables?

The next step is to consider nonlinear models. Can you help.

Thank you.

George K Zestos