Hi there,

I have a panel dataset which contains four variables: house price I(0), housing stock I(1), wage I(0) and employment I(1). Also, two I(1) variables are cointegrated with trend.
I want to do a panel dynamic model, but now I have two I(0) and two I(1) variables. Can someone suggest which PVAR or PVECM is more suitable?
If it is PVECM, can I set I(1) variables as endogenous but I(0) variables as exogenous? But it doesn't seem convincing as price and stock are interdependent intuitively.
Please advise.

Thanks & Regards,
Michelle