Non-stationarity and estimation

[Ausland2011]

Hi,

I am trying to estimate a regression where all variables are I(1) with the exception of one, which is I(2). In order to assure that all variables have the same order of integration, is the only solution transforming the variable that is I(2) in first differences in order to make it I(1)? Would it be ok to have all other variables in levels and one in first differences?

Thank you very much!