estimating time series model with I(0),I(1),I(2) processes

[bouki_81]

Hi....
I want to estimate a model for inflation, but I get the dependent variable has I(2) process and the other variables are mixed between I(0) and I(1) process, and I've been told that neither the johansen coitegration test nor the ARDL model are valid in this case, and I know to run johansen cointegration test the variables should be integrated at the same order, and for the ARDL model the dependent variable should be integrated at order one, so is there an other approach to deal with this case?

I would appreciate any advices.
thanks in advance

[solarin]

Try stationarity tests with structural breaks such as Zivot and Andrews (1992) Lee and Strazicich (2003, 2004). These may turn the series to at most I(1). I think you can use ARDL, if the series are I(0) and I(1).