Engle-Granger two step on stationary series

[chrisrik]

Hi guys!
I've got two variables, e which is the currency (price of one currency in another) and i which is the interest rate differential (local versus foreign).
These two series are I(1) but so is the residuals, thus I can't just go ahead using error correction models.

But; if I use the change from periode t to t+1 in e (delta e) and delta i, both these series are I(0) and so is the residuals; is it okay to just go ahead with the Engle-Granger's two-step procedure?

Am I correct to say that there is some kind of debate around doing it this way?

Thanks for any help

[random_access]

if u deal with the data in the level then they r usually integrated of order 1 i.e.I(1) and usually the residual will be integrated of degree zero i.e. I (0)......i cannot understand how all r integrated of degree one.
if u deal with the change then u have differenced the data so it is stationry in the level mostly.