Engle Granger Vs adf test on residuals

[Elghark]

Hi everyone, this is my very first post, so in case of any mistake i apologize in advance.
This is my problem: i was making some analysis on a currency pairs in order to develop a pairs trading strategy. I was pretty sure about the fact that the following two procedures were the same:
1)make a two step engle granger test by click ing the eviews test
2) i personally made the ols equation of the two currencies and once saved the residuals, I made the adf test on these.
However my results lead to two opposite directions: the first procedure shows no cointegration while the second show stationarity of residuals. Doesn't it mean they are cointegrated?

[dakila]

If you want to use procedure 2 you should use EG critical values. Because of generated residuals, you should different c.v table from ADF test c.v table.

[Elghark]

Thanks for your reply. To be honest I was suspicious about the very same reason. Now you confirmed it.
I downloaded both ADF and Engle Granger Tables for critical values and in fact I now see significant differencies.
Maybe I was just disappointed for tests not confirming my hopes of cointegration between EurJpY vs ChfJpY......just have to look for another good idea

Thanks again