EViews.com

How can I obtain the residuals from the standard Johansen cointegration test?
I want to experiment with different lag lengths so that the residuals are normal and not serially correlated.

You could make a VAR with the same co-integrating relationship as that given by the co-integration test. Then from the VAR you can make the residuals.

You could make a VAR with the same co-integrating relationship as that given by the co-integration test. Then from the VAR you can make the residuals.

Great! I was just trying to do this. Just to be 100% sure: the residuals I get from the VAR are the same as the Johansen cointegration residuals?

And something last: To get the residuals from the Johansen cointegration test including lags 1 to 3 for example, I must estimate the VAR with
1 to 4 lags and get those?

Thank you very much!!!