Cross correlation - first difference

[j52]

Hello,

is there a way to efficiently test the cross correlation of d(y) and d(x) series without having first to create the differences using series dy=d(y)? Or is it even possible to program an automatic selection using the best lag structur according to aic or bic?

[EViews Gareth]

Make a group containing d(y) and d(x), rather than containing y and x.

[j52]

yes, but do I first have to define the new series dy and dx or can I open the group with y and x and somehow adjust them to d(y) and d(x)?
Rigth now, I have the series y and x. Define them as series dy=d(y) and series dx=d(x). Then open dy and dx as a group to run the cross correlation. Is there a way to do it faster?

[EViews Gareth]

Yes, make a group containing just d(y) and d(x).

Code: Select all

group mygroup d(y) d(x)

[j52]

Well, I guess there is then no way around creating the dy series. But thanks anyway.

[EViews Gareth]

Um, I just showed you how to do it without creating the series dy and dx.

[j52]

:) okay, my mistake. But Iguess the effort is the same if I create the groups or the dx series.