Page 1 of 1

Datatype for Cointegration

Posted: Mon May 23, 2016 11:59 am
by ballpen
Hey Everyone :(

I'm currently working on a project and need some help for my cointegration test.
I want to use the Engel-Granger Cointegration Test.
I have Level Variables (GDP,X,M,C,I,G, all in local currency), the real interest rate, and the real exchange rate to the USDollar.
I Know that the variables for testing for cointegration have to be level variables and they all have to be non-stationary.

i used the Unit Root Test with augmented dicky fuller to test for stationarity, the real exchange rate and the real interest rate are stationary.
My question is, do i have to use the differences of all Variables now, meaning i have do generate the differences of all 8 Variables?
Do i also have to use des LOG of the six level Variables?
I can't find a satisfying answer in the literature. There are only examples with all variables being stationary or non-stationary.
Do i also have to include a trend for some Variables (e.g. GDP)? I know that there is a button for trend when testing for a unit root, but I don't know when to use it.

Thank you very much, your help is really appreciated!