Page 1 of 1

Cointegration: I(0) variable included in ce

Posted: Mon May 18, 2009 12:30 pm
by hum2004
Dear all,

i do have the following question and woud apprecate your kind help very much:

If one has a cointegration relation - say of 3 variables - and applies johansen`s procedure:

-What sense does it make to include a I(0) variable in the conintegration relation (ce)?
-what does change if one included this variable in the ce? (do test statistics change, do critical values (lamda trace and/or lamda max) change, does the interpreation change? do diagnostic test statistics have to be adjusted in any way?

Thank you very much for your response.

hum2004