Hi friends, I would really appreciate if somebody helps me give better understanding on subject matter. Suppose I have a set of Integrated variables (say max order is 1). I understand there may be some cases when there could be some linear combinations which are stationary and this phenomena is called as co-integration. Now suppose I have a vector of Random walks (which is also integrated of order 1 each), where error terms are not necessarily independent but of course uncorrelated over time i.e. error terms are WN. Here my question is still can we find some linear combinations (like co-integrated variables) which are stationary? And if not, why this is not. Please forgive me if my question seems very naive. However although intuitively I feel that no such linear relationships exist for RWs, however I find hard to explain to 2nd person. Any help please? Thanks [[alternative HTML version deleted]]