1
$\begingroup$

I am currently studying whether stock markets in the GCC region are cointegrated. All series are I(1); however, the Johansen test provides no robust evidence of cointegration. I then used the Bai Perron test for multiple breaks and found that there is about one break per year. Reading similar literature there are typically two ways of proceeding. First, to use the Gregory Hansen test if there is one break or use the Johansen procedure on sub-samples. However, as my sample is from 2012-2016, I feel as though multiple breaks and sub-samples would lead to unreliable inference. Please advise me on how I could proceed with this analysis. I only have access to EViews.

Update: I found the solution: Incase anyone may be in the same situation in the future, I used the Kejriwal and Perron test which isn't available without having to use coding and programming. Also, re-examine your data as there were a few issues there regarding market timing.

$\endgroup$
  • $\begingroup$ I did not receive the notification of your comment. Next time try writing "@RichardHardy" in front. Sorry, but I am pretty busy these days. I would have helped already if I had been able to. Your situation seems tricky. Could you plot your data and include some test results? Try to be as clear as possible. I probably will not give you an answer, but having some extra information would help other users. Also, is your goal to answer the question whether the series are cointegrated? Or is it just an intermediate step toward some other goal? $\endgroup$ – Richard Hardy Aug 9 '16 at 20:11
  • $\begingroup$ Incase anyone may be in the same situation in the future, I used the Kejriwal and Perron test which isn't available without having to use coding and programming. Also, re-examine your data as there were a few issues there regarding market timing. $\endgroup$ – Vladmir Putin Aug 10 '16 at 19:00
  • $\begingroup$ An update and a cleanup of your post would still help. $\endgroup$ – Richard Hardy Aug 10 '16 at 19:28
  • $\begingroup$ I did provide an update.. Sorry I am not sure how exactly you mean to clean up as I am still relatively new here. You can delete the post if you want as i've found the solution as I listed above. $\endgroup$ – Vladmir Putin Aug 10 '16 at 19:30
  • $\begingroup$ Now that you edited the post it's better. What you could do more is address the remaining points in my first comment, if you have time. If not, let it be. By cleanup I meant trying to write things explicitly so that it would be easy to follow for anyone. That requires some effort but could gather upvotes and serve the user community better. $\endgroup$ – Richard Hardy Aug 10 '16 at 19:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.