Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 2 down vote favorite

I am testing for cointegration all the pairs from a set of 100 stocks. I run an Ordinary Least Square Regression on each pair and then I test for the existence of unit roots in the residuals. I am using the ADF test with no trend and no intercept. I am surprised that I found a big amount of pairs which the test on the residual rejected the null hypothesis of having a unit root, with p-values smaller than 0.01. Can I use the Dick Fuller Statistic to rank those pairs? All the series have the same length.

Thank you!

share|improve this question
It depends on what you are after. ADF test tells you how strongly you can reject the null hypothesis of a unit root, so if you want to rank the pairs by how strongly you reject the null, then sure. – Akavall Jul 1 '12 at 12:40
I would think that as long as you validated the Gaussian Assumptions underlying the t test you should be good to go. I am currently expanding the ECM solution to deal with Gaussian Assumptions/Violations using AUTOBOX a time series / econometric package I helped develop. – IrishStat Jul 1 '12 at 15:27
@Akavall, I want to rank them by how strong they are cointegrated. As the pairs are very different, have different statistical caracteristics, i was not sure if a DF t-statistic for one pair could be comparable with a t-statistic for another pair. – Thiago Steiner Alfeu Jul 1 '12 at 19:19
@IrishStat, thank you. I agree with you that validating the Gaussian Assumptions is a very important matter in dealing with the t statistic. How should I validate those assumptions? – Thiago Steiner Alfeu Jul 1 '12 at 19:29
@Thiago Please see my answer to stats.stackexchange.com/questions/31275/… – IrishStat Jul 1 '12 at 19:41
show 2 more comments

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.