I started off using Granger causality test but since the two time series are not stationary and are also co-integrated. Is there another causality test that I can perform which is similar to Granger causality but works on non-stationary time series as well

  • $\begingroup$ Do you have "Causal" causality or Granger causality in mind? The difference is explained here. If it's the first, then probably not, unless the circumstances are fairly unusual. $\endgroup$
    – dimitriy
    Mar 26 at 4:38
  • $\begingroup$ I'm looking for a way to show that X causes Y (does not have to be granger causality). Or at least some way that gets me close to saying that. $\endgroup$
    – hussamh10
    Mar 26 at 4:43
  • $\begingroup$ I'm not very knowledgeable on time series, but since granger causality seems to only work on stationary time series and more often than not practical time series are non-stationary doesn't it make granger causility applicable on a niche or is it almost always required to convert stationary to non-stationary time series? $\endgroup$
    – hussamh10
    Mar 26 at 4:51
  • $\begingroup$ Unfortunately, Granger causality has very little to do with causality. It's a misnaming from a time when we were less careful about such things and has led to much confusion. It does have its uses in forecasting. However, without additional context that provides some restrictions, there is no rigorous procedure that recovers causal relationships from data, stationary or non-stationary. $\endgroup$
    – dimitriy
    Mar 26 at 4:58
  • $\begingroup$ Ok that does clear things up a bit. Thanks a lot for the help! $\endgroup$
    – hussamh10
    Mar 26 at 4:59

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.