With the original Chow test, you can only test two regressions vs a general one to see a structural change.

Is there a Chow test that can test structural change in 3,4...N regressions for structural changes versus a 'mother' regression at the same time?


Yes, there are extensions of the chow test for multiple breaks, if this is your question?

Bai and Perron investigated test, algorithms, and their distribution, look for example at:

If you are using R, the package strucchange has many nice features to estimate and test for multiple structural breaks

  • $\begingroup$ Yes, this is exactly my question. That R package is of no use for me as I want to check for structural differences by enacted policies in ARIMA models, whose class is not accepted by strucchange. I had to program the traditional Chow test myself, and took me quite some time, as I'm not a programmer by trade. $\endgroup$ – erasmortg Feb 19 '15 at 19:17
  • $\begingroup$ then you could as a gross estimate use strucchange to estimate the breaks on a similar AR() model, where I guess you will need to feed yourself the lags as regressors. Might give you already a rough idea of the possible number of breaks, their location, etc... $\endgroup$ – Matifou Feb 19 '15 at 19:24

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.