As written in the title, I am looking for the stability condition of a vector error correction model (VECM).

I have found this phrase:

the companion matrix of a VECM with 𝐾 endogenous variables and 𝑟 cointegrating equations has 𝐾 − 𝑟 unit eigenvalues. If the process is stable, the moduli of the remaining 𝑟 eigenvalues are strictly less than one.

Is this the correct condition? If so, could anyone tell me how we define "the companion matrix of a VECM"? Is it the coefficient matrix of X_(t-1) ?

In the end, how could I test this condition in R? Is there a package to do this?


migrated from stackoverflow.com Aug 30 '16 at 1:47

This question came from our site for professional and enthusiast programmers.


I maybe have an answer which I'm not sure about : I think that I should transfer my VECM to a VAR form, and then I could analyse the Eigenvalues of the compagnon matrix of VAR. There should be r unit roots where r is the number of cointegration relationships, and the module of the rest Eigenvalue should be strictly inferior to 1.

The definition of a compagnon matrix could be found at page 7 and 8 of this link : http://economia.unipv.it/pagp/pagine_personali/erossi/rossi_VAR_PhD.pdf

Please confirm if my idea is correct. Also, if my question is not clear enough, please tell me.


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.