Page 1 of 1

Retrieve Sij matrics from Johansen Estimation

Posted: Mon May 17, 2021 7:26 am
by jzpan

In testing and estimating a VECM, Johansen method uses the product moment matrices Sij to derive the eigenvalues and eigenvectors by solving the determinantal equation:

| lambda*S11 - S10*S00^-1 *S01 | = 0

I assume EViews procedure uses the same formula by taking advantage of the reduced rank regression.

My question is, is there a way in EViews to retrieve the Sij matrices after it has run the Johansen estimation of a VEC model?

The reason I need Sij is that I'm trying to derive the orthogonal complements of both the adjustment vector Alpha and the cointegration vector Beta (Alpha^perp and Beta^perp), which will be later used for the permanent and transitory decomposition of a multivariate time series system.

I'd appreciate any help or information anyone can provide.

Thank you,