Exogenous variable in VECM

[basher]

Hi,

I am curious to know how Eviews incorporates the exogenous regressors in VECM estimation since such an introduction affects the distribution of the rank test statistics unless they are I(0) and introduced in a specific fashion. See, for example, ANDERS RAHBEK and ROCCO MOSCONI, Cointegration rank inference with stationary regressors in VAR models, Econometrics Journal (1999), volume 2, pp. 76-91. http://old.dig.polimi.it/ita/comunita/ch/45/ej99.pdf

Thank you.

Basher

[EViews Gareth]

We compute the statistics in the usual way in the co-integration test, but put up a warning saying that they may not be valid when exogenous variables are present.

[basher]

So it seems it is not possible to compute Johansen cointegration test with a I(0) exogenous regressor in Eviews? Thank you.