generalized vs. "orthogonalized" impulse responses

[Dthurley]

I have two questions :
(1) How do you decide in whether to use generalized or orthogonalized impulse responses ...based on paper by Pesaran & Shin (1998) ? I have seen work here people have tested first for "non-diagonal error variance matrix" but I don't see any tests like that in EVIEWS
(2) If one decides to use the generalized impulse response, then, should variance decomposition be also based on the same assumption that the shocks are comtemporaneouly corrlated? If that's the case, where in EVIEWS do we have that option?

[ppp007]

up