8.3 Vector Moving-Average Models
A vector moving-average model of order q, or VMA(q), is in the form
where 
is a k-dimensional vector, 
are k × k matrices, and 
is the MA matrix polynomial in the back-shift operator B. Similar to the univariate case, VMA(q) processes are weakly stationary provided that the covariance matrix 
of 
exists. Taking expectation of Eq. (8.23), we obtain that 
. Thus, the constant vector 
is the mean vector of 
for a VMA model.
Let 
be the mean-corrected VAR(q) process. ...
Get Analysis of Financial Time Series, Third Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
