The RATS Software Forum

Sorry, the previous is incorrect, because in general covariances are not additive, variances are.

Therefore, for analytical DVECH I can only generate a 1-step similar to https://estima.com/webhelp/topics/mvgar ... edure.html, not multi-step.

That's not correct; the covariance matrix of the sum of a set of uncorrelated random vectors is the sum of the covariances.