Hi James and Johannes,
Coincidentally I am after the same thing. Footnote 30 in Iacoviello and Neri(2010) explains how they construct their moments and a very similar question has been asked in this legacy post.
Essentially the first part can be replicated through moments_varendo, as done by James. Johannes, in the legacy post you point out that the second part is not really necessary due to linearity, hence no artificial time series is needed, just the theoretical moments. However, in terms of replicating the moments of the above paper, is it somehow possible to simulate data for the each posterior draw as described by Iacoviello and Neri(2010)?
Thanks.
Rob