Missing data and BVAR model ''a la Sims''

Dear all,
I want to ask how to estimate the density of the BVAR model in Dynare when we have missing data (I use Mixed Frequency). I would like to use this to compare with the DSGE model. I know that Dynare can not calculate the density of the BVAR model if there is missing data.

  1. Is there another solution?
  2. Can we use smoothed variables as observation data? For example, if the observation equation is as follows:
    y_obs = y + y (-1) + y (-2) + y (-3)
    Can we use y (smoothed) as observation data?
    Thanks in advance.

There are mixed frequency VARs:
Eraker, Bjørn, Ching Wai (Jeremy) Chiu, Andrew T. Foerster, Tae Bong Kim, and Hernán D. Seoane (2015). “Bayesian Mixed Frequency VARs”. Journal of Financial Econometrics 13 (3), 698–721.
We use that approach in Born/Pfeifer (2017) “Uncertainty-driven business cycles: assessing the markup channel”.

Thanks Johannes for your reply.
Can we apply it in Dynare? If yes, how? If not, should we use a particular code or program?
I have seen your paper and paper referred to,
I think it needs to build a code. I found a code for the mixed frequency VAR for estimating Ghysels (2012) ([http://www2.kobe-u.ac.jp/~motegi/MFVAR_toolbox.zip]) but not BVAR.
If you can provide me the code I would be grateful.
Thanks again

I have to clean up the code a bit and then should be able to provide you with it. It’s Matlab-based.

OK, Thank you very much Dear Johannes.

Sorry for any inconvenience sir @jpfeifer
I just want to remind you of the code because I need it very soon.
Thank you again.

Please send me an email


Dear Johannes,
I met the same problems as Salo. I was really appreciate if you could share the codes to me.
This is my email: chteng23@163.com

I sent the code to you.

Dear Professor Pfeifer
I’d appreciate if you send me the code:


Thanks in advance

The paper has been published: Uncertainty‐driven business cycles: Assessing the markup channel | Born | Quantitative Economics
You can find the replication files there.

Dear Professor Pfeifer

Thanks very much.