# Var-dsge

**URL:** <https://forum.dynare.org/t/var-dsge/5450>\
**Category:** DSGE-VAR\
**Created:** [17 April 2016 08:49 UTC](https://forum.dynare.org/t/var-dsge/5450 "2016-04-17T08:49:56Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![fsacault](https://forum.dynare.org/letter_avatar_proxy/v4/letter/f/65b543/32.png) [@fsacault](https://forum.dynare.org/u/fsacault)\
**Post date:** [17 April 2016 08:49 UTC](https://forum.dynare.org/t/var-dsge/5450/1 "2016-04-17T08:49:56Z")

</div>

Hi, I am a Phd student from French Polynesia University,  
I read somewhere (in the UserGuide ?) that Dynare can solve Var-DSGE,  
I am very interested by this method, in particular I read the paper from Marco Del Negro nd Frank Schorfheide, _Priors from General Equilibrium Models fro Vars_  
Where can I find an example of Var-Dsge solved by Dynare ?  
Thanks

---

<div class="post-metadata">

**Author:** ![jpfeifer](https://forum.dynare.org/user_avatar/forum.dynare.org/jpfeifer/32/5044_2.png) [@jpfeifer](https://forum.dynare.org/u/jpfeifer)\
**Post date:** [17 April 2016 16:37 UTC](https://forum.dynare.org/t/var-dsge/5450/2 "2016-04-17T16:37:34Z")

</div>

See [github.com/DynareTeam/dynare/tree/master/tests/dsge-var](https://github.com/DynareTeam/dynare/tree/master/tests/dsge-var)

---

<div class="post-metadata">

**Author:** ![fsacault](https://forum.dynare.org/letter_avatar_proxy/v4/letter/f/65b543/32.png) [@fsacault](https://forum.dynare.org/u/fsacault)\
**Post date:** [18 April 2016 08:41 UTC](https://forum.dynare.org/t/var-dsge/5450/3 "2016-04-18T08:41:40Z")

</div>

Thanks Johannes Pfeifer

---

<div class="post-metadata">

**Author:** ![fsacault](https://forum.dynare.org/letter_avatar_proxy/v4/letter/f/65b543/32.png) [@fsacault](https://forum.dynare.org/u/fsacault)\
**Post date:** [21 April 2016 18:49 UTC](https://forum.dynare.org/t/var-dsge/5450/4 "2016-04-21T18:49:32Z")

</div>

I understand that var-dsge is:

- a way for taking into account of all your endogenous variables for which data series are available (meanwhile on DSGE alone, the number of varobs may be one or two variables and anyhow limited by the number of shocks) : it seems to me you lose the information brought by the data not included in varobs; for estimating a posteriori the parameters;
- consequently as said by Shorfeide, var-dsge results are more adequate to data than dsge alone, however what may be very puzzling in var-dsge method is when you merge simulated data from theoretical data to observed data for estimating the a posteriori parameters, don’t you distort the observed reality ?

thanks to enlighten me on this  
Kind regards,  
Franky

---

<div class="post-metadata">

**Author:** ![jpfeifer](https://forum.dynare.org/user_avatar/forum.dynare.org/jpfeifer/32/5044_2.png) [@jpfeifer](https://forum.dynare.org/u/jpfeifer)\
**Post date:** [22 April 2016 09:35 UTC](https://forum.dynare.org/t/var-dsge/5450/5 "2016-04-22T09:35:20Z")

</div>

Please carefully read the Del Negro/Schorfheide papers on DSGE-VARs. The crucial element is that DSGE-VARs allow relaxing the rigid cross-equation restrictions of DSGE-models while still using informative priors in a VAR. This is very powerful for forecasting. The first point you mention is not true. You always need at least as many shocks as observables to avoid stochastic singularity.

---

<div class="post-metadata">

**Author:** ![fsacault](https://forum.dynare.org/letter_avatar_proxy/v4/letter/f/65b543/32.png) [@fsacault](https://forum.dynare.org/u/fsacault)\
**Post date:** [23 April 2016 11:13 UTC](https://forum.dynare.org/t/var-dsge/5450/6 "2016-04-23T11:13:17Z")

</div>

thanks
