# Negative initial values

**URL:** <https://forum.dynare.org/t/negative-initial-values/13749>\
**Category:** Stochastic simulations\
**Created:** [4 May 2019 11:27 UTC](https://forum.dynare.org/t/negative-initial-values/13749 "2019-05-04T11:27:55Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![mikegouv](https://forum.dynare.org/user_avatar/forum.dynare.org/mikegouv/32/9467_2.png) [@mikegouv](https://forum.dynare.org/u/mikegouv)\
**Post date:** [4 May 2019 11:27 UTC](https://forum.dynare.org/t/negative-initial-values/13749/1 "2019-05-04T11:27:55Z")

</div>

Hi,

I have a question regarding negative initial values and how to interpet them. For example, in Gerali et al. (2010) model found in MMB base, a lot of initial values used to solve the model are negative, like consumption, housing, interest rate etc. The model is also expressed in logs - exp().

So does this initial values and their sign actually mean something?

Kind regards

---

<div class="post-metadata">

**Author:** ![stepan-a](https://forum.dynare.org/user_avatar/forum.dynare.org/stepan-a/32/6652_2.png) [@stepan-a](https://forum.dynare.org/u/stepan-a)\
**Post date:** [4 May 2019 18:45 UTC](https://forum.dynare.org/t/negative-initial-values/13749/2 "2019-05-04T18:45:33Z")

</div>

Hi,

If the model is expressed in logs (i.e. each variable X is replaced by e^x), then the steady state is for the variables in logs (x, not X). So it is possible to have a negative value in the steady state if the level of the original variable (X) is smaller than one, and it makes sense to use a negative initial guess for these variables.

Best,  
Stéphane.

---

<div class="post-metadata">

**Author:** ![mikegouv](https://forum.dynare.org/user_avatar/forum.dynare.org/mikegouv/32/9467_2.png) [@mikegouv](https://forum.dynare.org/u/mikegouv)\
**Post date:** [15 May 2019 07:38 UTC](https://forum.dynare.org/t/negative-initial-values/13749/3 "2019-05-15T07:38:08Z")

</div>

@stepan-a thank you very much for your help