# Initial value - Conditional forecasting

**URL:** https://forum.dynare.org/t/initial-value-conditional-forecasting/15738
**Category:** ML/Bayesian estimation
**Created:** [4 May 2020 10:13 UTC](https://forum.dynare.org/t/initial-value-conditional-forecasting/15738 "2020-05-04T10:13:19Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![guillaume.s](https://forum.dynare.org/letter_avatar_proxy/v4/letter/g/439d5e/32.png) [@guillaume.s](https://forum.dynare.org/u/guillaume.s)
#### Post date: [4 May 2020 10:13 UTC](https://forum.dynare.org/t/initial-value-conditional-forecasting/15738/1 "2020-05-04T10:13:19Z")

</div>

Dear Professors,

I would like to perform a conditional forecast but with the last historical value of the data as the initial point of the forecast instead of the steady state value as it does by default. I know that the unconditional forecasting (for example with the estimation function) does it but I am wondering if there is a way to obtain the same initial value for the conditional forecast.

Thanks in advance,  
Guillaume

---

<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: [4 May 2020 11:39 UTC](https://forum.dynare.org/t/initial-value-conditional-forecasting/15738/2 "2020-05-04T11:39:32Z")

</div>

If your `parameter_set` is different from `calibration`, Dynare should run the Kalman smoother on your data to extract the state values at the end of the sample for forecasting.