Dear all,
I would like to know if there is a way to use the conditional_forecast command without having to re-estimate the model, by simply loading the existing results.
I am using Octave and already have my estimation results saved. It would be helpful to modify the paths of the endogenous variables for sensitivity tests of my projections. However, since the model takes a considerable amount of time to estimate, the process becomes quite slow for each new endogenous path test.
Thank you,
Heitor
Usually, that should be rather straightforward to do. What do you want to load from estimation? Just the parameters? Or also the smoothed states at the end of the sample?
Hi Professor,
It would be helpful to load the smoothed states. I can load the estimation results into Octave without any issues. What I haven’t been able to do, though, is run the conditional forecast without having to re-estimate the model (even though the results are already uploaded).
Thanks!
The easiest way would be to run the calib_smoother to get the state estimates without reestimating.