Starting point using simult_ with order=3 and pruning

Stochastic simulations
1

I am running into the following problem.

  1. I use simult_ to simulate a model to the 3rd order for a large number of periods, with pruning, with all shocks set to 0, starting at the deterministic steady state (DSS).
  2. The simulation converges to a point where all variables remain constant. Let me call that point the stochastic steady state (SSS).
  3. I run simult_ again, without shocks, this time starting the simulation at the SSS (inputted as the first argument of simult_).

I would expect the simulation to remain at the SSS, but that is not the case. The end point does not match the initial point and there are some dynamics along the way. This does not seem correct – if the previous simulation (starting at the DSS) settled at SSS, then when I start at SSS, I would expect it to remain at SSS.

I believe this has to do with the pruning option, and in particular how yhat1,yhat2, yhat3 are set, but was wondering if anyone knows how to fix it.

Thank you!

2

This is tricky. See IRF: give the possibility to choose the starting point (at least at order 2 and above) (#116) · Issues · Dynare / dynare · GitLab

1 Like