Need more robust steady state algorithm for reforms in SS.py

[rickecon]

I ran into a strange case today in the OG-ZAF repository (see this comment in PR thread for details) in which the baseline steady-state solved but the reform steady-state did not. This is a very rare occurrence. Usually, the baseline steady-state values make good initial guesses for the outer loop of the reform steady state. However, in this one case, it didn't work.

Currently, the robust steady-state solution method with the list of initial guess perturbations is only for the baseline case. In the reform case, every option in SS.py should also follow the robust SS computation algorithm we use in the baseline. This is straightforward to implement, and I think it will get rid of the current method when use_new_guesses=True that just reverts the initial guess to the default for the baseline.

cc: @jdebacker

[rickecon]

@jdebacker correctly pointed out in this comment in the OG-ZAF repository that the enhancement suggested in this issue is not necessary. Closing this issue.