Forecasts using Ramsey optimal policy

Optimal Policy
1

Hello, I would like to make a forecast from a given state using Ramsey Optimal Policy. Is this possible?
I have tried to find this in the documentation, but I could not.

I also thought about stacking impulse responses to form a forecast, but in the context of optimal policy, I think they are not linearly stackable if I use a quadratic loss function.

2

You should be able to call the

command after the ramsey_policy
command (alternatively after a combination of ramsey_model and stoch_simul)

3

Thank you

4

Hi Professor,

I have a related question. Is it possible to combine calib_smoother and smoother2histval with optimal policy commands to do forecasts starting from the last observation? Also, is it possible to do a conditional forecast under an optimal policy setup?

In the attached template, I tried using the conditional_forecast command after solving for the Ramsey and discretion optimal policy. The code runs, but I’m not getting the expected results. The file works when I solve for osr though.

Thank you vm for the help.

test.mod (2.2 KB)

5

Usually, that should work. The issue in your case is that under optimal policy, y is insulated from y_ shocks:

POLICY AND TRANSITION FUNCTIONS
                                   y       inflation               r
y(-1)                              0               0        7.333333
inflation(-1)              -0.238274        0.569909        1.583964
MULT_1(-1)                  0.261812        0.062962       -4.580661
MULT_2(-1)                  0.055655        0.026219       -0.633517
y_                                 0               0       16.666667
inf_                       -0.496404        1.187311        3.299924

Thus, the desired conditional forecasts do not make sense.

1 Like
6

I see, seems like I’m getting a divine coincidence under optimal policy in this toy model.
I’ll break it to make things more interesting.
Thank you vm, professor.