Loss function and OSR command

Optimal Policy
8

Hi all. I am conducting a policy frontier as guided here: Policy frontier. Running into this thread,
I now have several questions (and confusions).

  1. I wonder what is the exact definition of ‘inefficient steady state’ under this thread? NK models are inefficient by construction, does this mean that a policy frontier analysis using OSR is invalid?

  2. In which condition(s) one might run into ‘inefficient steady state’?

  3. When should one concerns about ‘inefficient steady state’ and avoid using OSR command to derive the policy frontier.

Very much looking forward to be enlightened!