Goal-oriented Agent, Environment, Policy delay effect model

How can we better think about the advantages and disadvantages of quick/slow policy response?

Given ...

- the policy measurement may be better at a big picture level

- the policy response is possibly larger

- but policy interventions are likely slower and potentially less accurate than the agent's response