Dear community,

When I run osr under oder 2 and 1 I get the same optimal parameters’ result. I wonder if order 2 is truly working and if yes can I do welfare anlysis with ?

Thank you in advance

Dear community,

When I run osr under oder 2 and 1 I get the same optimal parameters’ result. I wonder if order 2 is truly working and if yes can I do welfare anlysis with ?

Thank you in advance

As documented in the manual, `osr`

only works with `order=1`

. Regarding welfare: if your welfare objective falls into linear-quadratic class, then the answer is yes. Otherwise, the answer is no (e.g. with distorted steady states)

Thank you Professor for your reply.

I’m not able to answer that question. With price stickiness à la calvo in both domestic and imported goods firms and no real wage rigidities, I guess the answer is no?

I would like to know please if it is possible to use the result of osr order 1 and make stoch-sim order 2 to assess welfare loss ?

Thank you

Can I use this code to maximize my welfare as objective function ?

https://forum.dynare.org/t/optimal-policy-parameters-in-a-non-linear-model/11772/2?u=i.lagrine

Yes, that code can be used.

Thank you very much Professor. Just one more question, how can I assess whether my welfare objective falls into linear quadratic class ? I’m reading Benigno papers

The question is whether the steady state is undistorted. That is often hard to judge. For that reason, I would always go for a full second-order approximation.

Ok thank you for your priceless help!

Professor Pfeifer,

I remark that the optimization routine used in the code above meant to minimize the objective function i.e. welfare, while I aim to maximize the welfare. I tried to replace welfare by -welfare but the results are not satisfing. What do you think ?

No, you can see in

```
outvalue=-oo_.mean(strmatch('omega_e',var_list_,'exact')); %extract Welfare
```

that **minus** the welfare is minimized, i.e. welfare is maximized.

I didn’t pay attention to the minus sign! I am embarrassed, thank you very much