# Comparing the Welfare under Ramsey Policy and OSR

**URL:** <https://forum.dynare.org/t/comparing-the-welfare-under-ramsey-policy-and-osr/15675>\
**Category:** Optimal Policy\
**Created:** [24 April 2020 20:00 UTC](https://forum.dynare.org/t/comparing-the-welfare-under-ramsey-policy-and-osr/15675 "2020-04-24T20:00:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Chashika](https://forum.dynare.org/letter_avatar_proxy/v4/letter/c/49beb7/32.png) [@Chashika](https://forum.dynare.org/u/Chashika)\
**Post date:** [24 April 2020 20:00 UTC](https://forum.dynare.org/t/comparing-the-welfare-under-ramsey-policy-and-osr/15675/1 "2020-04-24T20:00:40Z")

</div>

Dear Prof. Pfeifer,

I have been trying all the resources here to find a solution to my problem but I could not.

I developed a DSGE model similar to [Airaudo and Olivero (2019)](https://onlinelibrary-wiley-com.manchester.idm.oclc.org/doi/full/10.1111/jmcb.12598). They use Ramsey Policy. I developped the model and code.

Now I want to compare the Ramsey Policy (RP) vs Simple Optimal Rule (OSR). i.e. instead of adopting a more complex RP I want to know if we can find the optimal parameter values for the Taylor rule that might deliver a closer welfare gains to RP.

I use the `ramsey_policy` command to run the RP and `osr` command to run the optimal simple rule. The entire model is same in both cases and the `initial values` in RP has been adjusted as you have advised (i.e. conditional on R\_ss).

Following (Airaudo and Olivero, 2019), the central bank’s objective is obtained by taking a second-order approximation to the representative agent’s welfare. Which will give me the following Loss function;

`Loss = (1/2)*(coeff_pi * (pi^2) + coeff_y_gap * (Y_g^2))`

I specify this loss function under `planner_objective` in the `ramsey_policy` command and I use `coeff_pi` and `coeff_y_gap` as `optim_weights` in `osr` command. If my understanding is correct this implies: `loss = coeff_pi*var(pi) + coeff_y_gap*var(Y_g)`.

Next, I want to compare the welfare under these TWO policies. i.e. Adopting a Ramsey Policy vs OSR to see which policy I should recommend. In doing that I want to make a welfare comparison.

From this post: [Dynare forums • View topic - OSR vs. Ramsey problem](https://archives.dynare.org/phpBB3/viewtopic.php%3Fp=25515.html)  
I understand that I cannot simply use the value of the `objective function` in `ramsey_policy` and `osr` to compare these two policies.

But [Schmitt-Grohé and Uribe (2007)](https://www.sciencedirect.com/science/article/abs/pii/S030439320600167X) Table 2, they do a Welfare comparison between Ramsey Policy and OSR. Please kindly advice how can I do a similar comparison using `ramsey_policy` and 'osr` commands in Dynare. Unfortunately I cannot follow the method they have done, because they do not use Dynare as you may know.

**Question: 1**  
**I am really grateful to you if you could kindly advice me how I should carryout the welfare comparison between these two policies (welfare under RP vs OSR).**

* * *

**Question: 2**  
Should I use,  
`Loss = (1/2)*(coeff_pi * (pi^2) + coeff_y_gap * (Y_g^2))`  
function which I obtained by taking a second-order approximation to the representative agent’s welfare or the HH’s Utility function as the `planner_objective` in the `ramsey_policy`?

I am extremely sorry for asking such a long questions, I was trying to figure it out myself but I am completely lost.

Thank you very much for your time professor…

Kind regards  
Chashika