Page 1 of 1

Optim Weights in Optimal Policy Rules

PostPosted: Thu Sep 17, 2015 5:06 pm
by Nicolas Pinell
Hi I want to know how can I optimize the weight of lambda because I use:
Code: Select all
optim_weights;
y 1;
p lambda1;
end;

but I have to put the value of lambda1 above and the results changes everytime I change the value of lambda1.
Is there a way to make the program find the value of lambda1 that minimizes the loss function??
Thanks

Re: Optim Weights in Optimal Policy Rules

PostPosted: Thu Sep 17, 2015 7:19 pm
by jpfeifer
Your question makes no sense. You are telling Dynare to find the minimum of
y^2+lambda*p^2
Clearly, as you are considering the sum of two positive things, the minimum is for lambda equal to 0. Think about it this way: lambda measures the strength of the tradeoff between fluctuations in y and p. Of course the policy maker would be happy if there would be no tradeoff (p=0), but it makes no sense to search for the minimizing value.

Re: Optim Weights in Optimal Policy Rules

PostPosted: Thu Sep 17, 2015 7:25 pm
by Nicolas Pinell
Yes I understand your point. But what value of lambda1 do I have to use?

Re: Optim Weights in Optimal Policy Rules

PostPosted: Thu Sep 17, 2015 7:41 pm
by jpfeifer
There are no guidelines. Either you pick some loss function in an adhoc-way or you try to derive the parameter from microfoundations as e.g. in Gali's (2008) textbook.