Issues with an AR(1) process interacting with IRFs

[treyaro]

Hey guys

I am working on introducing a macroprudential rule into a really basic economy with housing.

We are trying to introduce a noise disturbance 'u' into our rule that simulates the effect of imperfect information on macroprudential policy setting.

For some reason, our value for noise ('u') which is defined as an AR(1) process as found in the mod file, has no effect on the IRFs when introduced into our policy rule.

I suppose we have two questions - 1) Are we defining the process correctly? 2) Are we implementing it correctly into the model?

I would appreciate any help. I have attached our .mod file.

Many thanks.

[jpfeifer]

You did not set any shock variance for

Code: Select all

eps_u

. You need to include it in

Code: Select all

shocks;
var eps_e; stderr 100*STDERR_AE;
end;


Otherwise, the variance is assumed to be 0.