Perfect foresight with imposed values for an "endogenous" variable

Hi all,

I want to estimate what is the implied interest path in Smets and Wouters (2003) model given an exogenous path for inflation and some initial conditions.

For this purpose, I am doing a perfect foresight simulation considering inflation as exogenous (i.e. within varexo) and abstracting from new realizations of every shock except the monetary policy shock, which now I consider as endogenous to keep the same number of equations and unknowns. Then, in the shock block, I impose the path of inflation I want.

The code works but the outcome is very volatile and it is extremely different from what someone might expect, especially the nominal interest rate since it is impossible to have a smooth path of inflation considering the sharp movements in R. Any idea of why this might be happening?

Attached are the codes I am using.
SW03_perf_foresight.mod (2.6 KB)


Why do you think this is suspicious? You are providing an arbitrary path for inflation in your model that is known to agents. Now they need to make the best out of that path. Apparently that path creates a spiky real interest rate.