K-order approximation: Dynare vs Dynare++

Hi, all

I am trying to use both dynare and dynare++ to get the higher order approximation of the simple ‘example1.mod’ file included in the official release of dynare.

What confusing me is that dynare and dynare++ give different IRF results: the IRF of 2nd order approximation given by dynare++ is rather choppy.

Why is that so? Am I doing anything wrong?

thanks!

Increase the number of replications