Question about Shock Sizes

Stochastic simulations
1

Hello all!

I recently estimated my first DSGE with dynare and am very excited about it. It’s just the basic new Keynesian from Galí chapter 3 but it is so cool looking at the posteriors/etc.

I have a question about the IRFs. When I set say

var eps_v = 0.01;

for an interest rate shock, is that a 1% shock in the interest rate? I am seeing some conflicting things online, and a few examples I also see code where

v = rho_v•v(-1) + 100•eps_v;

where eps_v = 0.01;

I was told this is simply scaling the IRFs, however when I do this vs removing the 100x and make

eps_v = 1

my order 2 IRFs look very different.

Any help would be greatly appreciated!

2

At first order, everything is linear and you can scale by 100. At higher order that is not allowed. 0.01 is 1 percent, 1 is 100 percent. See e.g. Shocks for IRFs in a third order approximation - #5 by jpfeifer