Dynare relative irf

Hi, I try to use dynare to relative irf to plot impulse response function. But it seems that the result is wrong.
For example, say two variables k and n.
If k change 68%, and n change 32% then I define knratio=k/n
Dynare told me that this knratio change -3.1%

I thought that use dynare relative irf should be the same as log linearization. So that I can put my model into the code directly as it shows in model.
But it seems that it is wrong.
Is there anyone who has the same problem as me?

relative_irf only normalizes the variables by the shock size so that the shock sizes are comparable. It does not rescale the other variables by their steady state values to make them percentage deviations!

The respective Dynare code for this option used in stoch_simul.m is

if options_.relative_irf y = 100*y/cs(i,i); end
where y is the standard IRF without relative_irf and cs is the cholesky factorization of the covariance matrix of the shocks, i.e. the IRFs are scaled by the standard deviation of the respective shock. If you want the model to be log-linearized, you have to put everything in exp() (search the forum).

So, does it mean that relative irf could be used when comparing one variable’s responses to different kind of shocks while standard irf could be used when explaining variables’ responses to one shock?

Yes and no. It always depends on the interpretation you are after. If you want to have the same unit shock, you can use it. If you want to compare one standard deviation shocks, then not.