Standard deviation simulated moments

Hello all,

I have a very trivial question and hope someone can help me.
I am conducting a volatility analysis in a standard DSGE model written in exp(form).

For instance, I have a variable, expressed through a Law of motion, which in steady state equals 20% (0.2).

Of course the simulated mean in first order is log(0.2)=-1.6094.
My question is now if the standard deviation was 0.1 (10%), regards that to the “level-value” 20% steady state level, or to the log-value -1.6094.

Thanks very much in advance!
Best regards

Dear Mia,

If the model is log-linearized (as you do with the exp transform, or as you could do with the loglinear option), the reported standard deviation is for the log of the variable.