Choice of units and second order approximation in Dynare

[jmohimon]

Dear Dynare users,

In a non-linear model expressed in exp-logs, variables are returned as percentage deviation from steady state. If I read a 0.5 in an IRF or in simulated variable, should I interpret it as a 0.5% deviation from steady state or as a 50% deviation from steady state?

I have seen both in linearized models, which I think does not matter as long as we keep consistency between shocks size and variables interpretation. However, with a second or higher order of approximation, since the size of shocks matters, I would guess it is important to stick to Dynare’s choice of units.

Kind regards,
Jolan

[jpfeifer]

You might want to have a look at See "Remark 10 (Scaling With a Factor 100)" in Pfeifer (2013): "A Guide to Specifying Observation Equations for the Estimation of DSGE Models" at https://sites.google.com/site/pfeiferec ... ations.pdf

The choice of 0.01 as 1% is not a Dynare convention, but mathematics.

[jmohimon]

Thank you!