Some Non Technical Questions - Stationarity & Steady States

[danishussalam]

Dear All,

I am working on my Master's thesis and would be very grateful if someone can answer these non-technical questions for me.

1) In dynare user guide chapter 6, there is a solved example Schorfhoide (2000). The equilibrium conditions of the model were non stationary so both the model variables and observed variables were either de-trended or were used in levels to make them stationary.

Usually academic papers now present stationary symmetric equilibrium conditions in the appendix. So for instance if I undertake a replication exercise of any such paper, I will write only those stationary symmetric equilibrium conditions in dynare right? without the need of using de-trending or any other method for making variables stationary?

2) Authors these days also present a steady state section in which they explicitly define steady state equations. Are these the same equations that I am suppose to write when I make an external MATLAB program to calculate exact steady state?

3) When authors say "the group of parameters that are calibrated" they are basically referring to the estimation exercise of models right? either by Bayesian or MLE.

[jpfeifer]

Most answers are in Pfeifer(2013): "A Guide to Specifying Observation Equations for the Estimation of DSGE Models" https://sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf
1. You can bo it both ways, but usually it is preferrable to eliminate unit roots from the model. Hence, use stationarized equilibrium conditions.
2. Yes.
3. No, they are referring to parameters that are fixed and set before estimation. Those parameters are not estimated.
Note that fs2000.mod in the examples of Dynare 4.4.2 contains a steady state model-block and is detrended.

[danishussalam]

Thanks you. I just got the document printed today and will start reading today.