Detrending data before estimating the DSGE model?

[JX2016]

Hi, everyone. I am just a rookie who has another question.

previously, I posted here with questions about my model:

viewtopic.php?f=1&t=8454

Now I plan to experiment on doing Bayesian Estimation for that model. I wonder if it is necessary to detrend the data before running the bayesian estimations?

If yes, is the regular HP-filter appropriate for my model ? (I know that sometimes 1-sided HP-filter is preferred, sometimes 1st-differecing is better....etc.)

[jpfeifer]

Please have a look at 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. In general, you have to use a transformation of the data that corresponds to the stationary model variables. Therefore, some filter (first difference, one-sided HP filter, linear trend) must be used.