Dynare forums

Hi Guys,
In all the papers concerning the Bayesian estimation approach, there are always some parameters which need to be fixed (to be calibrated rather than be imposed some prior and be estimated). Is there any efficient rule to judge which parameters to be calibrated or to be estimated?

Hi
There is nothing set in stone about which parameters to estimate and which ones to fix. Typically, people fix parameters for which we have some info from long run sample means. For example, the great ratios and the share of capital in production or the share of imports in GDP. Since we use filtered data (or growth rates) for the estimation, it is unlikely that these long run parameters are estimated at values which make sense. So it is good to fix them. Estimating these parameters can yield pretty crazy results. From my own experience, estimating the openness parameter in a DSGE model for a very open economy, drove the parameter to almost zero. So I fixed it at the sample mean in the data.

Reuben

See also the Remark "Calibration vs. Estimation" 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.

Hi Reuben,
I'm grateful for your help, thanks very much!

Yours
Zhiteng

reubenpjacob wrote:Hi
There is nothing set in stone about which parameters to estimate and which ones to fix. Typically, people fix parameters for which we have some info from long run sample means. For example, the great ratios and the share of capital in production or the share of imports in GDP. Since we use filtered data (or growth rates) for the estimation, it is unlikely that these long run parameters are estimated at values which make sense. So it is good to fix them. Estimating these parameters can yield pretty crazy results. From my own experience, estimating the openness parameter in a DSGE model for a very open economy, drove the parameter to almost zero. So I fixed it at the sample mean in the data.

Reuben

Hi, Professor Pfeifer,
it's nice of you to summarize these, which would definitely be useful in modelling and coding. Thanks very much.

Yours
Zhiteng

jpfeifer wrote:See also the Remark "Calibration vs. Estimation" 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.