I’m new using dynare. I need to estimate a DSGE model using bayesian estimation or maximum likelihood.

My question is regarding the command named “varobs”. I check that this command declares the list of observed endogenous variables which are loaded in the datafile and are used to estimate the parameters.

I check in all examples of estimation of DSGE models using bayesian or maximum likelihood estimation that in “varobs” declares a minimum amount of endogenous variables. Allways declare less variables than the include in “var”.

For example I attach one example from Lawrence Christiano. At the code we have 7 endogenous variables declared in “var”: pie x r rstar da tau dy.
But in “varobs” It declares only: pie dy.

What is the criteria to define the variables to include in “varobs”? Why x r rstar da and tau are excluded from varobs?

Thanks for your time, kind regards. Lautaro cggest.mod (3.48 KB)

I’m relatively new to Dynare myself, so hopefully someone will correct me if I make a mistake.

“varobs” is used to define the variables for which you have data, and want the estimation to take that data into account. The variables listed under “varobs” do not have to be endogenous variables from your model, they only need to have data available to be included in the estimation. One example would be oil import prices in a small open economy - you can use the WTI oil price as a varobs in your model, but it would be declared under varexo if you didn’t have data for it.

The short answer is: varobs includes a (potentially exhaustive) subset of the variables declared in the “var” statement for which you have data and which you want to use for estimation. See also “Remark 7 (Which Observables for Estimation?)” of Pfeifer(2013): “A Guide to Specifying Observation Equations for the Estimation of DSGE Models” sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf.