Observable variables input in DSGE estimation


I read a PDF file by professor Johans Pfiefer about the observable variables in DSGE models estimation, in a simple form we insert observable variables names in a typical form of X_obs in endogenous variables section, and in the model section then we insert names of these variables in the form of

varobs X_obs

but my question is that in some log-linear models I can see measurement equasions in the form of growth rate with steady state value for GDP, consumption,Investment ,…

I used observable variables in the model in the form of

varobs Y C I

and I didn’t use Y_obs C_obs I_obs in endogenous variables section and model section. When I run this DSGE model these is not any problem or error in Dynare.

My question is that my work in employing of observable variables in the model is right or not?? (my DSGE model is log-linear and these is not unit root in shocks equations).

  1. Sorry, but I am not following here. What do you mean with

Growth rates are differences over time, not relative to steady state.
2. If you have a log-linear model where variables are percentage deviations from their steady state and your data is also in percentage deviations from your concept of steady state in the data, then you don’t need separate equations/variables, because the objects already defined in the model are directly observed.

NEWRBCDATAFILE.csv (3.7 KB) RBC_MA2020.mod (7.3 KB)

I sent you professor my DSGE file and my data file. My data are GDP, consumption and Capital stock. The data are quarterly and seasonally adjusted. Then I transformed them to the logarithmic form and after that I derived cycle component of these data with HP filter then I included this cycle component in the NEWRBCDATAFILE.csv

ABCD_test.m (3.7 KB)

As you can see I didn’t use Y_obs C_obs K_obs in the endogenous variables section or in model section.I can run this dynare code in MATLAB without any problem at all.

My question is that my work is true or not?

In some models researchers write the measurement equasions of the model in this form:


in the above equations Y-Y(-1) is GDP growth rate and lambday is steady state of the growth rate as I know.