Data of unobservables

Hello together,

I have one small question. I need the data my model calculates for unobservable endogenous variables with their respective timestamp. Is there a possibility to export these data for further analysis?

Thank you!

What exactly is the question. With the smoother-option you can request smoothed unobserved variables. Or is it about exporting the results from Matlab to another program?

Thanks for the quick reply. It is more about whether it is possible after the ML estimation with e.g. consumption, inflation and federal funds rate as inputs to get the time series of e.g. wages according to the model so that their development can be compared with the real one.

Yes, that is an output of the smoother option.

Thank you, that was exactly what I was looking for. Is there a possibility to export this time series to excel?

Matlab offers xlswrite.

Thank you very much, everything worked. I have one small question, is e.g. w_obs or simply w the right time series simulation of the wage development?

That depends on what you are trying to do. Most of the time. w_obs is the equivalent to the data.

1 Like

Okay perfect, thank you. If my input data, e.g. c_obs only contains the cyclical component of c, will w_obs also contain only the cyclical component?

Most likely yes, but I don’t know your model.

1 Like

Thank you! :slight_smile:

I have one last question. Somehow I cannot access the time series of the endogenous variables of the model in my workspace. Or are they hidden somewhere? As I open the oo_SmoothedVariables file, I am only seeing the observed variables!

Which set of variables did you request at the beginning of estimation? Only the observed ones?

I need the smoothed time series of lambda and the interest rate for further computations. I just inserted “smoother” into my estimation command. But if I open oo_ in the workspace, I only find time series related to my data inputs.

If you have a lot of variables in your model, Dynare will ask you at the beginning of estimation for which variables you want the smoother to be computed. Did you select

[2] Consider all the observed endogenous variables

instead of

[1] Consider all the endogenous variables.


After fixing a small issue at the estimation command, I finally got the question regarding the variables. Thank you very much, Prof. Dr. Pfeifer.