Dynare forums

Hi fellow Dynare Users,

I am trying to calibrate the standard small open economy (Schmitt-Grohe and Uribe (2003) model) to match the Australian data. At the moment, I have chosen the parameter values such that it matches the data. The theoretical moments produced by the model matches the data quite well. However, for the same set of parameters when I simulate the model the simulated moments are very different. I know that in SOE models the use of Solow residuals for the productivity process are controversial . So, sometimes the persistent and std of TFP are chosen so that they match the serial correlation and std of output in the data, which is what i have done. So, my questions are:

1. how can i choose the parameter values for the productivity process such that it matches the data well I simulate the model?
2. How do I get the empirical moments from 100 simulations of the model for 135 periods?
3. Is it better to look at the theoretical moments or simulated moments when analyzing the performance of a model?

Attached is my mod file.

Thank you so much,

There should not be big difference between theoretical and simulated moments. I personally prefer theoretical moments because they are independent of particular seeds.
If you want replications of your simulation, see the Hansen_1985.mod at https://github.com/JohannesPfeifer/DSGE_mod

Thank you Johannes, you are amazing!