Estimation with non stationary variables

I am trying to estimate a basic Forecasting and Policy Analysis System (FPAS) with log of GDP, CPI and level of nominal interest as observables. I have tried to follow Dynare example of fs2000_nonstationary.mod. While fs2000 files are executed finely, my program gives following error:

terminate called after throwing an instance of ‘SymbolTable::NoTypeSpecificIDException’

Data and .mod files are attached. Help will be much appreciated!
data.xls (35 KB)
non_stationary.mod (2.1 KB)

This is a bug. You can keep track via Investigate "terminate called after throwing an instance of 'SymbolTable::NoTypeSpecificIDException'" (#113) · Issues · Dynare / preprocessor · GitLab

Thank you for spending time on query!
I have estimated a stationary version of the model by doing appropriate changes in code and observable variables. Data and .mod files are attached.

Can we safely assume that estimated posterior distributions of parameters would remain same whether estimated through a non-stationary or stationary model?
data.xls (32.5 KB)
stationary1_without_u.mod (4.4 KB)

Yes, if you do a proper detrending, you will get the same detrended FOCs as Dynare and estimation results will be identical.

1 Like