Model (linear) The steady state cannot be computed

General DSGE Modeling Replication Attempts
1

Dear all;

I’m trying to use Gali and Monacelli’s (2005) model for a small open economy. However, I wanted to make some changes to fit the features of the economy I’m trying to model, which is an oil-exporting economy. The modifications to the original model are : ​
1/Households can only purchase domestic bonds ( so I abstract from international risk sharing).
2/ There are two types of firms in the economy;
- A nontradable goods-producing firm (which follows Gali and Monacelli’s monopolistic competitive firms with price-setting à la Calvo);
- Oil-producing firm.
3/ Oil is the only exported good.
4/ The monetary policy instrument is the effective exchange rate. (experimental rule)

The problem I face is; when I run my mod file I get an error message saying :
Warning: Matrix is singular to working precision.

In evaluate_steady_state (line 261)
In model_diagnostics (line 74)
In my_model.driver (line 333)
In dynare (line 293)
MODEL_DIAGNOSTICS: The steady-state cannot be computed
MODEL_DIAGNOSTICS: Steady-state contains NaNs

Error using print_info (line 32)
The steady state has NaNs or Inf.

Error in steady (line 102)
print_info(info,options_.noprint, options_);

Error in my_model.driver (line 334)
resid; steady;

Error in dynare (line 293)
evalin(‘base’,[fname ‘.driver’]) ;

I’m working with a log linearized model, and I went through my equations and my mod file several times but can’t find the problem, could you please help me.

I attach here the mod. file I’m working on. SOE_oil_exporting.mod (6.2 KB)

2
  1. Your model fails to compute the steady state due to
ynnat_t = - sigma_y + gama_y * s_t + kapa_y * a_t;

containing a constant terms. Either have the constants consistently in the model or leave them all out. If you delete the -sigma_y it will work.
2. Start from a working model and add one feature at a time. As is, it is impossible to debug.

3

Thank you very much professor Pfeifer for your quick and helpful comment, indeed once I took off the constant term the steady-state was computed.