Cannot find the steady state

Steady state
1

yj.mod (2.6 KB)

Hello,
I am working on a Iacoviello-type DSGE model with fixed-rate loans,
and I keep receiving the error message saying that it is impossible to find the steady state.
Can anybody give me a piece of advice?
Thanks.

Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 1.744234e-27.

In trust_region>dogleg (line 202)
In trust_region (line 112)
In dynare_solve (line 185)
In evaluate_steady_state (line 223)
In steady_ (line 55)
In steady (line 81)
In yj (line 342)
In dynare (line 235)

Residuals of the static equations:

Equation number 1 : -0.013242
Equation number 2 : 0.58711
Equation number 3 : -0.025877
Equation number 4 : -1.0355
Equation number 5 : 0.040951
Equation number 6 : 2.0228
Equation number 7 : -0.004217
Equation number 8 : -0.056312
Equation number 9 : 0.97798
Equation number 10 : -0.1
Equation number 11 : 0
Equation number 12 : 0
Equation number 13 : 0.065044
Equation number 14 : 0.13009
Equation number 15 : -0.8
Equation number 16 : 0.020874
Equation number 17 : 0.025876
Equation number 18 : -0.0048349
Equation number 19 : 0.006919
Equation number 20 : -0.015296
Equation number 21 : 0
Equation number 22 : -1
Equation number 23 : -0.052171
Equation number 24 : -0.087578
Equation number 25 : 0

Error using print_info (line 83)
Impossible to find the steady state. Either the model doesn’t have a steady state, there are an infinity of steady states, or the
guess values are too far from the solution

Error in steady (line 104)
print_info(info,options_.noprint, options_);

Error in yj (line 342)
steady;

Error in dynare (line 235)
evalin(‘base’,fname) ;

4

Are you sure the model is correct?

m=log(MSS-M)-PHIMY*y-PHIMQ*q-PHIMB*lv+eps_m;

seems strange as m will depend on the level of y

5

Thanks for your comment!

In my model, y is log deviation from steady state and exp(y) is level.

After I changed some equations into log-linearized version, my model worked.
(I am not sure why because they depict basically the same relationships among variables. Perhaps Matlab is not perfect and I helped it to solve the problem by changing some equations into log-linearized version.)

Now I have another problem.
I would like to do optimal policy analysis,
but when I change values of policy parameters, most of the time i got the error again saying that it is impossible to find the steady state. My model is very sensitive to parameter values.
Could you advise on why this happens?

Thank you.

6
  1. If y is the percentage deviation from steady state, then exp(y)=1 in steady state. That does not look like a true level to me.
  2. Nonlinear models are very sensitive to the starting values.