Dynare forums

Thank you for your help!!!!

The nonlinear fiscal rule is the following:
gt=gt-1-phib(bgyt-1-bgybar)-phibb(bgyt-bgyt-1)

where g =government spending
bgy =public debt-to-gdp ratio
bgybar=ss public debt-to-gdp ratio

I tried to simplify the rule and I also change the coefficient values but nothing works. The problem appears when I have public debt in period t and t-1 with government spending at t-1. Government spending shows up in the market clearing condition, and in the government budget constraint only.

This suggests the presence of a conceptual problem. Try to work with the easiest version of your model and the most simple rule. I am also puzzled by the unit root between

Code: Select all

gt=gt-1+...

That part does not seem standard.

Thanks for your response!!!

Thank you for your response!!!

I think that there is something weird in the calibration. The time period in the model corresponds to one quarter in the data. Following some papers I consider b/Y=0.8*4. Then it seems that government spending react too much to public debt... if I consider 0.8 only everything works but then there is no difference if I look at the IRF with the case in which there is no public sector.

I attach the non-linear version of the model.
Thanks!

Getting a version with feedback rules that satisfies the BK conditions can be tricky (as you are experiencing). That was the reason I suggested to start with a simpler version where it is easier to find the uniqueness and determinacy parameter region. And yes, too much feedback can indeed happen.

I resolved the model again and I get the following error message. I am very thankful for your help.

There are 3 eigenvalue(s) larger than 1 in modulus
for 4 forward-looking variable(s)

The rank condition ISN'T verified!

Error using print_info (line 45)
Blanchard Kahn conditions are not satisfied: indeterminacy

Error in stoch_simul (line 98)
print_info(info, options_.noprint, options_);

Error in stackelbergml (line 329)
info = stoch_simul(var_list_);

Error in dynare (line 180)
evalin('base',fname)

Again, I can only tell you to check the model timing, specificiation and parameterization.