Blanchard Kahn conditions are not satisfied: indeterminacy

[pinopino]

Dear all,
I was running a DSGE model based on Dynare. But I come up with the error: "Blanchard Kahn conditions are not satisfied: indeterminacy".
Please find the mod file in the attachment.
I will be glad if one can help me,

Many thanks in advance,

Pinar

[jpfeifer]

You should start with Gali's textbook, chapter 7 and make sure that that model runs. Then start modifying it.
In particular, the last part of your dynamic IS curve looks strange.

[ku_francois]

Hi Everyone,
I am new to dynare and struggling. I have a linear model that works fine when i model the AR(1) process for government expenditure as g = Phi_G*b(-1)+eps_s.
Where (B) is short term bonds, but once I model government expenditure as g = Phi_G*g(-1)+eps_s I get the message below from matlab. I would appreciate help. I really need the help. I have attached the .mod file.

??? Error using ==> print_info at 43
Blanchard Kahn conditions are not satisfied: indeterminacy

Error in ==> stoch_simul at 81
print_info(info, options_.noprint);

Error in ==> threeyr at 307
info = stoch_simul(var_list_);

Error in ==> dynare at 120
evalin('base',fname) ;

Thanks in advance

[jpfeifer]

When I use the mod-file with

Code: Select all

g = phi_G*(g(-1))+eps_s;


instead of

Code: Select all

g = phi_G*(g(-1)+b_l(-1))+eps_s;


it stops with the same BK violation. This suggests that there is still a problem with your bond-pricing equation.

[ku_francois]

Thanks so much for the prompt reply, I appreciate it. I sent the wrong file the last time. Same question holds and please find attached new file.
This one runs but doesn't with

Code: Select all

g=Phi_G*g(-1)+eps_s

. My question is, is it ok to specify government expenditure process as

Code: Select all

g=Phi_G*b(-1)+eps_s

, where b(-1) are bond holdings.

Thanks for you time.

[jpfeifer]

Again, I don't know your model. You need to think harder about the economic intuition of what is going on in your model.

To me it looks a bit as if Ricardian equivalance is at work. The government budget is always balanced by implicit lump-sum taxation and only the present value of debt matters. Thus, one cannot determine the split between taxes and bonds as they are perfect substitutes. This would give indeterminacy. By making spending a function of debt, you suddenly add an equation determining debt.

[ku_francois]

I am really appreciative of your time. Thanks and will be working on the intuition. So technically if I can find a sound economic interpretation of the formulation for the equation determining debt, the model is fine?

Thanks again.

[jpfeifer]

You might want to look into the literature on fiscal rules. Having spending depend on debt is pretty commom. See e.g. my JEDC paper.

[ku_francois]

Dear jpfeifer,

Thanks so much for this. I really appreciate the effort and time.
To be honest, I don't know how you guys keep track of all these questions and answer each one of them. Thanks for all your patience and especially the time.

Well appreciated