Backus,Kehoe,Kydland model

[dy2700us]

I am trying to modify basic BKK (1995) model by adding two more state variables , which are capital stocks. Instead of combining domestic and foreign intermediate goods to generate final investment, investment is done at intermediate level, hence there are 4 capital stocks ( 2 per country).

I got error message saying "Rank condition isn't verified" and "Blanchard and Khan condition is not satisfied - rank failure" for some parameter values, while "STEADY convergence problem" for most other parameter values. Could you advice what could be wrong, or whether we need some tricks to solve this type of model? Is the problem about additional state variables? (I solved the basic model and it was fine in Dynare).

MOD file is attached.

Thank you

[admin]

From what I see, you have an analytical expression for the steady state, but you didn't attach dy_bkkcap_ss.m, so I couldn't try to run the model.

Also, the *.mod file is too large for me to see a problem just glancing at it.

You should include the two following lines before steady in your *.mod file

resid(1);
return;

This will give your the residual of each equation evaluated with the values set by dy_bkkcap_ss. When you have a nonzero residual, you know that this particular equation in the dynamic model of the *.mod file doesn't aggree with the analytical steady state.
This can give you a clue of what is wrong.

Best

Michel

[dy2700us]

Thank you very much for the tip. I check the residual and indeed one problem was that there are nonzero residuals. I correct those and add adjustment costs to capital. Now it works! (without adjustment costs, I got error message saying "The model doesn't determine the current variables uniquely" or "Matrix is close to singular or badly scaled.")

Thank you again.

[ebaykal]

Dear Michel,

I am having problem about the Blanchard-Khan Condition. The Matlab gives me the following error

There are 13 eigenvalue(s) larger than 1 in modulus
for 12 forward-looking variable(s)

The rank condition ISN'T verified!

??? Error using ==> print_info at 21
Blanchard Kahn conditions are not satisfied: no stable equilibrium.

I tried
resid(1);

to see which equation is causing the problem. When I change parameter eta from zero to something 0.009 I get BK condition problem. The mod file works fine with eta=0.00 or at most eta=0.003 but not larger. The funny thing that residuals are the same when mod file works as those when it does not work.

I attached the mod file. I do not have a clue as to how to solve the problem. I greatly appreciate any help or suggestion.


P.S. I also copied you answer to this issue from a post in forum regarding the same issue.

From what I see, you have an analytical expression for the steady state, but you didn't attach dy_bkkcap_ss.m, so I couldn't try to run the model.
Also, the *.mod file is too large for me to see a problem just glancing at it.

You should include the two following lines before steady in your *.mod file

resid(1);
return;

This will give your the residual of each equation evaluated with the values set by dy_bkkcap_ss. When you have a nonzero residual, you know that this particular equation in the dynamic model of the *.mod file doesn't aggree with the analytical steady state.
This can give you a clue of what is wrong.

Best

Michel