Dynare forums

Hey,

I'm trying to replicate the small open economy model from "Hanna Freystätter (2010): Financial market disturbances as sources of business cycle fluctuations in Finland" for my master thesis.
When I try to simulate it the following error occurs:

STEADY: numerical initial values incompatible with the following equations
Columns 1 through 14

1 2 3 4 6 7 8 10 12 16 22 25 26 27

Columns 15 through 19

30 31 32 34 38

Does anybody know what I did wrong? I attached the mod file and the model equations (first two pages of the pdf file).

I guess the error stems from a change I made in one equation. There is one equation with a function inside: S(.). I didn't know how to implement this function into dynare so I replaced it with a parameter.

I would be very thankful, if anybody could help me.

Cheers,
Henk

Hello Henk,

the error "STEADY: numerical initial values incompatible with the following equations" means that the initial values you provided Dynare to calculare the steay state of your model are not correct. That is, when Dynare calculates the steady state and it plugs in the values you provided, some of the equations are not verified.

If you enter the command "resid(1);" after "steady;" in your mod file, dynare will show you the residuals of the static equations and, if the steady state is correct, they should all be zero. If you changed something in the original model, it is possible that its steady state does not coincide with the steady state of the original model. In that case, you have to recalculate the steady state by hand.

Hope this helps.
Best regards

Hi Mary,

thanks a lot for your help, I appreciate it.

In the end I decided to implement the log-linearized version of the model into Dynare, since I'm more familiar with this procedure. I double-checked the derivations again and again, but I can't replicate all IRFs. It's kind of annoying since I'm 99% sure that there are no mistakes in the linearizations. The only equation I'm worried about is the "Country Borrowing Premium: gamma = exp(-kappa(a(t) - a_ss))" which closes the model. Since a = B/p and B = 0 in the steady state the log-linearized equation "gamma = -kappa*a_ss*a(t)" would become 0, which makes absolutely no sense, so I changed this to "gamma = -kappa*a(t)".

I would appreciate it if you or someone else could help me with this. This time i added the paper and my mod file. Thanks in advance!

Best regards,
Henk

have you resolved the problems?