Matrix close to singular or badly scaled

Stochastic simulations
1

Hi all,

while solving and simulating a New Keynesian open economy DSGE with incomplete pass-through,
I obtained a multiple of warnings that reads as follows:

""Warning: Matrix is close to singular or badly scaled.
Results may be inaccurate. RCOND = 1.150177e-017.

In th_autocovariances at 182
In disp_th_moments at 37
In stoch_simul at 73
In Sisay_OENK at 345
In dynare at 132"".

What should I do to overcome this problem? I attached the .mod file.
Any tip is highly appreciated.

(BTW, I obtained a solution to my previous query from a an earlier response posted by jpfeifer Oct 4/2010).

With best regards,

Boja :confused:
Small Open economy NK.mod (4.57 KB)

2

Hi all,
(Sorry, I re-posted the mod file).
while solving and simulating a New Keynesian open economy DSGE with incomplete pass-through,
I obtained a multiple of warnings that reads as follows:

""Warning: Matrix is close to singular or badly scaled.
Results may be inaccurate. RCOND = 1.150177e-017.

In th_autocovariances at 182
In disp_th_moments at 37
In stoch_simul at 73
In Sisay_OENK at 345
In dynare at 132"".

What should I do to overcome this problem? I attached the .mod file.
Any tip is highly appreciated.

(BTW, I obtained a solution to my previous query from a an earlier response posted by jpfeifer Oct 4/2010).

With best regards,

Boja :confused:
Boja_SOENK.mod.mod (4.57 KB)

3

This is only a warning so it is not necessary a source of worry.

Are you getting sensible results?

4

Hi Sebastien, Thank you so much!
Some of the impulse response functions are not converging back to their
initial value (=0) after a shock.
The whole idea is this:
I am trying to compare two models -

  1. standard production functions in both tradable and non-tradable sector - this works well!
  2. The second model with a production function in non-tradable sector the ability of firms
    to import intermediate inputs is constrained by availability of foreign exchange.
    It is for this model that some of the IRFs fail to converge.

I really appreciate any tip.

Boja