Page 1 of 1

Almost running

PostPosted: Sat May 18, 2013 5:10 pm
by 15291057
I get the following error message from Dynare

"Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 6.644063e-28.
> In dyn_first_order_solver at 217
In stochastic_solvers at 172
In resol at 118
In stoch_simul at 76
In code at 278
In dynare at 120"

The rank condition is however verified. What is this and how do I solve it!! Thanx.
I attach the mod file

Re: Almost running

PostPosted: Tue May 21, 2013 2:23 pm
by fabiover
Hi,

the model is not in (log-)linear form (while you stated that is it: model (linear) )

for instance, you have K = K(-1)*(1-delta) + I, this is the capital accumulation equation in levels

the model runs with Dynare 4.3.0 but it says that "All endogenous are constant or non stationary, not displaying correlations and auto-correlations"

Re: Matrix is close to singular or badly scaled

PostPosted: Fri Mar 28, 2014 1:54 pm
by selima
I just want to ask a general question about the message error

Warning: Matrix is close to singular or badly scaled.
Results may be inaccurate. RCOND = 9.966866e-017.
> In csolve at 90
In dynare_solve at 166
In evaluate_steady_state at 66
In steady_ at 54
In steady at 81
In essaifiscal at 2014
In dynare at 174

I have obtained this error message but after it I obtain the
STEADY-STATE RESULTS, the Residuals of the static equations( which are0),The rank condition is verified ;the MATRIX OF COVARIANCE OF EXOGENOUS SHOCKS,the
POLICY AND TRANSITION FUNCTIONS,the THEORETICAL MOMENTS,the variance decomposition and matrix of correlation. Can the result obtained be correct after this error message? thank you for answering me

Re: Almost running

PostPosted: Fri Mar 28, 2014 2:12 pm
by jpfeifer
Unless you have very good reasons to trust your results (so that the error message is just do to numerical accuracy such as extreme non-linearities), this is a bad sign. What does model_diagnostics say?

Re: Almost running

PostPosted: Fri Apr 04, 2014 3:23 pm
by selima
Sorry for the late but model_diagnostics (M_,options_,oo_) say nothing... I think there is no error.
Thank you Jpfeifer

Re: Almost running

PostPosted: Fri Apr 04, 2014 4:46 pm
by jpfeifer
I looked over the message again. As it only refers to steady state computation, but not to the actual model solution, you can ignore it.

Re: Matrix is close to singular or badly scaled

PostPosted: Fri May 01, 2015 6:27 pm
by selima
Dear Jpfeifer
concerninng the problem that I face with the oscillating IRF , I don't find solution even without the budget constraint and changing the fiscal rule modelisation of government spending and taxe as well as the data .

I change the model .I now work with the one of (Gali and Monacelli ,2005; Monacelli ,2005 ; Justiniano and Preston,2010) with some modifications.
In fact I face this message :

Matrix is close to singular or badly scaled
.Results may be innacuuate


..Is it a bug ? should I change the version? I am working on 4.4.3
.please help me ...this is my last code to finish my thesis..

Re: Almost running

PostPosted: Fri May 01, 2015 7:04 pm
by selima
??? Error using ==> chol
Matrix must be positive definite.
I obtain this error message :

Error in ==> gmhmaxlik at 197
dd = transpose(chol(CovJump));

Error in ==> dynare_estimation_1 at 437
[xparam1,PostVar,Scale,PostMean] = ...

Error in ==> dynare_estimation at 89
dynare_estimation_1(var_list,dname);

Error in ==> article2modif at 520
dynare_estimation(var_list_);

Error in ==> dynare at 180
evalin('base',fname) ;