error file mod

[dominik]

Using the unstable version, I get the same error.


ESTIMATION_CHECKS: There was an error in computing the likelihood for initial parameter values.
ESTIMATION_CHECKS: You should try using the calibrated version of the model as starting values. To do
ESTIMATION_CHECKS: this, add an empty estimated_params_init-block with use_calibration option immediately before the estimation
ESTIMATION_CHECKS: command (and after the estimated_params-block so that it does not get overwritten):


Error using chol
Matrix must be positive definite.
Error in sequential_importance_particle_filter (line 84)
Q_lower_triangular_cholesky = chol(Q)';
Error in non_linear_dsge_likelihood (line 298)
LIK = feval(DynareOptions.particle.algorithm,ReducedForm,Y,start,DynareOptions.particle,DynareOptions.threads);
Error in initial_estimation_checks (line 120)
[fval,junk1,junk2,a,b,c,d] =
feval(objective_function,xparam1,DynareDataset,DatasetInfo,DynareOptions,Model,EstimatedParameters,BayesInfo,BoundsInfo,DynareResults);
Error in dynare_estimation_1 (line 157)
oo_ =
initial_estimation_checks(objective_function,xparam1,dataset_,dataset_info,M_,estim_params_,options_,bayestopt_,bounds,oo_);
Error in dynare_estimation (line 104)
dynare_estimation_1(var_list,dname);
Error in EST_LEV_XIshock_8807old52 (line 1163)
oo_recursive_=dynare_estimation(var_list_);
Error in dynare (line 223)
evalin('base',fname) ;

[dominik]

I attached the wrong data file above, sorry.

[jpfeifer]

We are looking into it and will get back to you.

[dominik]

many thanks!

[StephaneAdjemian]

Hi, One problem is that you declare shocks which are not active (ie with zero variance). This explains
why matrix Q is not full rank. You should first try to remove these shocks from the mod file.

Best,
Stéphane.

[dominik]

thanks a lot Stefane and Johannes! that solved the issue!

[dominik]

Unfortunatley another issue came up. Now the routine stops at the mode. This also sometimes happens for first order but there the mode is typically obviously wrong. Here im not sure why i get what i get.

1210 , 16941 : 1.0000000000000e+08 +(1e-08,1e+00) | 1.05e+01 | 35:1.2e-04 18:8.6e-04
1220 , 17081 : 1.0000000000000e+08 +(3e-08,1e+00) | 1.08e+01 | 35:1.2e-04 18:8.7e-04
1230 , 17221 : 1.0000000000000e+08 +(1e-08,1e+00) | 1.10e+01 | 35:1.3e-04 18:9.2e-04
1240 , 17361 : 1.0000000000000e+08 +(7e-08,1e+00) | 1.11e+01 | 35:1.3e-04 18:9.4e-04
1250 , 17501 : 1.0000000000000e+08 +(6e-08,1e+00) | 1.13e+01 | 35:1.1e-04 18:8.5e-04
1254 , 17557 : 1.0000000000000e+08 +(3e-08,1e+00) | 1.13e+01 | 35:1.1e-04 18:8.4e-04
#Fevals: f(returned x) | bestever.f | stopflag
17558: 1.00000000000e+08 | 1.00000000000e+08 | warnequalfunvalhisttolhistfun
mean solution: +6.9e-01 +1.7e+00 +2.1e+00 +4.9e+00 +9.4e-02 -7.4e-02 +8.6e-01 +5.4e-01 +7.4e-01 +9.2e-01 +1.5e-01 +1.5e+00 +1.5e+00 +1.2e+00 +7.7e+00 +2.0e+00 +3.7e-01 +1.7e+00 +2.6e+00 +4.4e-01 +1.1e+00 +9.6e-01 +9.6e-01 +5.3e-02 +1.1e-01 +3.9e-01 +1.0e+00 +9.8e-01 +7.9e-03 +2.8e-02 +3.6e-01 +2.1e+00 +4.7e-01 +4.9e-01 -4.9e-02 +8.6e-01 -1.2e-01 +1.4e+00 +1.1e+00
std deviation: 6.2e-04 6.9e-04 5.8e-04 6.4e-04 6.3e-04 6.5e-04 6.0e-04 7.8e-04 6.7e-04 5.0e-04 6.3e-04 6.6e-04 6.4e-04 6.2e-04 6.4e-04 5.8e-04 7.3e-04 8.4e-04 7.5e-04 6.7e-04 6.0e-04 7.0e-04 5.8e-04 6.3e-04 6.7e-04 7.3e-04 6.6e-04 7.0e-04 8.0e-04 6.5e-04 5.4e-04 7.4e-04 5.9e-04 6.6e-04 1.1e-04 6.7e-04 5.7e-04 7.5e-04 7.9e-04
use plotcmaesdat.m for plotting the output at any time (option LogModulo must not be zero)

Objective function at mode: 100000000.000000

POSTERIOR KERNEL OPTIMIZATION PROBLEM!
(minus) the hessian matrix at the "mode" is not positive definite!
=> posterior variance of the estimated parameters are not positive.
You should try to change the initial values of the parameters using
the estimated_params_init block, or use another optimization routine.
Warning: The results below are most likely wrong!
> In dynare_estimation_1 (line 694)
In dynare_estimation (line 89)
In EST_LEV_XIshock_8807old46 (line 1071)
In dynare (line 180)
Warning: Matrix is singular to working precision.
> In dynare_estimation_1 (line 709)
In dynare_estimation (line 89)
In EST_LEV_XIshock_8807old46 (line 1071)
In dynare (line 180)

RESULTS FROM POSTERIOR ESTIMATION
parameters
prior mean mode s.d. prior pstdev

sigma_A 0.000 0.4506 Inf unif 1.0000
sigma_B 25.000 1.8159 Inf unif 14.4338
sigma_G 0.000 2.1110 Inf unif 10.0000
sigma_I 0.000 5.4240 Inf unif 10.0000
sigma_P 0.000 0.0669 Inf unif 1.0000
sigma_R 0.000 0.1172 Inf unif 1.0000
sigma_W 0.000 0.5674 Inf unif 1.0000
rho_r 0.500 0.3594 Inf beta 0.2000
rho_w 0.500 0.7337 Inf beta 0.2000
rho_a 0.500 0.7609 Inf beta 0.2000
rho_b 0.500 0.1723 Inf beta 0.2000
rho_g 0.500 0.8273 Inf beta 0.2000
rho_p 0.500 0.9999 Inf beta 0.2000
rho_i 0.500 0.6673 Inf beta 0.2000
phi_i 4.000 7.4144 Inf norm 1.5000
sig_c 1.500 2.2438 Inf gamm 0.3750
h 0.700 0.5208 Inf norm 0.1000
xi_w 0.500 0.6226 Inf beta 0.1000
sig_l 2.000 2.3353 Inf gamm 0.7500
xi_p 0.500 0.4535 Inf beta 0.1000
gamma_w 0.500 0.4635 Inf beta 0.1500
gamma_p 0.500 0.3232 Inf beta 0.1500
corGA 0.500 0.6891 Inf beta 0.2000
ma_p 0.500 0.8013 Inf beta 0.2000
ma_w 0.500 0.8132 Inf beta 0.2000
alph 0.300 0.3161 Inf norm 0.0500
mup 1.250 1.3938 Inf norm 0.1200
ctrend 0.400 0.3537 Inf norm 0.1000
lbar 0.000 -0.1587 Inf norm 2.0000
betta_rate 0.250 0.4565 Inf norm 0.1000
pbar 0.620 0.6055 Inf gamm 0.1000
r_pie_est 1.500 1.9198 Inf norm 0.2500
rho 0.750 0.7225 Inf beta 0.1000
r_y_est 0.120 0.1946 Inf norm 0.0500
xiparam 0.015 -0.0128 Inf norm 0.0100
scrap 0.500 0.8293 Inf norm 0.1000
kt 0.080 0.1018 Inf norm 0.0500
sigma_xi 5.000 0.6678 Inf unif 2.8868
rho_xi 0.500 0.9040 Inf beta 0.2000


Log data density [Laplace approximation] is NaN.

Error using dynare_estimation_1 (line 769)
Estimation:: Mode value(s) of sigma_A are outside parameter bounds. Potentially, you should
set prior_trunc=0.
Error in dynare_estimation (line 89)
dynare_estimation_1(var_list,dname);
Error in EST_LEV_XIshock_8807old46 (line 1071)
dynare_estimation(var_list_);
Error in dynare (line 180)
evalin('base',fname) ;

[dominik]

Ok found another problem, unlike the Kahlman filter the particle filter does not accept missing values in the data file.

[jpfeifer]

Yes, that is currently a restriction