(minus) the hessian matrix at the "mode" is not positive def

[yysun1986]

Hi, I am a student from japan. I am using dynare 4.3.3 about DSGE model. Now I got a problem.
The problem is the hessian matrix at the "mode" is not positive definite!
I want to get some help from you. Thank you very much!

This is my error message:

Improvement on iteration 13 = 0.000000000
improvement < crit termination
Objective function at mode: 8920.745465

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 at 480
In dynare_estimation at 70
In bkk at 291
In dynare at 120

MODE CHECK

Fval obtained by the minimization routine: 8920.745465

Most negative variance -739270.224944 for parameter 31 (rho_p = 0.996524)

RESULTS FROM POSTERIOR MAXIMIZATION
parameters
prior mean mode s.d. t-stat prior pstdev

sigma 1.000 1.4950 0.0000 0.0000 gamm 0.3750
theta 0.700 0.4029 0.0530 7.5992 beta 0.1500
chi 2.000 4.4011 0.3858 11.4083 gamm 0.7500
zetainv 4.000 7.1003 0.6431 11.0406 gamm 1.5000
mu 1.000 2.0975 0.0000 0.0000 gamm 1.0000
phi 0.075 0.0205 0.0001 258.2868 gamm 0.0125
gammaw 0.500 0.2970 0.0000 0.0000 beta 0.2500
xiw 0.375 0.5621 0.0095 59.3178 beta 0.1000
gammap 0.500 0.5934 0.0000 0.0000 beta 0.2500
xip 0.375 0.7411 0.0000 0.0000 beta 0.1000
lambdap 0.150 0.1809 0.0000 0.0000 gamm 0.0500
zstar 0.190 0.1730 0.0000 0.0000 gamm 0.0500
lstar 0.000 -0.0001 0.0500 0.0012 norm 0.0500
pistar 0.175 0.1746 0.0420 4.1532 gamm 0.0500
rstar 0.498 0.5175 0.0161 32.0470 gamm 0.0500
phir 0.800 0.6979 0.0000 0.0000 beta 0.1000
phipi 1.700 1.6919 0.0000 0.0000 gamm 0.1000
phiy 0.125 0.0016 0.0001 19.8237 gamm 0.0500
rho_z 0.500 0.0974 0.0129 7.5301 beta 0.2000
rho_b 0.500 0.6994 0.0136 51.5989 beta 0.2000
rho_i 0.500 0.5059 0.0000 0.0000 beta 0.2000
rho_g 0.500 0.9204 0.0000 0.0000 beta 0.2000
rho_w 0.500 0.2739 0.0206 13.2918 beta 0.2000
rho_p 0.500 0.9965 0.0000 0.0000 beta 0.2000
rho_r 0.500 0.4886 0.0135 36.0770 beta 0.2000
standard deviation of shocks
prior mean mode s.d. t-stat prior pstdev

e_z 0.500 1.6069 0.0306 52.5631 invg Inf
e_b 0.500 3.1015 0.0163 189.9076 invg Inf
e_i 0.500 4.8005 0.2274 21.1091 invg Inf
e_g 0.500 0.4763 0.0016 294.7348 invg Inf
e_w 0.500 0.5886 0.0148 39.6924 invg Inf
e_p 0.500 0.2403 0.0000 0.0000 invg Inf
e_r 0.500 0.1181 0.0038 31.4832 invg Inf

Log data density [Laplace approximation] is -9023.391922.

??? Error using ==> chol
Matrix must be positive definite.

Error in ==> metropolis_hastings_initialization at 68
d = chol(vv);

Error in ==> random_walk_metropolis_hastings at 69
[ ix2, ilogpo2, ModelName, MhDirectoryName, fblck, fline, npar, nblck, nruns, NewFile, MAX_nruns, d ] = ...

Error in ==> dynare_estimation_1 at 931
feval(options_.posterior_sampling_method,objective_function,options_.proposal_distribution,xparam1,invhess,bounds,dataset_,options_,M_,estim_params_,bayestopt_,oo_);

Error in ==> dynare_estimation at 70
dynare_estimation_1(var_list,dname);

Error in ==> bkk at 291
dynare_estimation(var_list_);

Error in ==> dynare at 120
evalin('base',fname) ;

can you help me , thanks!

[jpfeifer]

Search the forum. Short answer: test identification, try mode_compute=6,8, or 9 or try more informative priors.

[yysun1986]

sir,in my mod file I did not use the "mode_compute" , the data about informative priors is only these.

[yysun1986]

Thank you very much for your help, my problem is solved!