Dear all,

I am trying to run a Bayesian DSGE model but receiving the following error:

[quote]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 458

In dynare_estimation at 70

In linear_model at 428

In dynare at 120 [/quote]

The command window stopped here:

[quote]Error using chol

Matrix must be positive definite.

Error in metropolis_hastings_initialization (line 68)

d = chol(vv);

…[/quote]

The /identification;/ in dynare output as:

[quote]

WARNING !!!

The rank of H (model) is deficient!

psi is not identified in the model!

[dJ/d(psi)=0 for all tau elements in the model solution!]

WARNING !!!

The rank of J (moments) is deficient!

psi is not identified by J moments!

[dJ/d(psi)=0 for all J moments!]

[rhot,et] are PAIRWISE collinear (with tol = 1.e-10) !

ei is collinear w.r.t. all other params!

phipi is collinear w.r.t. all other params!

phiy is collinear w.r.t. all other params!

rhoi is collinear w.r.t. all other params![/quote]

I think I can understand that psi, the inverse elasticity of labor supply, is difficult to identify in the model, especially with limited number of observable. The priors of rhot, ei, phipi, phiy, rhoi, are similar to SW(2007):

[quote] prior mean mode s.d. t-stat prior pstdev

h 0.700 0.7382 0.0063 116.5629 gamm 0.1000

psi 0.500 0.5000 0.2500 2.0000 norm 0.2500

phipi 1.500 1.4385 0.0351 40.9271 norm 0.2500

phiy 0.125 0.1279 0.0502 2.5495 norm 0.0500

rhoa 0.850 0.9944 0.0004 2445.2232 beta 0.2000

rhog 0.850 0.9857 0.0007 1364.3778 beta 0.2000

rhot 0.850 1.0000 0.0000 0.0000 beta 0.2000

rhosg 0.500 0.5000 0.2426 2.0607 beta 0.2000

rhobg 0.500 0.5000 0.2123 2.3552 beta 0.2000

rhoi 0.750 0.9388 0.0047 197.9226 beta 0.1000

standard deviation of shocks

prior mean mode s.d. t-stat prior pstdev

ea 0.400 0.5441 0.0059 91.9176 invg 2.0000

eg 0.400 0.4231 0.0299 14.1266 invg 2.0000

ei 0.400 0.4143 0.0000 0.0000 invg 2.0000

esg 0.400 0.4000 0.0000 0.0000 invg 2.0000

ebg 0.400 0.4000 0.0000 0.0000 invg 2.0000

et 0.400 0.4359 0.0730 5.9671 invg 2.0000[/quote]

Can anyone in this forum please give me some clue to fix the posterior kernel optimization problem?

Thanks a lot for your time.

pb.mod (8.78 KB)

dataUS.xls (33 KB)