Posterior kernel optimization problem

I am doing DSGE model in Dynare for my dissertation, and I am stuck with some problems that I am trying to figure why for few days. But I still can’t fix it.

Two of my main questions are:

1.I got this:

(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 optimisation routine.’

2. In the RESULTS FROM POSTERIOR ESTIMATION section, the standard errors I got are 'NaN'
 And I don't understand why there are no standard errors.


prior mean mode s.d. prior pstdev

rho_p 0.500 0.5999 NaN beta 0.2000
rho_m 0.500 0.5999 NaN beta 0.2000
rho_x 0.500 0.7781 NaN beta 0.2000

I attached the mod file and my data.

Thank you so much. (7.55 KB)

Hi Wayne,

Regarding your question 1, have you checked the previous posts in this forum? Lately people have asked a lot of questions (including myself) about the problem of finding the mode. Here are two links:

[Mode check plots look ok, but Hessian is still negative)

[Hessian computation)

Also, have you tried running different mode finders, via the mode_file=your_mode_file option? Sometimes it can take a while and running one optimiser is sometimes not enough.