Using mode_check to fix hessian

[Aldo]

I want to fix negative hessian, in order to do this I've used mode_check option in estimation command. But after runnning some hours it starts to show mode_checks plots for most of parameters (not all) and I am getting this message error:

Error using trustnleqn (line 28)
Objective function is returning undefined values at initial point. FSOLVE cannot continue.

Error in fsolve (line 368)
[x,FVAL,JACOB,EXITFLAG,OUTPUT,msgData]=...

Error in prometheo_steadystate (line 230)

I am suspecting this message is related to the hessian problem because most of mode_check plots are concave.
Please, how could I fix this? I will appreciate your help.

I attach my .mod, the mode computed with mode_compute=2, and mode_checks plots.

[jpfeifer]

This is an error appearing in your steady state file. Apparently, when trying to compute the posterior density for neighboring areas to your mode, the numerical solver fails with the starting value you provide in your file. Without a proper handling for this, Dynare will crash as it currently does. This already provides you some indication of what may be going on. If your mode is a corner solution, the Hessian will not be positive definite.