Final value of minus the log posterior: Question

[econ86]

I want to estimate my DSGE model. I first find a mode of the posterior density using csminwel (i.e. mode_compute=4)

I get the following results:

Initial value of the log posterior (or likelihood): -6278.8237

After several iterations the algorithm ends up with:

Final value of minus the log posterior (or likelihood):2995.576238

Is there any problem with this value? Should it be negative?

[jpfeifer]

No, there is no problem. Why do you think so? The initial value of MINUS the posterior was 6278.8237. We are now searching for the minimum, which was found at 2995.576238

[econ86]

When I add some additional features in the model and run the estimation I get:

Initial value of the log posterior (or likelihood): -3165.4263
Final value of minus the log posterior (or likelihood):-119.039952

So, final value of minus the log posterior (or likelihood) can be either positive or negative number?

[jpfeifer]

Yes, see http://www.dynare.org/phpBB3/viewtopic.php?f=1&t=2847