NaN in s.d. from mode_compute in 8 and 9, but not 6

Estimation ML/Bayesian estimation
1

Hi,

I modified SW(2003) model with money in. The mod file is based on Jpfeifer’s mod file from SW (2007). However, there is a problem with finding a proper mode. The mode’s s.d. shows NaN when mode_compute= 8, or 9, but it works when mode_compute=6. Even I put mode file (from mode_compute= 6) with higher order 8 or 9, it still happens with that.
The model’s identification and simulations all look fine. I am not sure it is the problem with the model’s equations or as I updated version to Mac 4.5.3. Could you help me to quick look please? Thank you very much for the help!
swpooledata1.mat (5.1 KB)
mswmodel.mod (5.8 KB)

Best wishes,
Michael

2

For some reason, various parameters are at the prior bounds. The Hessian cannot be positive definite in this case:

The following parameters are at the prior bound: cprobw, csadjcost, cprobp, csigl, crpie
Some potential solutions are:
   - Check your model for mistakes.
   - Check whether model and data are consistent (correct observation equation).
   - Shut off prior_trunc.
   - Change the optimization bounds.
   - Use a different mode_compute like 6 or 9.
   - Check whether the parameters estimated are identified.
   - Check prior shape (e.g. Inf density at bound(s)).
   - Increase the informativeness of the prior.