Bayesian estimation is not working

Estimation ML/Bayesian estimation
1

Dear Prof,

Greetings.

I am unable to figure out what’s wrong with the model I am working on. One of the warnings I get is:

Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 1.370991e-17.
I have extensively checked the forum for topics with respect to this but still not able to find the error(s)
testest33.mod (10.6 KB)
ecadata2.xlsx (21.1 KB)
.

I have attached the mod. file and data (I used the one-sided hp filter).

Thank you.

2

That seems to be due to pathological parameter draws. I would not worry too much about those right now. I would switch to mode_compute=5 and investigate the problems in the mode_check-plots. One issue seems to be an identification problem:

MOMENTS (ISKREV, 2010):
!!!WARNING!!!
The rank of J (Jacobian of first two moments) is deficient!

[rrho_tp,pssi] are PAIRWISE collinear!

3

Thank you for the response.

I removed Pii_obs as an observation.

The model still runs with Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 1.370991e-17. (both at mode_compute=6 and mode_compute=5)

I have tried several attempts:

  1. Tried to vary the observation equations- as log difference and steady state difference.
  2. Tried to change priors
  3. used different mode finders

I will try to modify the equations next
testest34.mod (11.5 KB)
ecadata2.xlsx (21.1 KB)