Dear Professor,
I am trying to calibrate 5 parameters in the model by matching moment using SMM. I get the following warning sign, and the estimated result show that model moments are different from empirical moments. What am I doing wrong?
OPTIMIZATION PROBLEM!
(minus) the hessian matrix at the “mode” is not positive definite!
=> 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 optimization routine.
Thank you!
Best,
Jeremy
emp_moments.mat (2.4 KB)
model5.mod (2.6 KB)
You did not provide the steady state helper.
I apoligize Professor. Here they are.
ss_helper_bigK.m (492 Bytes)
ss_helper_bigR.m (498 Bytes)
ss_helper_k.m (493 Bytes)
ss_helper_kstar.m (499 Bytes)
ss_helper_S.m (496 Bytes)
ss_helper2.m (596 Bytes)
Are you sure the matching between data and model is correct. The mean for your observables is quite far from the steady state values.
Do you mean there is a problem with my steady state value? Or that my initial parameters are too far? Or that my model doesn’t match data that well?
It can be both. Are you estimating parameters that affect the steady state of the two observables?