Using bayesmh to simulate a tobit model

[Javier Escobal]

Hi all,

I am trying to do some bayesion simulations after estimating a tobit model. Using the bayes prefix command will not work as the command "bayespredict" is not available after "bayes: tobit"
Thus, I am tryng to use bayesmh and estimate a tobit model introducing explicitly the Log-likelihood evaluator
However I am doing something wrong

This is my program using simulated data

clear all
/ * simulated data for tobit estimation */
set seed 12345
set obs 1000
gen x1=invnormal(runiform())
gen x2=invnormal(runiform())
scalar a0=.5
scalar a1=.3
scalar a2=.1
gen e=invnormal(runiform())
gen y=a0+a1*x1+a2*x2+e
gen y1=y
replace y1=0 if y<=0
******************************
capture program drop mytobit
program mytobit
version 17
args lnf beta sigma
quietly replace `lnf'= log(1-normal(`beta'/`sigma')) if $ML_y1==0
quietly replace `lnf'= log((1/`sigma')*normalden(($ML_y1-`beta')/`sigma')) if $ML_y1>0
end
******************************
bayesmh y1 x1 x2, llevaluator(mytobit) prior({y1:}, flat)
bayespredict {_ysim} if y1==0, saving(reps) rseed(123)


Thanks in adavance

Javier