Minimizing "Craggit" Command Iterations for Double Hurdle Model

[Alexis Adams]

Hello,

I'm currently using the Craggit command to do a Double Hurdle Cragg model analysis on estimates of married couple and single parent expenditures for children that are ages birth through age 17. I previously used the Tobit model because in the data, there are families that have zero expenditures, but the Tobit model failed the test for homoscedasticity and normality, which is required when running the Tobit model.

As an alternative, I ran a Double Hurdle Model, but after running the below Craggit command code, the iterations run for thousands of iterations and does not allow me to get a final result of the model. Is there a way to minimize these iterations? Also, is there a better command to use for the Double Hurdle model?

Thanks in advance for your help.

Alexis


Code:

use "D:\STATA\stata_crc.dta

keep if hw_kids & age_young_member < 16

global xlist income2 income3 num_child1 num_child3 age_y35 age_y68 age_y911 age_y1214 age_y1517

craggit nz_yc_clothing_new $xlist [pweight = finlwt21], second(yc_clothing_new $xlist )

[FernandoRios]

HI Alexis,
THe number of iterations problem could be caused by a many possible problems. I would suggest to start by estimating the model separately first to see if there is any obvious culprit.
Keep in mind that a craggit can also be estimated using a truncated regression and a probit regression. That may give you a better clue of what is going on.
HTH
Fernando