estimating multiple-hurdle model simultaneously

[Anwesha Chakrabarti]

Dear STATA users,

I have the following set-up of my data (see attachment) and trying to find what command to use that suits my case: First stage is a binary decision to participate or not (y1= 0, 1). Second stage is an ordered decision, whether to add to a baseline donation once, twice or thrice (y2=0,1,2,3). And the third decision is the donation amount conditional on the four possible outcomes from the second stage (y= y3, y4, y5, y6). I want to estimate the three stages together. I read that this could be done using gsem and cmp in STATA. I have some success with cmp (by David Roodmand : Estimating fully observed recursive mixed-process models with cmp Stata J., 11 (2) (2011), pp. 159-206).

I used the following command:

 
 cmp(respond = x1 x2 x3)(multiple_add = x2 x3 x4)(ln_contribution= x3 x4 x5), ind($cmp_probit $cmp_oprobit $cmp_cont) nonrtolerance vce(robust)

I get the estimation result for the three stages. But the third stage only produces one outcome y. Is there a way to estimate the model so that I can get y3, y4, y5 and y6 separately in the third state of estimation? Is it possible to combine -if- command in the third stage somehow? I wold greatly appreciate any help with this.

Thanks,
Anwesha

[Emmanuel Mwau]

I have a dilemma over the application of the hurdle model.
Here is a description of the data that I want to analyze:
1. Binary info on household purchase of different milk products
2. If products is purchased above, I have the quantities purchased and prices
3. Of the quantities purchased, who consumed in the household and the quantities for each member (Intrahousehold allocation)

My objective is to analyze this data to produce empirical evidence on determinants of milk purchases and intrahousehold allocations - I have sociodemographic household information too

Previous studies have employed double hurdle to analyse participation and determinants of the purchase, but my worry is how to go beyond this stage to look at the intrahousehold allocations, with an interest of how sociodemographic factors influence children access to milk

I welcome suggestions on how to go about this

Thanks and Regards
Emmanuel