margins after hausman-taylor (xthtaylor in Stata13)

[Ourega-Zoe Ejebu]

Hi there,
I am running a Hausman-Taylor regression to investigate the relationship between cheap alcohol unit and other demographic and economic control variables, as per formula below:

xthtaylor Weekly_cheap_units Low_income Intermediate_income Urban_property Comfortable_off Moderate_means Hard_pressed ///
Total_number_of_adults Household_with_children5 Age3040 Age4050 Age5060 Age60p Household_alcohol_level2 Household_alcohol_level3 ///
Year_of_purchase2 Year_of_purchase3 Total_alcohol_spending Total_grocery_spending Beer_lager Cider Fabs Other_wines Spirits Promotion ///
if Social_Class!=6 ,endog(Low_income Household_alcohol_level2 Household_alcohol_level3)

Thereafter, I'd like to use the command 'margins' to obtain:
1) the predicted quantity of purchasing cheap alcohol
2) the probability to purchase cheap alchol

I used the commands as below respectively:
1)margins if Low_income==1 & Household_alcohol_level==0, predict(ystar(0,.))
2)margins if Low_income==1 & Household_alcohol_level==0 , predict(pr(0,.))

However, Stata tells me I cannot use the option 'predict' after using xthtaylor.
Please, could you tell me how to obtain the margins as explained above with a Hausman-Taylor model.

Thank you
Ourega- Zoe Ejebu