I have the dataset on employment sector of households as;
A. Wage in agri
B. Wage in non agri
C. Self in agri
D. Self in non-agri

using these four variable, I want to compute livelihood diversification strategies for each household:

1. Purely agriculture
(Those households in A & C=1, rest 0)

2. Ag and wage in non agri
(Those households in A &B&C but also those households in A&B and B&C =1, rest 0)

3. Ag and self in non agri
(Those households in A&C&D but also those households in A&C and C&D=1, rest 0)

4. Ag and self and wage in non agri
(Those households in A&B&C&D but also those households in A&B&C and A&C&D=1, rest 0)

5. Purely non farm
(Those households in A & C=1, rest 0)

how should I do it in STATA so that I get 4 mutually non exclusive livelihood strategies.

also, I’m trying to compare the impact of these livelihood strategies on total consumption expenditure. Since livelihood diversification is endogenous, my first thought was to do the regression using 2SLS estimator. But using binary endogenous response variable would lead to forbidden regression. So which estimator should I use that also corrects this endogeneity?