I'm conducting about gender bias on household education expenditure, but i don't know exactly Stata command to calculate The outlay equivalent ratios using Deaton method, so I calculate manually. And I don't know if the calculation steps are correct or not?
And another problem I don't know what command the author used to calculate the result table 2 and 3? (as attached image) 
reg share_edu male_children female_children x1 x2 x3, robust
scalar alpha_edu= _b[_cons]
scalar beta_edu= _b[ln_ppc_exp]
scalar eta_edu= _b[ln_hhsize]
scalar gamma_boy_edu= _b[male_children]
scalar gamma_girl_edu= _b[female_children]
matrix V_edu= e(V)
sum share_edu
scalar mean_w_edu= r(mean)
sum male_children female_children
scalar mean_male_children = r(mean)
scalar mean_female_children = r(mean)
scalar pi_boy_edu= (eta_edu- beta_edu+ gamma_boy_edu- (gamma_boy_edu*mean_male_children + gamma_girl_edu*mean_female_children)) / (beta_edu+ mean_w_edu)
display pi_boy_edu
scalar alpha_edu= _b[_cons]
scalar beta_edu= _b[ln_ppc_exp]
scalar eta_edu= _b[ln_hhsize]
scalar gamma_boy_edu= _b[male_children]
scalar gamma_girl_edu= _b[female_children]
matrix V_edu= e(V)
sum share_edu
scalar mean_w_edu= r(mean)
sum male_children female_children
scalar mean_male_children = r(mean)
scalar mean_female_children = r(mean)
scalar pi_boy_edu= (eta_edu- beta_edu+ gamma_boy_edu- (gamma_boy_edu*mean_male_children + gamma_girl_edu*mean_female_children)) / (beta_edu+ mean_w_edu)
display pi_boy_edu
Comment