Hi all,
I am estimating a demand system but I have many expenditure shares which are zero. To deal with this, some authors have used a two step procedure: first a probit to estimate the probability of consumption. Second, they calculate the cumulative distribution (φ(.)) and normal probability density functions (φ(.)) for each food group. This is then used to generate new consumption shares for all food groups.
Do you know how I can implement this two step procedures in STATA? I have ten food groups (s1 to s10) and 6 demographic variables.
I am estimating a demand system but I have many expenditure shares which are zero. To deal with this, some authors have used a two step procedure: first a probit to estimate the probability of consumption. Second, they calculate the cumulative distribution (φ(.)) and normal probability density functions (φ(.)) for each food group. This is then used to generate new consumption shares for all food groups.
Do you know how I can implement this two step procedures in STATA? I have ten food groups (s1 to s10) and 6 demographic variables.