Hello,
I am currently estimating a conditional logit which renders me negative marginal utility of income for some income ranges.
In Python I solved this problem in the past by imposing non-linear constraints on my maximisation problem that the marginal utility should be positive over a grid of points in which I evaluated the first derivative of my utility function.
The utility function is a third order polynomial in income and labour. Is there a way to do something similar in stata? The non-linear constraint would therefore be a second order polynomial in income and labour.
I saw that you can pass linear constraints to the ml command which is used by the clogit command, but I don't see this option for non-linear constraints.
Does anybody know if there is a maximisation routine that allows me to pass non linear constraints?
Best regards and many thanks in advance!
Hannes
I am currently estimating a conditional logit which renders me negative marginal utility of income for some income ranges.
In Python I solved this problem in the past by imposing non-linear constraints on my maximisation problem that the marginal utility should be positive over a grid of points in which I evaluated the first derivative of my utility function.
The utility function is a third order polynomial in income and labour. Is there a way to do something similar in stata? The non-linear constraint would therefore be a second order polynomial in income and labour.
I saw that you can pass linear constraints to the ml command which is used by the clogit command, but I don't see this option for non-linear constraints.
Does anybody know if there is a maximisation routine that allows me to pass non linear constraints?
Best regards and many thanks in advance!
Hannes
