Hello!
I am currently trying to use Mc Fadden's choice model in order to estimate how husband's and wife's labor supply are affected by fertility.
To do so I am using cmclogit. I am using as choice variable a categorical string variable consisting of the 25 possible combinations of labor hours division in the family (hwhh); as alternative variables I am using income (y), leisure time male (lm aka 80-hours worked male) and leisure time female (lf aka 80 hours worked female); and as case variables I am using number of children (nchild), number of childre under 6 (childu6), and education and age of both partners. I am trying to use a quadratic utility function for this.
The problem is that when I run the model I get infinite iterations with the same loglikelihood and the "not concave" message. Another huge proble is that so many of my variables get dropped because of collinearity.
cmset hhid hwhh
cmclogit observed y lm lf y2 lm2 lf2 y_lm y_lf lm_lf, casevars(nchild childu6 age_male age_female education_male e
> ducation_female)
note: lf omitted because of collinearity.
note: lm2 omitted because of collinearity.
note: lf2 omitted because of collinearity.
note: lm_lf omitted because of collinearity.
Iteration 0: Log likelihood = -763.11224 (not concave)
Iteration 1: Log likelihood = -706.72031 (not concave)
Iteration 2: Log likelihood = -698.39649 (not concave)
Iteration 3: Log likelihood = -693.98142 (not concave)
Iteration 4: Log likelihood = -693.43404 (not concave)
Iteration 5: Log likelihood = -692.52519 (not concave)
Iteration 6: Log likelihood = -692.11163 (not concave)
Iteration 7: Log likelihood = -692.0161 (not concave)
Iteration 8: Log likelihood = -691.93109 (not concave)
Iteration 9: Log likelihood = -691.88114 (not concave)
Iteration 10: Log likelihood = -691.84558 (not concave)
Iteration 11: Log likelihood = -691.82056 (not concave)
Iteration 12: Log likelihood = -691.81971 (not concave)
Iteration 13: Log likelihood = -691.81013 (not concave)
Iteration 14: Log likelihood = -691.78225 (not concave)
....
....
....
Iteration 139: Log likelihood = -691.52465 (not concave)
Iteration 140: Log likelihood = -691.52465 (not concave)
Iteration 141: Log likelihood = -691.52465 (not concave)
Iteration 142: Log likelihood = -691.52465 (not concave)
I am stuck and I don't know why this is happening or how to tackle it. If anyone knows how I should procede to find out what the issue is, it would be greatly appreciated.
Thank you a lot.
Annalisa
I am currently trying to use Mc Fadden's choice model in order to estimate how husband's and wife's labor supply are affected by fertility.
To do so I am using cmclogit. I am using as choice variable a categorical string variable consisting of the 25 possible combinations of labor hours division in the family (hwhh); as alternative variables I am using income (y), leisure time male (lm aka 80-hours worked male) and leisure time female (lf aka 80 hours worked female); and as case variables I am using number of children (nchild), number of childre under 6 (childu6), and education and age of both partners. I am trying to use a quadratic utility function for this.
The problem is that when I run the model I get infinite iterations with the same loglikelihood and the "not concave" message. Another huge proble is that so many of my variables get dropped because of collinearity.
cmset hhid hwhh
cmclogit observed y lm lf y2 lm2 lf2 y_lm y_lf lm_lf, casevars(nchild childu6 age_male age_female education_male e
> ducation_female)
note: lf omitted because of collinearity.
note: lm2 omitted because of collinearity.
note: lf2 omitted because of collinearity.
note: lm_lf omitted because of collinearity.
Iteration 0: Log likelihood = -763.11224 (not concave)
Iteration 1: Log likelihood = -706.72031 (not concave)
Iteration 2: Log likelihood = -698.39649 (not concave)
Iteration 3: Log likelihood = -693.98142 (not concave)
Iteration 4: Log likelihood = -693.43404 (not concave)
Iteration 5: Log likelihood = -692.52519 (not concave)
Iteration 6: Log likelihood = -692.11163 (not concave)
Iteration 7: Log likelihood = -692.0161 (not concave)
Iteration 8: Log likelihood = -691.93109 (not concave)
Iteration 9: Log likelihood = -691.88114 (not concave)
Iteration 10: Log likelihood = -691.84558 (not concave)
Iteration 11: Log likelihood = -691.82056 (not concave)
Iteration 12: Log likelihood = -691.81971 (not concave)
Iteration 13: Log likelihood = -691.81013 (not concave)
Iteration 14: Log likelihood = -691.78225 (not concave)
....
....
....
Iteration 139: Log likelihood = -691.52465 (not concave)
Iteration 140: Log likelihood = -691.52465 (not concave)
Iteration 141: Log likelihood = -691.52465 (not concave)
Iteration 142: Log likelihood = -691.52465 (not concave)
I am stuck and I don't know why this is happening or how to tackle it. If anyone knows how I should procede to find out what the issue is, it would be greatly appreciated.
Thank you a lot.
Annalisa
Comment