I used logit2 command (stata) available on Mitchell Peterson page for 2 different models for an unbalanced panel.
http://www.kellogg.northwestern.edu/.../se/logit2.ado
I could get results for model 1 but model 2 could not be run due to NON CONVERGENCE/INABILITY TO MAXIMIZE LIKELIHOOD FUNCTION. I had included 2 variables in model2 (idiosyncratic risk and systematic risk), in addition to the ones used in Model1.
One possible source can be multicollinearity but the tests do not point towards this (I used Collin package). Can someone please enlighten me what I should recheck or what can be the possiblereason(s) for this problem.
http://www.kellogg.northwestern.edu/.../se/logit2.ado
I could get results for model 1 but model 2 could not be run due to NON CONVERGENCE/INABILITY TO MAXIMIZE LIKELIHOOD FUNCTION. I had included 2 variables in model2 (idiosyncratic risk and systematic risk), in addition to the ones used in Model1.
One possible source can be multicollinearity but the tests do not point towards this (I used Collin package). Can someone please enlighten me what I should recheck or what can be the possiblereason(s) for this problem.
