Dear All,
I work on credit risk and use a dataset of application and performance data
for households. I check for selection and, since being approved for a loan
and default or no default are both binary variables, I was thinking about
using a heckprob model. I am concerned that coefficients are not necessarily
consistent. I heard people arguing that the heckman model produces
consistent estimates because of the OLS second stage equation, but that a
probit model does not necessarily provide consistent estimates. Other say
the exact opposite. Can you give some advice? Maybe you can give me some
references of models which were used to implement heckprob and give some
reasons as to why the estimates are consistent.
Thank you!
Best wishes,
Sascha
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
