Hypothetical default rate

[Sanne Otagho]

Dear Statalist,

I have a dataset containing loans which are followed through time.
I am running a logistic regression on a default variable which is 1 if a loan defaults and 0 otherwise. I'm investigating the treatment effect from participating in a refinance program so a dummy variable for the program is added in the regression.
I've runned the correct logistic model and have the regression results in log odd ratios. Now I want to determine the hypothetical default rate, if loans which participated in the program did not participate. So I want the default rate with the realized values of the other independent variables and the counterfactual (no treatment).

I tried some things with predict and adjust but that did not work out.

Could someone help me with figuring this out?

Kind regards,
Sanne

[Phil Bromiley]

If you're working with logistic but want to work in odds, logit might be easier. You should also look at the treatment manual.
I suspect you can replace the value of treatment for the loans that did participate with non participate and then run the predicted probabilities.