The inclusion of time-different covariates into survival Investigation has led to raised predictions of some time to default in behavioural credit score scoring versions. Nevertheless, when these time-various covariates are endogenous, there are two significant troubles: estimation bias of the survival model and not enough a prediction framework fo