The inclusion of time-various covariates into survival Evaluation has led to higher predictions of the time to default in behavioural credit score scoring products. Having said that, when these time-different covariates are endogenous, There's two main troubles: estimation bias on the survival model and not enough a prediction framework for future