I have an upcoming job interview as a statistician for a firm that sets credit card limits. In preparation, I'm trying to learn how credit limits are determined.
I know that there are vast data-sets where the best regression/ML model wins, but I'm looking for what is actually done in financial institutions and why. Any papers, insight, etc. are appreciated.
Also, how do firms correct for the significant selection bias caused by having performance data only on approved accounts?