Overview
Banking industry has been the early adopters of predictive modelling techniques starting with credit scoring system in late 50’s. Ever since, all banks in the world use credit score for their underwriting, account management etc. for their high-volume loan products. Now all the non-banking financial institutions, auto-finance etc. companies are also using these tools for automated credit screening purposes.
Application (A) Score / Credit Score: The purpose is to facilitate credit extension for new to bank customers. All entities (not necessarily banks) that do lending, have to have some kind of A Score aka Credit Score.
Behaviour Score: Used for account management purposes e.g. increase credit line, top-up loans, approve POS transactions, etc. Typically, since the entire portfolio comes under the purview of account management, an in-house tool Behaviour Score is typically used, otherwise cost becomes prohibitive if one uses external scores unless it is mandated by regulators.
Collection Score: Used to create collection queues to prioritise collection of outstanding amounts from delinquent customers. This score has a very specific usage which is collection queuing.
Large banks have their internal risk analytics team who builds these scorecards and they are primarily used for unsecured loans e.g. Credit Card, Personal Loan etc. Many others also uses external vendors to develop these models. Developing, maintaining and, if necessary, re-development of these tools is one of the major activities of the risk management departments. They are the primary automation tools to manage the high-volume loan products.
This course is part of the Artificial Intelligence and Data Science series offered by NUS-ISS.