Understanding credit risk using liability variables: Axis Bank

Abstract

In today’s cutthroat competition, it is very important that the bank manage its credit risk efficiently. This project is one step forward in that direction. Focus of this project was ETB customers. I have analyzed their behaviour and found out some variables which are key to understand their performance. Some of the findings of the project are 1. Vintage is a very good predictor of risk. Risk reduces as vintage increases. 2. Customers with higher Cash Debits as a percentage of total debits are riskier. 3. As the number of cash debits increases, so does the bad rate. 4. Debits (cash or direct) higher than average balance show higher risk. 5. Operating average balance shows good risk ordering. Higher the balance, lower the bad rate. 6. Customer with high credit deposits to low average balance ratio have a higher probability to default. 7. As total debits to total credits increases, risk increase. I have found good risk ordering among these variables. This information would be helpful for scorecard team to build scorecard models.