Analyzed delinquency risk of housing loan customers for a consumer bank

A comprehensive risk segmentation of the housing loan portfolio was conducted at both the account and the customer level by analyzing credit history data of over 20,000 active home loan customers excluding such accounts that were either closed or had turned dormant.

A holistic approach was used where all participants in a home loan account such as the main applicant, co-applicant, guarantor and associated persons were segmented into various risk buckets.

TransOrg used both internal data such as details of the customers that were available within the bank and external data such as past credit history of these customers, to make two different risk assessments:

• Analyzed the repayment patterns of prior loans taken by the customer both within and outside the bank.
• Conducted a vintage analysis of the proportion of customers in 30, 60 and 90 days past due (DPD) buckets and the distribution of risk within the entire home loan customer portfolio.

Client was generating data from several different sources throughout the customer journey hence TransOrg stitched the customer data from all available internal and external sources to create a customer 360° view.

Delinquency trends and customer payment patterns for various loans were tracked based on parameters such as:

1. Days Past Due (DPD) i.e., the number of days passed since the date when the repayment of loan was due.
2. Number of latest enquires made to the credit bureau i.e., to track the number of new loans that the customer maybe applying for in other banks or with the client while having an existing home loan account with the client.
3. The updated credit score of the customers as per the credit bureau.
4. Months on Book (MOB) or the number of months since the customer’s account opening date.

Each parameter was divided into three risk bands – High (3), Medium (2)
and Low (1). For example, a customer with more than three enquiries made to the credit bureau in the last three months was considered to be at a high risk of turning delinquent. Based on the distribution of customers in various bands TransOrg assigned weights to the parameters as per business requirements and changed the weightage periodically as per business use case. A weight
matrix was made for these variables, with ‘DPD in the past 6 months’ having the highest weight, and the ‘Number of Enquiries in the last 3 months’ having the lowest.

Finally, a composite risk value was assigned to each customer based on the sum product of their respective risk bands and the weightage of the respective risk parameter. Based on their composite risk scores the customers were divided into three risk buckets of High, Medium and Low risk.

TransOrg delivered tremendous business impact including key insights and suggestions such as:

• With a thorough risk segmentation of their customers the client can dig further and understand the average time between a loss event and the time when a problem was initially identified.
• Client regularized the customer portfolio risk segmentation process as a routine activity to be performed on a quarterly basis on its active customer portfolio. However, based on the business strategy, variables are refreshed on a quarterly basis by the client and the weights assigned to the parameters are altered accordingly.