The Challenges
At the time of lending to a customer, it is important for the financial institution to be able to gauge his ability to pay his dues both in the short and longer term. While credit score or application score is used to assess the creditworthiness of the customer, it is only when combined with a robust income estimation model that lenders can measure their actual ability to pay. With digitalization taking over traditional lending practices in a big way, income estimation models are extremely crucial for a better customer experience at the time of onboarding as there is no hassle of providing income documents, and loan disbursals can happen much faster. From a lender’s perspective as well, income estimation models reduce the cost and effort of verification and manual collection of documents, thus truly leveraging the benefits of a digital lending process.
The Solution
The Income Estimation Model was developed through multiple phases of segmentation, target variable selection, independent variable creation and predictive model development
Segmentation
Different segmentation have been explored based on customer profile such as product holding (HL, AL, PL, CC, CD, etc.), no. of products, geography, etc.
Target Variable Selection
Various transformations were tried on the income of customers like normalization, log transformation, etc. to arrive at the best suited target variable for model development.
Independent factor creation
Over 3000 variables were created using customer’s credit history with the bureau as well as demographic information.
Model Development and Validation
For each identified segment, a model was developed and validated on different lender types, products and regions to test its robustness and stability.
Result
Customer experience is key in today’s lending space. Lenders are extending their products and services to various income segments where verified income proof is either unavailable or complicated to retrieve. Moreover, customers are looking for a smoother experience at the time of onboarding.
For the financial institutions too, collection and verification of income documents is a time and cost intensive process. By automating this through an income estimation models, lenders are able to optimize on cost as well as reduce turn-around-time, while assessing the riskiness of a customer with greater confidence.
The Income Estimation Model is currently being used by 50+ institutions for their customer onboarding and management
Our Post
Recent Posts
