CASE STUDY
Application scorecard engagement
How our application scorecard model supported a lender assess the creditworthiness of potential borrowers
We worked with a regional bank to develop a loan application scorecard model to help them predict and score the likelihood of credit risk for potential borrowers. The lender now uses these scores to make decisions about which loans to approve and which to deny.
Key results
Improved lending decisions
Minimised credit risk exposure
Streamlined lending process
Our approach
​
Data analysis
As a starting point, the bank provided our team with a large dataset of historical loan applications, along with information about which loans were repaid or went into arrears, and the characteristics of each borrower. By analysing the data, including looking at the borrower’s credit score, employment status, income, debt-to-income ratio and loan purpose, we were able to identify the variables that were most predictive of credit risk. Using statistical techniques such as logistic regression, we built a model that could predict the likelihood of a loan being approved based on these variables.
Testing
Once the model was built we tested it using a holdout sample of loan applications that weren’t used in the original analysis. The testing showed the model performed well in predicting loan approval, with a high degree of accuracy and a low rate of false positives and false negatives.
Model development
The next step was to work with the lender to develop a scoring system based on the model. We assigned a score to each loan application based on the borrower’s characteristics and the model’s prediction of credit risk.
Monitoring and refinements
Over time, we’ve monitored the performance of the scorecard to ensure that it remained accurate and effective. Adjustments have been made to the model and scoring system as necessary, based on new data and changing market conditions.
Find out more
​
Happy Prime can support you with a scorecard development framework to classify risk within your portfolio. Utilising in-house and external data sources, our approach ensures that the most appropriate methodology is being used to better understand both new and existing customers by creating an optimal assessment of their credit risk.