Nationwide, the UK’s largest building society, has chosen credit risk and Artificial Intelligence (AI) experts Jaywing, to work with its analytical team on a pioneering analysis of different risk modelling techniques, as it explores the application of machine learning for application scoring.
Nationwide was seeking to boost the performance of the application risk models it uses to approve existing customers for unsecured loans. Recognising industry concerns around the transparency of some AI-based modelling techniques, it wanted to understand how Archetype, Jaywing’s proprietary AI-based modelling software, performed compared to other approaches.
Assessing the performance of Archetype against traditional linear regression showed that Archetype was able to generate a 6% uplift in the predictive power of the model, offering huge potential for reductions in bad debt or an increase in acceptance rates. The results were comparable with other AI-based modelling approaches, and the Archetype models were backed by Jaywing’s ability to control the neural network and explain the resulting model.
Commenting on the results, Matthew Jones, Head of Retail Modelling at Nationwide said: “Archetype clearly represents a powerful alternative in building credit scores. We were impressed by the ease of use, the rapidly-realised benefits and ability to enforce intuitive behaviour on key variables in neural networks. Jaywing’s approach and depth of knowledge has made us confident that the use of neural networks and similar technologies has reached a level of maturity where it can be deployed safely within the credit risk arena.”
As Archetype offers the analyst control over how the input variables are used, automatically producing charts and outputs which explain how the model behaves, it was easy for Jaywing to provide Nationwide a view of the most influential variables in the model and demonstrate how each data input had contributed to the overall performance. And with governance taking a key role in the design of the models, Archetype enables lenders to deploy better, compliant models, more frequently, and with less effort.
Nevan McBride, Risk Practice Director at Jaywing, commented: “Across the whole credit lifecycle, from application and behavioural scoring through to debt collection, we’ve seen Archetype produce significant score improvements every time.
“In Nationwide’s case, we’ve also proven that Archetype can go head-to-head with other AI-based techniques and get the similar level of benefit, including where the data that is fully constrained to guarantee that the resulting model will not produce erroneous results, even on unseen cases. We look forward to working with Nationwide as it explores other areas of machine learning modelling.”