OnApproach, a leading provider of Big Data and analytics solutions to credit unions, has released a whitepaper detailing how the combination of vintage and loan-level models offer new insights to the world of Retail Lending Analytics.
The whitepaper, titled “Predicting and Pricing for Risk: Vintage Loan-Level Models” is based on an interview conducted with Dr. Joe Breeden, Chief Data Scientist at Deep Future Analytics. In the whitepaper, Dr. Breeden discusses the importance of Big Data & Analytics at credit unions with an emphasis on Predictive Analytics for risk-based loan pricing. Dr. Breeden discusses the changing perspective of retail lending which includes the reconsideration of the FICO score, viewing it more as a component of risk analysis instead of a method for risk analysis.
“People are recognizing that a score doesn’t get you there, not by itself. So having a sound vintage modeling approach as your starting point is essential,” Breeden said, “I have been building what you might call loan-level vintage models, combining vintage and scoring, and those have also been very well-reviewed. So my expectation is that as they become more widely adopted and more thoroughly studied, I hope to see those [models] become the future best practice.”
Dr. Joe Breeden brings more than 20 years of experience leading financial institutions through financial modeling, allowing them to achieve real understanding of portfolio dynamics for retail lending where those problems originated. He is a published author on the subject of retail lending analytics and is currently a partner in Deep Future Analytics (DFA), a CUSO that is focused on predictive analytics relating to loans.
Dr. Breeden has worldwide experience in creating predictive models which include the Mexican Peso Crisis, Asian Economic Crisis, 2001 Global Recession, Hong Kong SARS Recession, US Mortgage Crisis, and the Global Financial Crisis. He currently serves as the associate editor for the Journal of Credit Risk and the Journal of Risk Model Validation.
Materializing the Past: Vintage Loan-Level Models can be downloaded here.