Lenders are struggling not only to build the infrastructure needed to handle the rising number of defaults, but also to identify borrowers with the greatest chance of becoming delinquent before they fall behind on their payments. We believe the science of predictive modeling, which financial institutions use on a variety of fronts today, can help mortgage lenders with both tasks.

Predictive modeling applies data mining techniques and algorithms to create mathematical formulas used to forecast and segment future events, such as identifying borrowers with a greater propensity to move to foreclosure in the next 90 days. By identifying such borrowers early in the pre-collections lifecycle, lenders can focus on their high risks and tailor their mitigation efforts accordingly. For example, instead of having pre-collections representatives directly contact all potentially delinquent borrowers — a costly process, to be sure — lenders can employ predictive modeling to segment those borrowers into multiple risk levels. The most likely to default could receive an active treatment, such as a live telephone call once they are past due, while the least likely to default may simply receive a passive treatment, such as a letter.

Giving lenders the ability to sift through thousands of loans and focus on those most at risk, predictive modeling can:

• Result in improved pre-collections outcomes
• Help people stay in their homes
• Improve the profitability of lenders through the effective allocation of resources and reduced charge-offs

Download the PDF below to learn more about the science of predictive modeling.