PT  - JOURNAL ARTICLE
AU  - Mark F. Milner
TI  - Using Predictive Analytics to Support MBS Servicing and Loss Mitigation
AID  - 10.3905/jsf.2010.16.2.028
DP  - 2010 Jul 31
TA  - The Journal of Structured Finance
PG  - 28--32
VI  - 16
IP  - 2
4099  - https://pm-research.com/content/16/2/28.short
4100  - https://pm-research.com/content/16/2/28.full
AB  - In order to minimize losses, mortgage servicers must identify and prioritize those troubled loans that pose the greatest risk to their portfolios. Drawing upon the rich stores of loan-level data residing in their servicing platforms, servicers can better identify their highest-risk loans though the use of predictive models and analytics. By segmenting their portfolios according to expected loss measures, mortgage servicers can more effectively target their limited resources where they have the potential to have the most risk mitigation impact. Predictive modeling allows servicers to project the impact a wide variety of corrective measures may have on the probability and severity of loss and choose their mitigation strategies accordingly. By developing servicer-defined or investor-defined goals and constraints, predictive modeling also enables servicers to structure optimized loan modifications.TOPICS: MBS and residential mortgage loans, CMBS and commercial mortgage loans