Case Study: Bank of America achieves 90% faster credit-risk scoring and forecasting with SAS

Bank of America avoids gridlock in credit risk scoring, forecasting

Bank of America’s Corporate Investments Group (CIG) is responsible for modeling credit risk across a large loan portfolio — including probability-of-default (PD) calculations on 9.5 million mortgages and forecasting losses for credit cards — while supporting roughly 59 million customer relationships. Adding credit‑card loss forecasting strained the bank’s shared‑services environment: multi‑terabyte datasets and long batch runs were slowing hedging decisions and limiting ad hoc analysis.

CIG moved to a dedicated platform running SAS Enterprise Risk Management on SAS Grid with SAS Scalable Performance Data Server on a 224‑core IBM BladeCenter and IBM XIV storage. The change cut PD calculation time from 96 hours to 4 hours (24x faster), reduced ad hoc job times by 90%, increased overall throughput (three times the prior speed), and sped some portfolio scoring jobs from 3 hours to 10 minutes — while handling 30 TB of source data and providing unfettered access for users.

Bank of America

Russell Condrich

Senior Vice President Corporate Investment Group