Preventing Fraud through Analytics

About the Program

Category: Human Services (Best in Category)

Year: 2015

The County of San Bernardino Transitional Assistance Department (TAD), in collaboration with the Statewide Automated Welfare System (SAWS) Consortium C-IV (C-IV) and Accenture, initiated the development of a analytics predictive model to prevent fraud and improve the efficiency of the county fraud investigation process. Statistical analyses were performed on a set of cases which had confirmed findings of fraud in the past. The data from these cases was obtained from the county’s SAWS C-IV database and the California Statewide Electronic Benefit Transfer (EBT) System. An analytics predictive model was developed from these analyses and applied to current Supplemental Nutritional Assistance Program (SNAP) cases. Investigations were performed on 500 random cases to determine the percentage of discrepancies found. The identification of discrepancies significantly increased from 6.1% prior to the application of analytics, to 20.8% with the predictive model in production. The use of analytics has resulted in increased fraud prevention activities, earlier identification of fraud and a proactive approach to identifying potential cases to investigate. Prior to leveraging analytics, fraud referrals were made by county staff or fraud was discovered through third party verification systems after the fraud had already occurred, and involved several months of discrepancies in each case.